Handbook of Adhesion Technology [2 ed.] 9783319554105, 9783319554112

Citation preview

Lucas F. M. da Silva Andreas Öchsner Robert D. Adams Editors

Handbook of Adhesion Technology Second Edition

Handbook of Adhesion Technology

Lucas F. M. da Silva • Andreas Öchsner Robert D. Adams Editors

Handbook of Adhesion Technology Second Edition

With 919 Figures and 106 Tables

Editors Lucas F. M. da Silva Department of Mechanical Engineering Faculty of Engineering University of Porto Porto, Portugal

Andreas Öchsner Faculty of Mechanical Engineering Esslingen University of Applied Sciences Esslingen, Germany

Robert D. Adams Department of Mechanical Engineering University of Bristol Bristol, UK Department of Engineering Science University of Oxford Oxford, UK

ISBN 978-3-319-55410-5 ISBN 978-3-319-55411-2 (eBook) ISBN 978-3-319-55412-9 (print and electronic bundle) https://doi.org/10.1007/978-3-319-55411-2 Library of Congress Control Number: 2018940005 1st edition: # Springer-Verlag Berlin Heidelberg 2011 # Springer International Publishing AG, part of Springer Nature 2018 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Printed on acid-free paper This Springer imprint is published by the registered company Springer International Publishing AG part of Springer Nature. The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Second Edition Preface

The first edition of Handbook of Adhesion Technology has been very well accepted by the adhesion community with over 25,000 downloads per year. This was the main motivation for the preparation of this 2nd edition. Also, there has been a general increase of the scientific and industrial community in this technology, which justifies an update. There are more papers published every year, there are new journals related to the field, there are more national and international conferences, and there are more applications. A revolution is currently taking place in the transport industry where fuel is being replaced by electric power. This change means that lighter structures need to be manufactured with the use of light materials and more efficient joining methods such as adhesive bonding. But, the use of adhesive bonding is not limited to the transport industry, and there are several other applications described in this handbook such as the use of adhesives in medicine. This 2nd edition of Handbook of Adhesion Technology is a complete revision of the 1st edition with an update of the methods that have been investigated recently, which are now fully accepted by the adhesion community. Themes that are now included in more detail include, for example, hybrid adhesives used for automotive applications, ecofriendly surface treatments, damage mechanics, joint durability prediction, functionally graded joints, adhesive selection for space applications, and, lastly, trends in bioadhesion. There is also a new chapter related to the application of adhesives in the oil industry. Besides these content changes, there has been a complete revision of all chapters in terms of text, figures, tables, and references for a more didactic character of this reference book. The editors would like to thank the authors for their patience with the preparation of this handbook. Finally, the editors especially thank Dr. Christoph Baumann, Ms. Tina Shelton, and Ms. Monika Garg, Springer editors, who helped enormously toward the success of this handbook. Department of Mechanical Engineering Faculty of Engineering University of Porto Porto, Portugal

Lucas F. M. da Silva

v

vi

Second Edition Preface

Faculty of Mechanical Engineering Esslingen University of Applied Sciences Germany

Andreas Öchsner

Department of Mechanical Engineering University of Bristol Bristol, UK

Robert D. Adams

Department of Engineering Science University of Oxford Oxford, UK

First Edition Preface

Adhesives have been used for thousands of years, but until 100 years ago, the vast majority was from natural products such as bones, skins, fish, milk, and plants. Since about 1900, adhesives based on synthetic polymers have been introduced, and today, there are many industrial uses of adhesives and sealants. It is difficult to imagine a product—in the home, in industry, in transportation, or anywhere else for that matter—that does not use adhesives or sealants in some manner. Adhesion technology is nowadays a common approach to many joining situations. There are many books and several international journals dealing with this subject. However, when the end-user needs to apply this technology, he does not have the time or the resources to thoroughly study the various technologies related to adhesives and sealants. It is not practical to go through volumes of text and product information looking for the specific methods or processes to apply. However, such information should be close-by if the need arises. The Handbook of Adhesion Technology is intended to fill a gap between the necessarily simplified treatment of the student textbook and the full and thorough treatment of the research monograph and review article. The subject is treated very comprehensively and with the most up to date information so that the end-user has in a single book a proper guidance and fundamental knowledge required for many adhesive bonding or sealing applications. The Handbook of Adhesion Technology is intended to be a book of reference in the field of adhesion. Essential information is provided for all those concerned with the adhesion phenomenon. Adhesion is a phenomenon of interest in diverse scientific disciplines and of importance in a wide range of technologies. Therefore, this handbook includes the background science (physics, chemistry, and materials science), engineering aspects of adhesion, and industry specific applications. It is arranged in a user-friendly format with ten main sections: theory of adhesion, surface treatments, adhesive and sealant materials, testing of adhesive properties, joint design, durability, manufacture, quality control, applications, and emerging areas. The sections contain each about five chapters written by internationally renowned authors who are authorities in their fields. This Handbook is intended to be a reference for people needing a quick, but authoritative, description of topics in the field of adhesion and the use of adhesives and sealants. It is intended for scientists and engineers of many different backgrounds who need to have an understanding of various aspects of adhesion vii

viii

First Edition Preface

technology. These will include those working in research or design, as well as others involved with marketing services. It is expected to be a valuable resource for both undergraduate and research students. The editors would like to thank the authors for their patience with the preparation of this Handbook. Finally, the editors especially thank Christoph Baumann and Tina Shelton, Springer editors, who helped enormously toward the success of this Handbook. Porto, Portugal Skudai, Johor, Malaysia Oxford, UK

Lucas F. M. da Silva Andreas Öchsner Robert D. Adams

Contents

Volume 1 1

Introduction to Adhesive Bonding Technology . . . . . . . . . . . . . . . . Lucas F. M. da Silva, Andreas Öchsner, and Robert D. Adams

Part I

Theory of Adhesion

1

.................................

9

2

Theories of Fundamental Adhesion David E. Packham

........................

11

3

Forces Involved in Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Maurice Brogly

43

4

Wetting of Solids . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Martin E. R. Shanahan and Wulff Possart

71

5

Spreading of Liquids on Substrates . . . . . . . . . . . . . . . . . . . . . . . . Günter Reiter

101

6

Thermodynamics of Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . Wulff Possart and Martin E. R. Shanahan

115

Part II

Surface Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

129

7

General Introduction to Surface Treatments . . . . . . . . . . . . . . . . . Gary Critchlow

131

8

Surface Treatments of Selected Materials . . . . . . . . . . . . . . . . . . . . Guy D. Davis

163

9

Surface Characterization and Its Role in Adhesion Science and Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . John F. Watts

197

ix

x

10

11

Contents

Use of Surface Analysis Methods to Probe the Interfacial Chemistry of Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . John F. Watts Organosilanes: Adhesion Promoters and Primers . . . . . . . . . . . . . Marie-Laure Abel

Part III

Adhesive and Sealant Materials

.....................

227 257

281

12

Classification of Adhesive and Sealant Materials . . . . . . . . . . . . . . Erol Sancaktar

283

13

Composition of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hyun-Joong Kim, Dong-Hyuk Lim, Hyeon-Deuk Hwang, and Byoung-Ho Lee

319

14

Adhesive Families . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Eric Papon

345

15

Pressure-Sensitive Adhesives (PSAs) Charles W. Paul and Eric Silverberg

.......................

373

16

Selection of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ewen J. C. Kellar

409

Part IV

Testing of Adhesive Properties . . . . . . . . . . . . . . . . . . . . . .

431

17

Physical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David A. Dillard

433

18

Thermal Properties of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . John Comyn

459

19

Failure Strength Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Lucas F. M. da Silva, Ricardo J. C. Carbas, and Mariana D. Banea

489

20

Fracture Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Bamber R. K. Blackman

523

21

Impact Tests Luca Goglio

..........................................

555

22

Special Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . David A. Dillard and Tetsuo Yamaguchi

593

Part V 23

Joint Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Constitutive Adhesive and Sealant Models . . . . . . . . . . . . . . . . . . . Erol Sancaktar

613 615

Contents

xi

24

Analytical Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Liyong Tong and Quantian Luo

665

25

Numerical Approach: Finite Element Analysis Ian A. Ashcroft and Aamir Mubashar

...............

701

26

Special Numerical Techniques to Joint Design . . . . . . . . . . . . . . . . Andreas Öchsner

741

27

Design Rules and Methods to Improve Joint Strength . . . . . . . . . . Lucas F. M. da Silva, Eduardo A. S. Marques, and Raul D. S. G. Campilho

773

28

Design with Sealants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Gregory L. Anderson

811

29

Design for Impact Loads . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Chiaki Sato

829

30

Vibration Damping of Adhesively Bonded Joints . . . . . . . . . . . . . . Robert D. Adams, Duncan G. A. Cooper, and Stuart Pearson

853

Volume 2 Part VI

Durability

.......................................

877

31

Effect of Water and Mechanical Stress on Durability Ian A. Ashcroft, John Comyn, and Aamir Mubashar

..........

879

32

Adhesives in Space Environment . . . . . . . . . . . . . . . . . . . . . . . . . . Sabine Dagras, Julien Eck, Claire Tonon, and Denis Lavielle

915

33

Fatigue Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Ian A. Ashcroft

941

34

Creep Load Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Paul Ludwig Geiß and Melanie Schumann

975

35

Durability of Nonstructural Adhesives . . . . . . . . . . . . . . . . . . . . . . 1009 James D. Palmer

Part VII

Manufacture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1029

36

Storage of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1031 Hans K. Engeldinger and Cai R. Lim

37

Preparation for Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1051 Andreas Lutz

xii

Contents

38

Equipment for Adhesive Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . 1091 Manfred Peschka

39

Environment and Safety . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1117 Ansgar van Halteren

Part VIII

Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1127

. . . . . . . . . . . . . . . . . . . . . . . . . 1129

40

Quality Control of Raw Materials Kazutami Wakabayashi

41

Processing Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1151 Kosuke Haraga

42

Nondestructive Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1171 Robert D. Adams

43

Techniques for Post-fracture Analysis Romain Créac’hcadec

Part IX

Applications

. . . . . . . . . . . . . . . . . . . . . . 1195

.....................................

1233

44

Adhesively Bonded Joints in Aircraft Structures . . . . . . . . . . . . . . 1235 L. John Hart-Smith

45

Aerospace Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1285 Christian Désagulier, Patrick Pérés, and Guy Larnac

46

Automotive Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1333 Klaus Dilger, Bernd Burchardt, and Michael Frauenhofer

47

Adhesive Bonding for Railway Application . . . . . . . . . . . . . . . . . . 1367 Yasuaki Suzuki

48

Marine Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1391 Peter Davies

49

Civil Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1419 Stefan Böhm, Martin Kahlmeyer, Andreas Winkel, and Ilko Hartung

50

Electrical Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1449 Kwang-Seok Kim, Jong-Woong Kim, and Seung-Boo Jung

51

Shoe Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1483 José Miguel Martín-Martínez

52

Oil Industry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1533 Silvio de Barros, Luiz C. M. Meniconi, Valber A. Perrut, and Carlos E. Reuther de Siqueira

Contents

Part X

xiii

Emerging Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1557

53

Molecular Dynamics Simulation and Molecular Orbital Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1559 Ya-Pu Zhao, Feng-Chao Wang, and Mei Chi

54

Bioadhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1597 Katharina Richter, Ingo Grunwald, and Janek von Byern

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1641 Stanislav N. Gorb and Lars Heepe

56

Adhesives with Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1677 Ambrose C. Taylor

57

Adhesive Dentistry John W. Nicholson

58

Adhesion in Medicine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1729 Robin A. Chivers

59

Recycling and Environmental Aspects . . . . . . . . . . . . . . . . . . . . . . 1751 Chiaki Sato

60

Adhesion Technology Recap: Current and Emerging Areas . . . . . 1775 Lucas F. M. da Silva, Andreas Öchsner, and Robert D. Adams

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1703

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1785

About the Editors

Lucas F. M. da Silva is currently Associate Professor with Aggregation at the Department of Mechanical Engineering of the Faculty of Engineering of the University of Porto and Director of the Integrated Master in Mechanical Engineering. He obtained his Ph.D. in bonding of composites from the University of Bristol (UK) in 2004. He leads the Adhesives Group, composed of post-docs, Ph.D. students, and M.Sc. students. He has published 206 ISI papers (174 as author and 32 as editor) and 23 books (9 as author and 14 as editor) mainly on adhesive joint mechanics. His papers were cited 4690 times and correspond to an h-index of 37 (SCOPUS, 06/02/2018). One of his papers obtained the SAGE Best Paper Award 2010 and Donald Julius Groen Prize 2010 (both awards given by the Institution of Mechanical Engineers). He received in 2013 and 2018 the Award of Scientific Excellence by the Faculty of Engineering of the University of Porto. He is editor-in-chief of The Journal of Adhesion, Proceedings of the Institution of Mechanical Engineers, Part L: Journal of Materials: Design and Applications, and University of Porto Journal of Engineering. He is also co-editor of two Springer book series (Advanced Structured Materials and Springer Briefs in Engineering: Computational Mechanics). He is member of the editorial board of International Journal of Adhesion and Adhesives and Journal of Adhesion Science and Technology. He is reviewer of 60 ISI journals. He organizes international conferences on adhesive bonding (Structural Adhesive Bonding every odd year and Industrial Applications of Adhesive Bonding every even year) and materials (Materials Design and Applications every even year), and founded the Portuguese xv

xvi

About the Editors

Adhesion Society that belongs to European Adhesion Societies Group (EURADH). He has seven pending patents. He developed a software for designing adhesive joints available online ( jointdesigner). He is consultant of several international companies (e.g., Alstom, Nagase Chemtex, John Deere). Andreas Öchsner is Full Professor for lightweight design and structural simulation at the Esslingen University of Applied Sciences, Germany. Having obtained a Diploma Degree (Dipl.-Ing.) in Aeronautical Engineering at the University of Stuttgart (1997), Germany, he spent the time from 1997 to 2003 at the University of Erlangen-Nuremberg as a research and teaching assistant to obtain his Doctor of Engineering Sciences (Dr.-Ing.). From 2003 to 2006, he worked as Assistant Professor in the Department of Mechanical Engineering and Head of the Cellular Metals Group affiliated with the University of Aveiro, Portugal. He spent 7 years (2007–2013) as a Full Professor at the Department of Applied Mechanics, Technical University of Malaysia, where he was also Head of the Advanced Materials and Structure Lab. From 2014 to 2017, he was a Full Professor at the School of Engineering, Griffith University, Australia, and Leader of the Mechanical Engineering Program (Head of Discipline and Program Director). His research interests are related to experimental and computational mechanics, cellular metals and thin structures and interphases. He has published over 450 scientific publications, comprising 13 research monographs, 23 book chapters, and four teaching books on finite element methods. He obtained more than 3000 citations in Google Scholar. He is the general Chairman of 12 international conferences on computational and experimental engineering (ACE-X series) and 14 international conferences in the area of heat and mass transfer (DSL series). His editorial work comprises posts as editor-in-chief of the international journal Continuum Mechanics and Thermodynamics (Springer), editor-inchief of the Springer book series on Advanced Structured Materials, and editor of SpringerBriefs in Applied Sciences and Technology: Computational Mechanics. His research activities were recognized in 2010 by the

About the Editors

xvii

award of a higher doctorate degree (D.Sc.) by the University of Newcastle, Australia. Robert D. Adams got his first degree, in Mechanical Engineering, from Imperial College, London, in 1962, and his Ph.D. from Cambridge University in 1967 and an Sc.D. in 2017. He was awarded a D.Sc. (Eng) from London University in 1986. Most of his academic career (1967–2005) was at Bristol University where he was made Full Professor in 1986. He was Head of the Department of Mechanical Engineering from 1994 to 1998 and Graduate Dean from 1998 to 2004. In 1991, he was awarded the Visiting Foreign Franqui Chair and Medal at the Free University of Brussels (VUB). He is now an Emeritus Professor of Applied Mechanics at the University of Bristol and a Visiting Professor at the Universities of Oxford and Oxford Brookes. He has published and edited several books and published about 200 papers in international refereed journals. In 2011, he was awarded the R.L. Patrick Fellowship of the US Adhesion Society. He is active in organizing conferences and publishing and was a founder member of EURADH and WCARP, the main European and World series on Adhesion. He has been joint editor-inchief of the International Journal of Adhesion and Adhesives since 1999. His main research area is on adhesively bonded joints, and he pioneered the application of finite element analysis for determining the stresses, strains, and strength of such joints. In addition to his adhesives research, he has also worked extensively on vibration properties in composites and developed low-velocity impact tests for nondestructively testing composites and sandwich structures.

Contributors

Marie-Laure Abel Faculty of Engineering and Physical Sciences (A1), University of Surrey, Guildford, Surrey, UK Robert D. Adams Department of Mechanical Engineering, University of Bristol, Bristol, UK Department of Engineering Science, University of Oxford, Oxford, UK Gregory L. Anderson R&D Laboratories, Propulsion Systems – Launch Systems Group, Orbital ATK, Corinne, UT, USA Ian A. Ashcroft Faculty of Engineering, University of Nottingham, Nottingham, UK Mariana D. Banea CEFET/RJ – Federal Center of Technological Education in Rio de Janeiro, Rio de Janeiro, Brazil Bamber R. K. Blackman Department of Mechanical Engineering, Imperial College London, South Kensington Campus, London, UK Stefan Böhm Department for Cutting and Joining Manufacturing Processes, University of Kassel, Institute for Production Technologies and Logistics, Kassel, Germany Maurice Brogly ENSCMu – LPIM – Equipe Chimie et Physico-Chimie des Polymères, Université de Haute Alsace, Mulhouse, France Bernd Burchardt Weiningen Zürich, Switzerland Raul D. S. G. Campilho Instituto Superior de Engenharia do Porto (ISEP), Instituto Politécnico do Porto, Porto, Portugal Ricardo J. C. Carbas Institute of Science and Innovation in Mechanical and Industrial Engineering (INEGI), Faculty of Engineering, University of Porto, Porto, Portugal Mei Chi State Key Laboratory of Nonlinear Mechanics (LNM), Institute of Mechanics, Chinese Academy of Sciences, Beijing, China Robin A. Chivers York, North Yorkshire, UK xix

xx

Contributors

John Comyn Materials Department, Loughborough University, Leicestershire, UK Romain Créac’hcadec ENSTA Bretagne, Institut de Recherche Dupuy de Lôme, FRE CNRS 3744, Brest, France Assemblages Multi-Matériaux, Institut de Recherche Dupuy De Lôme, Brest, France Duncan G. A. Cooper Department of Mechanical Engineering, University of Bristol, Bristol, UK Gary Critchlow Department of Materials, Loughborough University, Loughborough, Leicestershire, UK Lucas F. M. da Silva Department of Mechanical Engineering, Faculty of Engineering, University of Porto, Porto, Portugal Sabine Dagras Airbus Defence and Space, Toulouse, France Peter Davies Materials and Structures Group, Marine Structures Laboratory, Brest Centre, IFREMER, (French Ocean Research Institute), Plouzané, France Guy D. Davis ElectraWatch, Inc., Linthicum Heights, MD, USA Silvio de Barros Department of Mechanical Engineering, Federal Center of Technological Education in Rio de Janeiro – CEFET/RJ, Rio de Janeiro, RJ, Brazil Carlos E. Reuther de Siqueira Department of Offshore Engineering, Petróleo Brasileiro S.A. – PETROBRAS, Rio de Janeiro, RJ, Brazil Christian Désagulier ArianeGroup, Les Mureaux, France Klaus Dilger Institute of Welding and Joining, TU Braunschweig, Braunschweig, Germany David A. Dillard Biomedical Engineering and Mechanics Department, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA Julien Eck Airbus Defence and Space, Toulouse, France Hans K. Engeldinger Product Development HAF Tapes, tesa SE, Norderstedt, Germany Michael Frauenhofer Audi AG, Ingolstadt, Germany Paul Ludwig Geiß Faculty of Mechanical and Process Engineering, Workgroup Materials and Surface Technologies (AWOK), University of Kaiserslautern, Kaiserslautern, Germany Luca Goglio Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino, Italy Stanislav N. Gorb Department of Functional Morphology and Biomechanics, Zoological Institute at the University of Kiel, Kiel, Germany

Contributors

xxi

Ingo Grunwald Department of Adhesive Bonding Technology and Surfaces, Adhesives and Polymer Chemistry, Fraunhofer Institute for Manufacturing Technology and Advanced Materials (IFAM), Bremen, Germany Kosuke Haraga Tapes and Adhesives Department, Electronic Materials Division, Electronic Materials Business Unit, Denki Kagaku Kogyo Kabushiki Kaisha, Chuoku, Tokyo, Japan L. John Hart-Smith Boeing Company, Long Beach, CA, USA Ilko Hartung Dr. Kornder Anlagen- und Messtechnik GmbH & Co. KG, Bergheim, Germany Lars Heepe Department of Functional Morphology and Biomechanics, Zoological Institute at the University of Kiel, Kiel, Germany Hyeon-Deuk Hwang Laboratory of Adhesion & Bio-Composites, Program in Environmental Materials Science, Seoul National University, Seoul, Republic of Korea Seung-Boo Jung School of Advanced Materials Science & Engineering, Sungkyunkwan University, Suwon, Gyeonggi-do, Republic of Korea Martin Kahlmeyer Department for Cutting and Joining Manufacturing Processes, University of Kassel, Institute for Production Technologies and Logistics, Kassel, Germany Ewen J. C. Kellar TWI Ltd. (The Welding Institute), Cambridge, Cambridgeshire, UK Hyun-Joong Kim Laboratory of Adhesion & Bio-Composites, Program in Environmental Materials Science, Seoul National University, Seoul, Republic of Korea Jong-Woong Kim School of Advanced Materials Engineering, Chonbuk National University, Jeonju, Jeollabuk-do, Republic of Korea Kwang-Seok Kim Carbon & Light Materials Application Group, Korea Institute of Industrial Technology, Jeonju, Jeollabuk-do, Republic of Korea Guy Larnac ArianeGroup, Saint Médard en Jalles, France Denis Lavielle Airbus Defence and Space, Toulouse, France Byoung-Ho Lee Laboratory of Adhesion & Bio-Composites, Program in Environmental Materials Science, Seoul National University, Seoul, Republic of Korea Cai R. Lim Research & Development, tesa SE Hamburg, Norderstedt, Germany Dong-Hyuk Lim Laboratory of Adhesion & Bio-Composites, Program in Environmental Materials Science, Seoul National University, Seoul, Republic of Korea Quantian Luo School of Aerospace, Mechanical and Mechatronic Engineering, The University of Sydney, Sydney, NSW, Australia

xxii

Contributors

Andreas Lutz Automotive, Dow Europe GmbH, R&D Automotive Systems, Horgen, Switzerland Eduardo A. S. Marques Instituto de Ciência e Inovação em Engenharia Mec^anica e Engenharia Industrial (INEGI), Porto, Portugal José Miguel Martín-Martínez Adhesion and Adhesives Laboratory, University of Alicante, Alicante, Spain Luiz C. M. Meniconi Research and Development Center, CENPES, Petróleo Brasileiro S.A. – PETROBRAS, Rio de Janeiro, RJ, Brazil Aamir Mubashar Mechanical Engineering, Middle Eastern Technical University, Northern Cyprus Campus, Akara, Turkey John W. Nicholson Bluefield Centre for Biomaterials, London, UK Dental Physical Sciences, Institute of Dentistry, Queen Mary University of London, London, UK Andreas Öchsner Faculty of Mechanical Engineering, Esslingen University of Applied Sciences, Esslingen, Germany David E. Packham Materials Research Centre, University of Bath, Bath, UK James D. Palmer Staffordshire, UK Eric Papon Laboratoire des Polymères Organiques – UMR CNRS 5629, University of Bordeaux, Pessac, France Charles W. Paul Henkel Adhesives, Bridgewater, NJ, USA Stuart Pearson Department of Mechanical Engineering, University of Bristol, Bristol, UK Patrick Pérés ArianeGroup, Saint Médard en Jalles, France Valber A. Perrut Research and Development Center, CENPES, Petróleo Brasileiro S.A. – PETROBRAS, Rio de Janeiro, RJ, Brazil Metallurgical and Materials Engineering Department, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil Manfred Peschka Fraunhofer Institute, Bremen, Germany Wulff Possart Lehrstuhl Adhäsion und Interphasen in Polymeren, Universität des Saarlandes, Saarbrücken, Germany Günter Reiter Physikalisches Institut, Fakultät für Mathematik und Physik, Albert-Ludwigs Universität Freiburg, Freiburg, Germany Katharina Richter Department of Adhesive Bonding Technology and Surfaces, Adhesives and Polymer Chemistry, Fraunhofer Institute for Manufacturing Technology and Advanced Materials (IFAM), Bremen, Germany

Contributors

xxiii

Erol Sancaktar Department of Polymer Engineering, University of Akron, Akron, OH, USA Chiaki Sato Precision and Intelligence Laboratory, Tokyo Institute of Technology, Midori-ku, Yokohama, Japan Melanie Schumann Faculty of Mechanical and Process Engineering, Workgroup Materials and Surface Technologies (AWOK), University of Kaiserslautern, Kaiserslautern, Germany Martin E. R. Shanahan Institut de Mécanique et d’Ingénierie-Bordeaux (I2M), CNRS UMR 5295, Université de Bordeaux, Talence, France Eric Silverberg Henkel Adhesives, Bridgewater, NJ, USA Yasuaki Suzuki Suzuki Adhesion Institute of Technology, Ichinomiya, Aichi, Japan Ambrose C. Taylor Department of Mechanical Engineering, Imperial College London, London, UK Liyong Tong School of Aerospace, Mechanical and Mechatronic Engineering, The University of Sydney, Sydney, NSW, Australia Claire Tonon Airbus Defence and Space, Toulouse, France Ansgar van Halteren Industrieverband Klebstoffe e.V. RWI-Haus, Düsseldorf, Germany Janek von Byern Austrian Cluster for Tissue Regeneration, Ludwig Boltzmann Institute for Experimental and Clinical Traumatology, Vienna, Austria Faculty of Life Science, Core Facility Cell Imaging and Ultrastructure Research, University of Vienna, Vienna, Austria Kazutami Wakabayashi APS Research Co Ltd, Osaka, Japan Feng-Chao Wang State Key Laboratory of Nonlinear Mechanics (LNM), Institute of Mechanics, Chinese Academy of Sciences, Beijing, China John F. Watts The Surface Analysis Laboratory, Faculty of Engineering and Physical Sciences, University of Surrey, Guildford, Surrey, UK Andreas Winkel Department for Cutting and Joining Manufacturing Processes, University of Kassel, Institute for Production Technologies and Logistics, Kassel, Germany Tetsuo Yamaguchi Machine Elements and Design Engineering Laboratory, Department of Mechanical Engineering, School of Engineering, Kyushu University, Fukuoka, Japan Ya-Pu Zhao State Key Laboratory of Nonlinear Mechanics (LNM), Institute of Mechanics, Chinese Academy of Sciences, Beijing, China

1

Introduction to Adhesive Bonding Technology Lucas F. M. da Silva, Andreas Öchsner, and Robert D. Adams

Contents 1.1 Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.2 Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.3 Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1.4 Organization of the Handbook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 4 5 6 7

Abstract

This introductory chapter gives a brief description of adhesive bonding and adhesion-related phenomena. The major definitions of the terms associated to this technology such as adhesion, cohesion, adhesives, sealants, and adherends are given so that there is uniformity of language throughout the handbook. The reasons that drove the editors to prepare and revise this book are explained. Finally, the organization of the book is described.

L. F. M. da Silva (*) Department of Mechanical Engineering, Faculty of Engineering, University of Porto, Porto, Portugal e-mail: [email protected] A. Öchsner Faculty of Mechanical Engineering, Esslingen University of Applied Sciences, Esslingen, Germany e-mail: [email protected] R. D. Adams Department of Mechanical Engineering, University of Bristol, Bristol, UK Department of Engineering Science, University of Oxford, Oxford, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_1

1

2

1.1

L. F. M. da Silva et al.

Background

Adhesives have been used for thousands of years, but until 100 years ago, the vast majority was from natural products such as bones, skins, fish, milk, and plants. Since about 1900, adhesives based on synthetic polymers have been introduced, and today, there are many industrial uses of adhesives and sealants. It is difficult to imagine a product – in the home, in industry, in transportation, or anywhere else for that matter – that does not use adhesives or sealants in some manner. It is possible to distinguish three stages in the formation of a bonded joint. Initially, the adhesive must be in the solid state so that it can spread easily on the surface and wet properly the adherends to be bonded so that an intimate molecular contact is created between the adhesive and the surface. Secondly, in order for the adhesive to support the loads that it will encounter in service, the liquid adhesive must harden. Often, the adhesive is initially in the form of a monomer and polymerizes into a polymer with a high molecular weight. Pressure-sensitive adhesives are an exception in that they do not harden but remain permanently sticky. Finally, it is necessary to understand that the capacity of the joints to carry the load and the durability of the joint are affected by various factors such as the joint design, the way the loads are applied, and the environment that the joint will be subjected to. Thus, to obtain good results with adhesive bonding technology, it is necessary to have knowledge in various sciences including surface chemistry, chemistry and physics of polymers, materials engineering, mechanical engineering, etc. (Petrie 2000). Adhesively bonded joints are an increasing alternative to mechanical joints in engineering applications and provide many advantages over conventional mechanical fasteners. They provide a more uniform stress distribution along the bonded area which enables to have a higher stiffness and load transmission, reducing the weight and thus the cost. Figure 1 shows how the stress distribution in a bonded joint can give a higher stiffness and more uniform stress distribution in relation to a riveted joint. Due to the polymeric nature of the adhesive, adhesive joints provide good damping properties which also enable to have high fatigue strength. Adhesives can bond dissimilar materials with different coefficients of thermal expansion because the adhesive flexibility can compensate the difference. They bond thin plates very efficiently being one of the major applications of structural adhesives. Adhesives have a strength well below that of metals; however, when used to bond thin plates with a large bearing area, the strength of the adhesive is sufficient for structural applications. The adhesive application can be very efficient because it is easy to create an automatic process. The joint design is very flexible and new concepts and materials can be introduced. A good example is the sandwich structures where the core is made of honeycomb and the skin of a composite. They enable to have very smooth surface finishes because they avoid the use of holes for rivets or bolts or welding marks. They create an intimate contact between the bonded surfaces which is good in structural terms and also for corrosion resistance.

1

Introduction to Adhesive Bonding Technology

3

Fig. 1 Improved stiffness (left) and stress distribution (right) of adhesively bonded joints in relation to riveted joints

Adhesive bonding is also associated to some disadvantages that leave room for more technological research and development. It is necessary to reduce the peel and cleavage stresses because they concentrate the load in a small area giving poor joint strength. They have limited resistance to extreme temperature and humidity conditions due to the polymeric nature of the adhesive. The bonding is usually not instantaneous which requires the use of tools to maintain the substrates in position. In addition, the hardening needs temperature for many adhesives. This is a big economical disadvantage. To have a good interfacial strength and a durable joint, a careful surface preparation is necessary such as solvent cleaning, mechanical abrasion, or chemical treatments. The quality control is more difficult than mechanical fasteners because it is not possible to dismantle an adhesive bond. However, a series of nondestructive techniques are now available. The joint design tends to be complex in many cases. There are no simple rules such as in the case of bolts, rivets, or welding, and design engineers still do not trust this technique, especially when it comes to long-term strength. Applications related to adhesive bonding are today very diverse and can be found in virtually all types of industry. The aeronautical industry is one of the precursors of this technology and increasingly uses adhesives as more composites are introduced in airplanes. The rail and automotive industry is also turning to adhesives to produce lighter vehicles. Civil construction, shoes, and electronics are other examples. Emerging fields such as biology and medicine are also using developing processes based on cell adhesion and protein adhesion on surfaces which are important issues in, for example, biocompatibility of materials for prosthetics, artificial organs, and surgical glues.

4

L. F. M. da Silva et al.

This handbook is focused on adhesion science and technology. The topics of particular interest are surface preparation, joint configuration, adhesive properties, environmental conditions, analytical and finite element analyses of joints, and test methods.

1.2

Definitions

Adhesives have been used for many centuries. However, this method of joining only evolved significantly in the last 70 years. The main reason is because adhesives used in advanced technological applications are based on synthetic polymers whose development only occurred in the middle of the 1940s. The synthetic polymers possess properties that enable them to adhere easily the most materials and a strength capable to transmit considerable loads. According to Kinloch (1987), an adhesive may be defined as a material which when applied to surfaces of materials can join them together and resist separation. However, there are other substances which are outside this definition but which show the phenomenon of adhesion; these include paints and printing inks. A structural adhesive is an adhesive that can resist substantial loads and that is responsible for the strength and stiffness of the structure (Adams et al. 1997). The shear strength can vary from 5 MPa for a polyurethane to 50 MPa for an epoxy. Sealants are also an important part of this handbook and work under the same principles of adhesives. However, in this case the objective is to bond two surfaces, filling the gap between them and forming a barrier or protective layer. The materials to be bonded are called substrates. However, after bonding, the term generally used is adherend. Both adhesives and sealants work by adhesion phenomena. Adhesion is the attraction between two substances resulting from intermolecular forces that establish between them. This concept is different from that of cohesion which only involves intermolecular forces inside one substance. The intermolecular forces that exist in adhesion and cohesion are mainly van der Waals-type forces. Mechanical, electrical, and diffusion phenomena may also occur at the adhesion level (Possart 2005). Figure 2 illustrates the difference between adhesion and cohesion. Materials bonded with adhesives or sealants break by adhesion, cohesion, or a combination of the two. The region between the adhesive and the adherend is referred as the interphase. The interphase has chemical and physical characteristics different from those of the bulk adhesive or adherend. The nature of the interphase is a critical factor in the determination of the mechanical properties of the adhesive bond. The interface, different from the interphase, is a plane of contact between the surface and the two materials. It is within the interphase. It is useful to define and measure the surface energy. The interface is also designated by boundary layer. In the interphase, there might be several interfaces between different materials localized between the adhesive (or sealant) and the adherend. A primer is a substance which is often applied on the substrate to improve adhesion or protect the surfaces until the adhesive or sealant application. The joint is the whole part formed by the adherends, the adhesive or sealant, the primer (if present), and the interphases and interfaces associated to it, as shown in Fig. 3.

1

Introduction to Adhesive Bonding Technology

5

Fig. 2 Examples of cohesive and adhesive failures Fig. 3 Adhesive joint

1.3

Motivation

The adhesion technology is nowadays a common approach to many joining situations. There are many books such as those of Kinloch (1987), Dostal (1990), Adams et al. (1997), Petrie (2000), Packham (2005), Adams (2005), Possart (2005), Lacombe (2006), Tong and Steven (1999), Ebnesajjad (2008), da Silva and Öchsner

6

L. F. M. da Silva et al.

(2008), da Silva et al. (2011, 2012), and da Silva and Sato (2013), and several international journals (International Journal of Adhesion and Adhesives, The Journal of Adhesion, Journal of Adhesion Science and Technology, Applied Adhesion Science, and Reviews of Adhesion and Adhesives) dealing with this subject. However, when the end user needs to apply this technology, he does not have the time or the resources to thoroughly study the various technologies related to adhesives and sealants. It is not practical to go through volumes of text and product information looking for the specific methods or processes to apply. However, such information should be close by if the need arises. The Handbook of Adhesion Technology (second edition) is intended to fill a gap between the necessarily simplified treatment of the student textbook and the full and thorough treatment of the research monograph and review article. The subject is treated very comprehensively and with the most up-todate information so that the end user has in a single book a proper guidance and fundamental knowledge required for many adhesive bonding or sealing applications. This book is intended to be a reference for people needing a quick, but authoritative, description of topics in the field of adhesion and the use of adhesives and sealants. It is intended for scientists and engineers of many different backgrounds who need to have an understanding of various aspects of adhesion technology. These will include those working in research or design, as well as others involved with marketing services. It is expected to be a valuable resource for both undergraduate and research students. The first edition of 2011 is already 6 years old, and the editors felt the necessity to update the whole book and include new material. The fact that the book has been of great interest to the adhesion community was also an important motivation for a complete review.

1.4

Organization of the Handbook

The Handbook of Adhesion Technology, second edition, is intended to be a book of reference in the field of adhesion. Essential information is provided for all those concerned with the adhesion phenomenon. Adhesion is a phenomenon of interest in diverse scientific disciplines and of importance in a wide range of technologies. Therefore, this handbook includes the background science (physics, chemistry, and materials science), engineering aspects of adhesion, and industry-specific applications. It is arranged in a user-friendly format with ten main sections: theory of adhesion, surface treatments, adhesive and sealant materials, testing of adhesive properties, joint design, durability, manufacture, quality control, applications, and emerging areas. The sections contain each about five chapters written by internationally renowned authors who are authorities in their fields. The last chapter gives a brief summary of each chapter with information about the main changes in relation to the first edition. Every chapter has been arranged so that it can be studied independently as well as in conjunction with the others. For those who are interested in in-depth information, numerous sources have been listed at the end of each chapter. The references listed serve as both bibliography and additional reading sources.

1

Introduction to Adhesive Bonding Technology

7

References Adams RD (ed) (2005) Adhesive bonding: science, technology and applications. Woodhead Publishing Limited, Cambridge Adams RD, Comyn J, Wake WC (1997) Structural adhesive joints in engineering, 2nd edn. Chapman & Hall, London da Silva LFM, Öchsner A (eds) (2008) Modeling of adhesively bonded joints. Springer, Heidelberg da Silva LFM, Sato C (eds) (2013) Design of adhesive joints under humid conditions. Springer, Heidelberg da Silva LFM, Pirondi A, Öchsner A (eds) (2011) Hybrid adhesive joints. Springer, Heidelberg da Silva LFM, Dillard D, Blackman B, Adams RD (eds) (2012) Testing adhesive joints – best practices. Wiley, Weinheim Dostal CA (ed) (1990) Engineered materials handbook, vol 3: Adhesives and sealants. American Society for Materials, Metals Park Ebnesajjad S (ed) (2008) Adhesives technology handbook, 2nd edn. William Andrew Inc, New York Kinloch AJ (1987) Adhesion and adhesives: science and technology. Chapman & Hall, London Lacombe R (2006) Adhesion measurement methods – theory and practice. Taylor & Francis, New York Packham DE (2005) Handbook of adhesion, 2nd edn. Wiley, Chichester Petrie EM (2000) Handbook of adhesives and sealants. McGraw-Hill, New York Possart W (2005) Adhesion – current research and application. Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Tong L, Steven GP (1999) Analysis and design of structural bonded joints. Kluwer Academic, Boston

Part I Theory of Adhesion

2

Theories of Fundamental Adhesion David E. Packham

Contents 2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.1 Practical and Fundamental Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.2 Fracture Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.3 Fracture Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.1.4 Structure of This Chapter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2 Adsorption Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.1 Contact and Bond Formation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.2 Strength of the Adhesive Bond . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.3 Environmental Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.4 Kinetic Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.2.5 Rough Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3 Mechanical Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.1 Macro and Microroughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.3.2 Mechanism of Adhesion to Rough Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4 Electrostatic Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.1 Early Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.4.2 Particle Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5 Diffusion Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.1 Origins: Voyutskii and Colleagues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.2 Polymer–Polymer Compatibility . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.3 Interfacial Strength and Interdiffusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.4 Interfaces Between Incompatible Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.5 Copolymer Compatibilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.5.6 Diffusion Theory: Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6 Weak-Boundary-Layer Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.1 Practical Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.2 Polymer Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

12 12 13 14 15 15 15 19 19 20 20 21 21 23 25 25 27 27 27 28 30 32 33 35 36 36 37

D. E. Packham (*) Materials Research Centre, University of Bath, Bath, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_2

11

12

D. E. Packham

2.6.3 Formation of Weak Layers During Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.6.4 The Legacy of Bikerman . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

37 37 38 40

Abstract

The historical development and current status of the four classical theories of adhesion are first reviewed. The adsorption theory emphasizes the point that once adhesive and substrate come into contact forces of attraction will act between them. As long as the extent of wetting is good, these forces, whether primary bonds, such as covalent, or secondary van der Waals forces, are generally considered sufficient to give a high bond strength. Primary bonding may be necessary to achieve bond durability in a hostile environment. The mechanical theory focuses on interlocking between adhesive and a rough substrate surface. Again good wetting is required, or surface roughening is likely to lead to poor bond strength. It has been shown to apply to some surfaces rough on a macroscale as well as to microfibrous and microporous surfaces, such as anodized aluminum. The enhanced adhesion is associated with increasing plastic energy dissipation during fracture in the bulk adhesive. The electrostatic theory points to electrical phenomena such as sparking, which may be observed during the destruction of an adhesive bond, and consider the transfer of electrostatic charge between the adhesive and substrate. It regards the adhesive-substrate system as analogous to a parallel plate condenser. Estimates of the energy associated with this process are generally small compared with adhesion fracture energies, and the theory is much less vigorously supported than was the case some 50 years ago. The diffusion theory has attracted increasing interest since the development of reptation theory of polymer chain dynamics. It provides a model for polymer-topolymer adhesion, and gives an explanation of the time and molecular weight dependence of adhesion to polymers of various compatibilities. The role of weak boundary layers is discussed with emphasis on the importance of careful investigation of the locus of failure of an adhesive bond. In conclusion it is argued that the classical theories are best regarded as emphasizing a different aspect of a more comprehensive model which, in principle, relates molecular dispositions in the region of the interface to macroscopic properties of an adhesive joint.

2.1

Introduction

2.1.1

Practical and Fundamental Adhesion

The phenomenon of adhesion has been of interest for tens of thousands of years. Our remote ancestors were concerned to stick pigments to walls for cave painting and flint and bone to wood for tools and weapons (Pascoe 2005; Wadley et al. 2009).

2

Theories of Fundamental Adhesion

13

Aristotle remarked on the gecko’s ability to adhere to vertical surfaces; both Galileo and Newton were fascinated by the high forces of adhesion that could be observed in certain surfaces in intimate contact. As Newton realized, there are two different types of question to be asked about the phenomenon of adhesion. The first type asks about the forces and so on which have to be applied to separate the bonded components; the second concerns what holds the components together in the first place. These two types of question respectively concern practical and fundamental adhesion. Some authors have written as if they thought that fundamental and practical adhesion were the same. This tendency was more prevalent in the past, but is still occasionally encountered. Practical adhesion, then, is concerned with the magnitude of mechanical force or the quantity of energy which has to be applied to break an adhesive bond. It is relevant whether an adhesively bonded component breaks under load in service, or in some test machine in the laboratory. Practical adhesion is clearly of enormous importance and is considered in detail in this Handbook in ▶ Chaps. 20, “Fracture Tests,” ▶ 21, “Impact Tests,” ▶ 22, “Special Tests,” and ▶ 23, “Constitutive Adhesive and Sealant Models.” This chapter is concerned with fundamental adhesion. It is concerned with forces and mechanisms on, or approaching, the molecular scale, which are involved in holding together the different components of an adhesive bond. Newton recognized that the forces concerned in fundamental adhesion could be very large and challenged “experimental philosophy” to discover what they were! (Newton 1730). Serious scientific concern with fundamental adhesion is usually dated back to the classic work of McBain and Hopkins in the 1920s, and to the various theories were advanced during the middle part of the last century. There was a strong tendency during much of this period to regard different theories as rival explanations of the same phenomena, rather than as complementary aspects of a broader rationalization. Much of the literature of the time, and even some of the literature today, reflects this approach, and therefore it is appropriate, if somewhat anachronistic, in this chapter first to consider different theories individually. A synthesis will be given in the conclusions. It is obvious that adhesion at the fundamental level is a prerequisite for the existence of practical adhesion. What is the relationship between fundamental and practical adhesion? The answer to this question is by no means simple; indeed some authorities in the past have considered that there was no relation between the two. It is not appropriate here to attempt a detailed answer, but a simple, indeed simplified, account will be presented now. Practical adhesion is generally assessed in the laboratory by loading an adhesive joint and observing the force required to produce failure, cf. ▶ Chaps. 20, “Fracture Tests,” ▶ 21, “Impact Tests,” and ▶ 22, “Special Tests.” In some tests, the numerical value representing practical adhesion is a fracture energy (per unit area); in other tests, a fracture stress. Let us consider each in turn.

2.1.2

Fracture Energy

When an adhesive joint fails energy will have to be supplied to break the bonds where the joint fails, creating two new surfaces. These bonds broken may be,

14

D. E. Packham

according to circumstances, primary or secondary bonds. This energy required to form new surfaces is surface energy. If the failure is at the interface between phases 1 and 2, the appropriate surface energy term is the work of adhesion, WA, W A ¼ γ 1 þ γ 2  γ 12

(1)

where γ 1 and γ 2 are the surface energies of materials 1 and 2 respectively in contact with air, and γ 12 is the interfacial energy between phases 1 and 2. An adhesive joint often fails cohesively within one component, rather than at the interface. Under these circumstances, the corresponding term will be the work of cohesion, WC, given (for failure in phase 1) by W C ¼ 2γ 1

(2)

The work of adhesion and work of cohesion have been defined and discussed more fully in ▶ Chap. 5, “Spreading of Liquids on Substrates.” In practice the magnitude of the fracture energy, G, is (almost) always considerably greater than that of the surface energy term WA or WC (now written as G0), given by Eqs. 1 or 2. G0 may be some average of WA or WC, depending on the locus of failure. These Eqs. 1 and 2, are thermodynamic equations; however, fracture of an adhesive bond will rarely, if ever, be reversible as thermodynamics require. Moreover, during the fracture process other energy absorbing processes occur, for example, plastic and viscoelastic deformation. Thus G ¼ G0 þ ψ

(3)

where ψ represents these other energy absorbing processes. From a simple standpoint, Eq. 3 can be taken as representing the relationship between practical adhesion (represented by the measured G) and fundamental adhesion associated with molecular forces at the interface, G0. To repeat, usually ψ is very much larger than G0, and so practical fracture energies for adhesive joints are almost always orders of magnitude greater than work of adhesion or work of cohesion. It might be thought from this that the surface energy term G0 was not important as its magnitude was small. However G0 and ψ are coupled: as G0 increases ψ gets larger as well. Put simply, a very weak interface (low G0) would not be able to transmit a high stress needed to produce high ψ, say by plastic deformation. In some mechanically simple adhesive bonds, the loss term is directly proportional to the surface energy term. If G0 is doubled, ψ also is doubled, so the overall fracture energy, G, becomes twice as large.

2.1.3

Fracture Stress

Where the measure of practical adhesion is a fracture stress, similar considerations apply. A simple treatment of the physics of fracture (Griffith-Irwin theory, v. Good 1972) shows that the fracture stress σ f also depends on the fracture energy G:

2

Theories of Fundamental Adhesion

15 1

σ f ¼ kðEG=lÞ2

(4)

where k is a constant, l is the length of the critical crack which leads to fracture, and E is the effective modulus. When this equation was first put forward, it was considered that the term G was the surface energy, but is was soon recognized that this led to strength predictions which were much too low. It was realized that G represented the sum of all the energy absorbing processes involved in the fracture, i.e., once again Eq. 3 applied. Practical adhesion (now represented by σ f) is related to fundamental adhesion through Eqs. 3 and 4.

2.1.4

Structure of This Chapter

Contemporary discussion of the fundamentals of adhesion generally reflects the historical development of the subject. For convenience, this approach will be adopted here. Early work in the 1920s discussed two kinds of adhesion. Where surfaces were smooth what would now be described as Adsorption Theory applied, with rough or porous surfaces Mechanical Theory. The next two Sects. 2 and 3, of this chapter discuss these in turn. These theories were challenged by work originating in the Soviet Union. The Electrostatic Theory, considered in Sect. 4, laying emphasis on the transfer of charge across an interface, was put forward by Deryagin in the 1940s. A different challenge came from a different Soviet school which insisted on the fundamental importance of diffusion across an interface. The Diffusion Theory is the subject of Sect. 5. The weak-boundary-layer theory is essentially a theory of fracture, rather than of fundamental adhesion. However it is frequently discussed along with these other theories, so consideration is given to it here in Sect. 6. A great deal has been discovered about the mechanisms of adhesion since the middle years of the last century when the “classic” theories of adhesion were first proposed. The concluding section of the chapter will reassess them in the light of present understanding and question whether emphasis on one theory at the expense of another is still a useful way of rationalizing the results of experimental research in adhesion.

2.2

Adsorption Theory

2.2.1

Contact and Bond Formation

It is clear that the adhesive and substrate must come into contact for the possibility of the formation of an adhesive bond to exist. By far, the most common situation is one where a liquid adhesive comes into contact with a solid substrate. (That obviously means close contact on a molecular scale.) Factors that affect the extent of wetting and spreading of a liquid drop on a solid surface are, therefore, of primary

16

D. E. Packham

importance in any consideration of theories of adhesion. It is appropriate, then, that two other chapters in this handbook are devoted to these considerations. As these other chapters show, there is a well-developed thermodynamic theory of wetting and spreading This relates surface energies and surface tensions to contact angles and extent of wetting. Where the relevant surface energies and contact angles are known, or can be deduced, the extent of contact between a specific adhesive and substrate at equilibrium can be predicted. The essential idea of the adsorption theory of adhesion is that whenever there is contact between two materials at a molecular level, there will be adhesion. When two materials come into contact, there will be forces of attraction between them. What those forces will be will depend on the chemical nature of the surfaces of the materials concerned. If this is known, it may be possible to postulate, or even to prove experimentally, the formation of a specific type of bond. Thus, metallic bonds may be involved in contact between a metallic solder and substrate metal (Some authors discuss secondary bonding in adhesion under the heading “Adsorption Theory,” and treat primary bonding under a different heading, e.g., “Chemical Adhesion.”). Strong evidence for the presence of covalent bonds has been found, e.g., from some studies of the action of organosilane adhesion promoters used in bonding epoxies and other adhesives to metal and glass surfaces. It has long been suggested that the effectiveness of silanes was, in part, associated with covalent bond formation between the silane and a metal or glass surface, as indicated schematically in Fig. 1. Secondary ion mass spectroscopy (SIMS), which enables molecular fragments from

a

Reacts with adhesive X

Condenses with other silane molecule

RO

Si

Condenses with other silane molecule

OR

OR Reacts with substrate

b

Adhesive

O

Si O

X

X

X

O

Si

O

O

Si O

Substrate

Fig. 1 Typical structure (a) and postulated mode of action (b) for a silane adhesion promoter

2

Theories of Fundamental Adhesion

CH3O

17

Cδ+ Oδ–

δ–

C H

δ– O

C δ+

O δ–

O

O Acid E.G.

OCH3

O

Weak base

Strong base

Si

Al

Ni

H-bond

ionica

ion/dipole

Fig. 2 Interaction between PMMA and surfaces of differing acidic and basic nature (After Leadley and Watts (1997)). aIndicating hydrolysis of the ester and adsorption through the carboxylate anion

an interface to be examined, has produced evidence of the presence of such siliconoxygen-metal covalent bonds. Using high-resolution X-ray photoelectron spectroscopy (XPS), small changes can sometimes be detected in the molecular bonding of a chemical group at the interface. Leadley and Watts (1997) used the technique to examine the interface between polymethylmethacrylate (PMMA) and oxidized surfaces of metals or silicon. They found evidence for hydrogen bonding, ionic bonding, and ion-dipole interactions, depending on the strength of the acid or basic nature of the mineral surface (Fig. 2). All three are examples of electron donor–acceptor (i.e., Lewis acid–base) interactions. A broadly similar conclusion was reached by a different route by Schneider et al. (2002). Maleic anhydride is used as a curing agent for some types of epoxy resin. The authors studied its interaction with a model system consisting of a thin aluminum layer deposited on a silicon wafer. The aluminum was allowed to oxidize and was then equilibrated with a laboratory atmosphere, before a monolayer of maleic anhydride was deposited from solution in acetone. Reflection–absorption intra-red spectroscopy indicated that the anhydride was hydrolyzed to the acid and adsorbed onto the oxidized aluminum surface. As a result of comparing the observed absorption frequencies with those calculated by quantum mechanical modeling for various plausible structures, the authors argued for the presence of three different adsorption structures (Fig. 3) in the ratio 6:1:1 for Fig. 3a–c. The dominant structure involved a carboxylate ion adsorbed on two neighboring aluminum ions. Figure 3b shows similar carboxylate ion adsorption with additional hydrogen bond adsorption involving the second carboxylic acid group of the maleic acid molecule. The third structure, Fig. 3c, involves interaction between the carbonyl oxygen and aluminum in the substrate surface. All three structures again involve electron donor–acceptor (i.e., Lewis acid–base) interactions. Weak, secondary forces, resulting from molecular dipoles, also act between materials. They are often classified according to the nature of the interacting dipoles. Keesom orientation forces act between permanent dipoles, London dispersion forces

18

D. E. Packham H

a

O

b

H

O

O H-bond

H O

O

H-bond O H O

Al

Al

Al

H

Concentration

6

to

c

H

O

H-bond O H

H

H

O H

H

O

O

O

O

Al

Al

Al

Al

1

to

1

Fig. 3 Adsorption of maleic acid on model-oxidized aluminum surface (After Schneider et al. (2002))

Table 1 Classification of secondary forces Van der Waals

London Debye Keesom

Dispersion Induction Orientation

Dipolar interaction Transient/transient Permanent/induced Permanent/permanent

Table 2 Alternative classification of secondary forces (N.B. some authors, especially in earlier literature, have used t upon specifi to include hydrogen bonding and other electron donorersion forces are universal. They only require Van der Waals Polar forces

London Debye Keesom

Dispersion Induction Orientation

Dipolar interaction Transient/transient Permanent/induced Permanent/permanent

between transient dipoles, and Debye induction forces between a permanent and an induced dipole, see Tables 1 and 2. These are collectively known as van der Waals forces (but note alternative usage of this term, Table 2) and occur widely between materials. They are much less dependent upon specific chemical structure than primary bonds. Indeed, dispersion forces are universal. They only require the presence of a nucleus and of extranuclear electrons, so they act between all atomic and molecular species. Thus, whenever there is contact between two materials at a molecular level, there will be adhesion. Physical adsorption or chemical adsorption (chemisorption) will occur. Specific types of bonding may be present, but there will always, at least, be London dispersion forces. That is the essential idea of the adsorption theory of adhesion.

2

Theories of Fundamental Adhesion

2.2.2

19

Strength of the Adhesive Bond

Elementary chemistry books give bond dissociation energies for bonds of different types. Primary bonds are typically of the order of hundreds of kilojoules per mole, van der Waals bonds tens of kilojoules per mole. It is natural then that many have supposed that an adhesive bond involving primary bonds at the interface will be stronger than one relying simply on van der Waals interactions. However, Eq. 3 above implies that the situation is not so simple. The “strength of an adhesive bond” (practical adhesion) depends not just on the interfacial forces (reflected in G0) but also on other energy terms (ψ), and these latter are usually much greater in magnitude. It is possible to calculate the energy of interaction between model molecules, representing adhesives, and model surfaces. When the results of such calculations are compared with corresponding practically measured adhesion, the predicted strength is found to be much higher than the measured strength. Even when dispersion forces are the only interactions considered, the calculated value is generally an order of magnitude higher than the measured strength. This has led to the conclusion that dispersion forces alone are more than sufficient to account for practical strengths of adhesive bonds. The question to be answered when considering the strength of adhesive bonds is not “why are they so strong?” but “why are they not stronger?” Of course, an analogous question occurs when considering the strength of materials in general, and much of the science of materials is devoted to answering it!

2.2.3

Environmental Stability

The discussion so far points to the critical importance of wetting in the formation of an adhesive bond. Once wetting occurs, there are adhesive–substrate interactions – adsorption of some sort has occurred, and dispersion forces alone will be sufficient to give a strong bond. In a practical context, van der Waals adsorption may not form the basis for a satisfactory adhesive bond. The bond may not be stable in the service environment. This can be seen from a discussion of the work of adhesion considered above. If the work of adhesion (as defined in Eq. 1) is positive, the adhesive will make contact with the adhesive and adsorption will take place. However, in the presence of a medium m, Eq. 1 has to be modified, and the work of adhesion now becomes W A=m ¼ γ 1m þ γ 2m  γ 12

(5)

where γ 1m is the surface energy of material 1 in the presence of the medium m. If = γ 1m and/or γ 2m are sufficiently low, the work of adhesion (Eq. 5) may now become negative, and the bond fails. A widely encountered “medium” is water, either liquid or vapor, and this is strongly adsorbed onto many high-energy surfaces, lowering their surface energy.

20

D. E. Packham

There are many examples of adhesive bonds that fail in a humid environment because the adhesive is displaced by this mechanism from the substrate. Thus dispersion force adsorption may be sufficient, in principle, to provide good fundamental adhesion. In practice, it is often necessary to produce stronger bonds that are less susceptible to attack by medium m, e.g., less susceptible to hydrolysis. Silane adhesion promoters, discussed above, increase the hydrolytic stability of interfaces between adhesives and glass or metals through covalent bond formation.

2.2.4

Kinetic Effects

At the start of Sect. 2, it was pointed out that thermodynamic theories of wetting (cf. ▶ Chap. 6, “Thermodynamics of Adhesion”) enable a prediction to be made of the extent of contact between a liquid adhesive and a given substrate, provided information relating to interfacial energies and contact angles was available. Obviously a thermodynamic theory predicts the equilibrium state, and in practice, it is important to consider whether equilibrium will be attained. Typically, adhesives are liquids that solidify following application to the substrate. There is no reason to suppose that equilibrium wetting will always be achieved before they set. Whether equilibrium will be attained will depend on the relation between such factors as the forces causing the adhesive to spread and the rate of increase in viscosity with time. Moreover, the surface energy of the solidified adhesive will differ from that of the liquid, changing the equilibrium conditions. Such considerations are relevant to the application of the adsorption theory of adhesion.

2.2.5

Rough Surfaces

Rough surfaces are central to the concept of the mechanical theory of adhesion, discussed in Sect. 3. However, practical surfaces are always rough to some degree, and the discussion of the adsorption theory so far has tacitly assumed that the surface to be bonded was smooth. If this roughness is not very great, its effect on wetting will be small, but some of the surfaces used in adhesion technology are extremely rough on a macro or microscale. Significant roughness can seriously reduce the extent of wetting achieved at equilibrium. If wetting is reduced, adsorption, and as a consequence fundamental adhesion, will be reduced also. The extent of this effect depends on the roughness features. It is possible, with plausible assumptions about surface tension and contact angle, to calculate equilibrium contact with model rough surfaces. Figure 4 shows some results after the classic work of de Bruyne (1956). Obviously a reentrant (“ink bottle”) pore is particularly difficult to wet. The calculations on which Fig. 4 is based assume that equilibrium will be established. Considerations of the rate of setting of the adhesive and of its effect on viscosity and contact angle are relevant here. Depending on these factors, contact

2

Theories of Fundamental Adhesion

21

Fig. 4 Equilibrium penetration of a liquid into cylindrical and g will be small, but some de Bruyne (1956)

a

1 μm

b 1 μm 100 μm

100 μm 135°

90°

60 Penetration (μm)

A

30

B 0

0

45⬚ Contact angle

90⬚

between an adhesive and a particular rough surface may be further inhibited by kinetic effects (Packham 2002).

2.3

Mechanical Theory

2.3.1

Macro and Microroughness

Many of the features of the adsorption theory of adhesion, just discussed, can be traced back to McBain and Hopkins’ classical work in the 1920s, where they are referred to it as “specific adhesion” (McBain and Hopkins 1925; Packham 1998, 2002, 2003). They also described “mechanical adhesion” to porous substrates, such as wood, unglazed porcelain, pumice, and charcoal. They regarded it as obvious “that a good joint must result whenever a strong continuous film of partly embedded adhesive is formed in situ.” This is, in essence, the mechanical theory of adhesion. Clearly, this mechanical adhesion relies on contact between adhesive and substrate, and therefore adsorption forces, described in Sect. 2, will act between the two materials. Despite the “obvious” nature of the mechanical theory of adhesion, by the 1950s, the validity of the mechanism was largely denied by scientists working on adhesion. This denial was based on the results of studies in which adhesion to similar surfaces with different degrees of roughness was reported. These generally showed an inverse relationship between roughness and practical adhesion. These results, no doubt, bore witness to the difficulty of obtaining good wetting of a rough surface by a viscous adhesive, cf. Fig. 4. Thus, many joints to rough surfaces

22

D. E. Packham

have voids at the interface and asperities will act as points of stress concentration, which can lower the practical adhesion with a brittle adhesive. However, there are now many examples in the literature where good wetting is obtained with rough surfaces, and, indeed, stress concentrations can even enhance practical adhesion with a ductile adhesive, but initiating local plastic deformation, which increases the energy dissipated during failure – ψ in Eq. 3. Many successful prebonding treatments produce rough, microfibrous, or microporous surfaces. Some of these are shown in Fig. 5. The integrity of a sizable proportion of the world’s civil aviation fleet relies on adhesive bonding to anodized aluminum (Fig. 5a). A black, microfibrous oxide coating on copper is often used as a pretreament for bonding in electronic applications (Fig. 5c). PTFE is notoriously difficult to bond. Figure 5d shows the surface resulting from a successful bonding pretreatment involving irradiation by argon ions. Figure 5e shows an example, on a somewhat coarser scale, where the poor adhesion of copper to silica has been

c

b a The anodic coating Pore

Anodic cell

10 μm

Aluminium

e

TiW (i)

SiO2 Substrate Pd

(ii)

d Electroless Cu (iii)

(iv)

Fig. 5 Some rough surfaces resulting from pretreatment prior to adhesive bonding: (a) porous anodic oxide on aluminum (schematic); (b) dendrites of zinc electrodeposited onto a zinc surface; (c) black CuO layer produced on copper; (d) PTFE irradiated by argon ions (After Koh et al. 1997); (e) adhesion of copper to silica using a mechanical key (van der Putten 1993). (i) TiW islands deposited; (ii) Pd activator adsorbed and HF etching; (iii) electroless Cu deposited; (iv) Cu electrodeposited

2

Theories of Fundamental Adhesion

23

overcome by anchoring the copper to the silica surface using titanium tungstide “keys.” Control experiments have shown scientifically that very rough surfaces, such as those illustrated in Fig. 5, give high adhesion under circumstances where adhesion to a corresponding smooth surface is low. So it could reasonably be said that the mechanical theory of adhesion applies in these cases. Of course, this does not necessarily mean that the high adhesion is the only technological reason for using the pretreatment. For example, the anodization of aluminum enhances durability of an adhesive bond in humid environments: in the aviation context, this is more important than the actual level of practical adhesion achieved.

2.3.2

Mechanism of Adhesion to Rough Surfaces

On the phenomenological level, it is clear that there are many rough surfaces known to adhesion science and technology where the surface roughness plays an essential role in adhesion: the mechanical theory of adhesion applies. Rather than simply ascribing adhesion to “mechanical effects,” it is useful to explore why roughness can lead to good adhesion. Consider again Eq. 3. The surface energy term G0 is, from the standard definition of surface energy, the excess energy in the surface of the material, over that in the bulk material. In a simple way, this excess energy can be thought of as the energy necessary to break those bonds needed to produce the surface. Compare the bulk atom (B) in Fig. 6, with an atom S on a plane surface. In this schematic two-dimensional diagram, where the material is represented as a close-packed array of spherical atoms, B is bonded to six nearest neighbors, but S is only bonded to four. By extension of this argument, it can be seen that an atom (A) on an asperity will have a still higher surface energy. Thus, a rough surface, especially a very rough surface such as those illustrated in Fig. 5b–d, will have a higher surface energy than a corresponding smooth surface. Returning to Eq. 3, it must be remembered that the fracture energy G is fracture energy per unit area. Obviously, G0 and ψ are also energies per unit area. The area relevant here is the formal (ideal) area, i.e., the macroscopic area of the interface, assuming that there was no roughness. For a rough surface, the “true” area will be greater and the surface energy term (G0) will consequently be higher. For moderate increases in surface roughness, a proportional increase in adhesion (practical adhesion) has been demonstrated (Gent and Lai 1995). For very rough microporous and microfibrous surfaces, the effective increase in area becomes enormous: G0 can be raised to a very high value indeed. Many engineering surfaces are fractal in nature: for such surfaces, the “area” is, in principle, indefinitely large. The practical adhesion to such surfaces does not become indefinitely large, because a joint with a strong interfacial region will fail cohesively in some other region where G0 is locally smaller. The results presented in Table 3 show this effect in practice. Much higher fracture energies are found for adhesive bonds made to microfibrous surfaces than to

24

D. E. Packham

Fig. 6 Local environment of an atom in the bulk of a material (B), on a plane surface (S), and on an asperity on a rough surface (A) (Schematic representation)

S

B

A

Table 3 Adhesion of low density polyethylene (LDPE) to copper and of epoxy resin to electroformed zinc, assessed, respectively, by peel and single edge notched (SEN) tests (Evans and Packham 1979; Hine et al. 1984) Polymer LDPE LDPE Epoxy (unmodified) Epoxy (unmodified) Epoxy (toughenedd) Epoxy (toughenedd) a

Fracture energya J/m2 210  60

Fracture surface Close to substrate

1,620  140

Plastic deformation of PE

Substrate Copper (polished) Copper (microfibrousb) Zinc (flat)

Test 180 peel 180 peel SEN

Zinc (dendriticc)

SEN

670  290

Zinc (flat)

SEN

700  200

Zinc (dendriticc)

SEN

2,548  515

105  30

95% confidence limits given cf. Fig. 5c c cf. Fig. 5b d 15% CTBN rubber toughening agent incorporated b

Apparently adhesive Cohesive above dendrite tips Adhesive with residual polymer areas Cohesive

2

Theories of Fundamental Adhesion

25

Fig. 7 Adhesion of polyethylene to copper. Substrate surface after peeling from (a) polished copper and (b) copper with a microfibrous oxide surface

corresponding smooth surfaces. For the smooth surfaces, stresses are concentrated at the interface, and failure occurs at or close to the interface with little plastic energy dissipation (Fig. 7a); for the microfibrous surfaces, the stresses are concentrated at the fiber or dendrite tips causing yielding, which moves into the polymer, giving cohesive failure and higher fracture energy associated with the plastic deformation involved (Fig. 7b). It is interesting to note that, even with the brittle unmodified epoxy resin, cohesive failure occurs showing signs of plastic deformation associated with the dendrite tips (Fig. 8). Thus, the basic ideas, related by McBain and Hopkins in the 1920s, underlying both the adsorption and mechanical theories of adhesion still provide useful models for rationalizing observations in adhesion science. Two further adhesion theories, the electrostatic and diffusion theories, emerged from the Soviet Union in the middle of the twentieth century. We now turn to consider them, and how they are regarded today.

2.4

Electrostatic Theory

2.4.1

Early Formulation

The electrostatic theory was put forward in the 1940s by Deryagin (alternative transliteration Derjaguin) and colleagues in the Soviet Union (Deryagin and Krotova 1948). The interface is seen as analogous to the plates of an electrical condenser across which charge transfer occurs. A physical model was introduced representing the electric double layer as a three-dimensional region at the polymer–solid contact (Deryagin et al. 1973).

26

D. E. Packham

Fig. 8 Zinc with a dendritic surface bonded with unmodified epoxy resin: fracture surface showing plastic deformation of the epoxy just above the dendrite tips (Hine et al. 1984)

The basis of the theory is that free charges exist to some extent in any condensed material, even in the best dielectrics, and there will always be an electrochemical potential difference across the interface between two materials in contact, e.g., adhesive and substrate. Free electronic or ionic charge carriers will tend to move across the contact interface, and an electric double layer is established. This mechanism is considered quite distinct from any charge transfer, which may be associated with bonding at the interface, as discussed in Sect. 1.1. Electrical phenomena (e.g., sparking) can often be demonstrated to accompany the destruction of an adhesive bond. The work of separation of the plates of a condenser depends upon rate of separation and ambient gas pressure. Measured peel energy certainly depends on rate of separation, and some published results indicate a dependence upon gas pressure. Deryagin argued that this supported the electrostatic theory of adhesion. He published results showing good correlation between measured peel energy and condenser discharge energy, but the details of his argument have been the subject of adverse criticism, see, e.g., Wake (1982) and Kinloch (1987). More recently, Possart has used potential contrast scanning electron microscopy to study electric double layer formation associated with the interface between a thin layer of solvent-cast low-density polyethylene and aluminum (Possart 1988). While arguing that the electrical discharge energy of the double layer makes a contribution to the fracture energy, this study accepts that many other energy mechanisms also contribute. The work spent in mechanical testing surmounts the electrostatic interaction energy in the double layer by several orders of magnitude. The work of peeling was measured to be 1 J/m2. The specific electrostatic work of separation was calculated as 247 J/m2. There was a period in the mid-years of the last century when its supporters presented the electrostatic theory as a comprehensive theory of adhesion, replacing the adsorption theory. Many critics of the theory pointed out examples where a change of filler or of polymer might lead to the absence of electrical phenomena,

2

Theories of Fundamental Adhesion

27

despite there being similar levels of measured adhesion: most cases of conventional adhesion, it was argued, could be explained without recourse to the electrostatic theory. By the 1970s, Deryagin himself was prepared to accept that “an adhesive bond is always caused by either the forces of chemical bonding or by so-called van der Waals forces.” (Deryagin et al. 1978). The present writer’s perception is that the widely held view is that electrostatic contributions in conventional adhesion are likely to be small; this view, however, is not universally accepted.

2.4.2

Particle Adhesion

Although the detailed mechanism is distinct from that just discussed, it is relevant to note that electrostatic forces are still considered to be of relevance in adhesion of small particles. An advantage of this sort of adhesion is that it may be possible to manipulate the charged particles by external electric fields. Electrostatic precipitation is a well-established industrial technique. It is used to remove dust particles from a gas flow, e.g., from flue gases in a power station. Electrically charged dust particles are attracted to an electrode surface and are thus removed from the gas stream. In a similar way, the application of a strong field across the open base of a silo can be used to prevent the egress of charged dust particles. The modern photocopier relies for its function on both electrostatic and van der Waals forces. Charged toner particles initially adhere to charged regions on a moving photoconductor belt and are subsequently transferred by electrostatic forces to a paper sheet on which they are fused (Kendall 2001).

2.5

Diffusion Theory

2.5.1

Origins: Voyutskii and Colleagues

The second adhesion theory to emerge from the Soviet Union in the mid-years of the twentieth century was the diffusion theory. The basic principles were set out by S.S. Voyutskii and his colleagues. Their compatriot, R.M. Vasenin, contributed to its quantitative development using Fickian diffusion theory (Voyutskii 1963; Wake 1982). Much of Voyutskii’s original work was done on the self-adhesion (called autohesion) of unvulcanized rubbers. It was subsequently extended to polymer adhesion, more generally. The theory postulates that the molecules of the two parts of the specimen interdiffuse, so that the interface becomes diffuse and eventually disappears. For polymers in contact, Voyutskii studied the effects on adhesion of such variables as time, temperature, contact pressure, molecular weight, polarity, and crosslinking. He argued that the results proved that the adhesion was associated with the interdiffusion of polymer chains. Voyutskii’s 1963 book, Autohesion and Adhesion of High Polymers, is the locus classicus for the exposition of his ideas in English. Here he takes a robust,

28

D. E. Packham

comprehensive view of the scope of the diffusion theory arguing that it is superior both to adsorption theory and electrostatic theory. He even argues for its application to the adhesion of polymers to metals. As already indicated, the mid-decades of the twentieth century were the time when many authorities in the field of adhesion took an exclusive approach to theories of adhesion: if one theory was correct, another must be incorrect. In contrast, it is now much more widely recognized that different theories may have different areas of applicability, or even be better regarded as emphasizing a particular feature of a broader, overarching understanding. Consistent with the spirit of the 1950s and 1960s, the diffusion theory generally got an unsympathetic reaction among supporters of the adsorption theory, i.e., mainly workers in the West. But attitudes changed, and by 1970, the position could probably be summarized by saying that diffusion was generally accepted as a mechanism for adhesion between samples of the same polymer (autohesion) or between very similar polymers (Wake 1982), but it was not more broadly accepted, even for polymer–polymer adhesion. The reluctance to accept the applicability of diffusion theory to adhesion between chemically different polymers was because mutual solubility (emphasized by Voyutskii) was regarded as an essential feature of the mechanism, and most polymer pairs were known to be incompatible.

2.5.2

Polymer–Polymer Compatibility

It is often important to know whether two phases are mutually soluble, i.e., will mix on the molecular level to form a true solution. The question of mutual solubility – compatibility – can be approached in terms of thermodynamic criteria. The two phases will be compatible if the Gibbs free energy for mixing is negative. The Gibbs free energy of mixing ΔGm is related to the enthalpy (heat) of mixing ΔHm and entropy of mixing ΔSm by the usual second Law equation: ΔGm ¼ ΔH m  TΔSm

(6)

where T is the absolute temperature. As mixing always increases disorder, ΔSm is positive, it follows that the – TΔSm term is always negative, and therefore favors mixing. However, where polymers are involved, the long chain molecules mean that the entropy gain is much smaller than when compounds with small molecules are mixed. The – TΔSm term, although negative, tends to be small in magnitude. One approach to the discussion of compatibility in polymer systems is the FloryHuggins theory (Berg 2002). This considers the number of ways in which chain segments of the polymers may be distributed among identical unit cells in a hypothetical lattice. The theory gives an expression for the free energy of mixing of polymers 1 and 2 (volume fractions ϕ1 and ϕ2, degrees of polymerization x1 and x2) in terms of the Flory-Huggins interaction parameter, χ. It may be written as:

2

Theories of Fundamental Adhesion

29

ΔGmix ¼ kT ½χϕ1 ϕ2 þ ðϕ1 =x1 Þlnϕ1 þ ðϕ2 =x2 Þlnϕ2 

(7)

ΔGmix =kT ¼ χϕ1 ϕ2 þ ðϕ1 =x1 Þlnϕ1 þ ðϕ2 =x2 Þlnϕ2 enthalpy configurational entropy

(8)

where k is Boltzmann’s constant and T absolute temperature. In this particular form of the equation, the free energy change is expressed per total number of polymer chain segments. The interaction parameter, χ, is a dimensionless temperature-dependent parameter. The last two terms in Eqs. 7 and 8, involving ln ϕ, represent the configurational entropy of mixing. This is negative and so favors the formation of a solution, but, as explained, it is small in magnitude because of the limited potential gain in entropy in mixing long chain molecules. The first term, involving χ, was originally conceived as expressing the enthalpy of mixing and is usually positive, meaning that for most polymer–polymer combinations ΔGmix overall is positive, and so the mixture is heterogeneous. Even polymers as similar as polyethylene and polypropylene are incompatible: the value of χ at 140  C has been calculated to be 0.011. Such thermodynamic treatments point to the incompatibility of most polymer/ polymer combinations, and practical studies with polymer blends lead to the same conclusion. Mutual solubility, regarded as an essential feature of the diffusion theory, would rarely be achieved, restricting the applicability of the theory to autohesion, and perhaps to adhesion between polymers of closely similar chemical structure.

Developments in Molecular Dynamics Following the work on molecular dynamics by de Gennes (1998) and by Doi and Edwards (1986) leading to the development of reptation theory, a much more detailed description of the motion of polymer chains can now be given (Wool 1995, 2002, 2005). Consider two samples of an amorphous polymer brought into contact above the glass transition temperatures. The conformations of the chains at the interface will tend to relax. Five different mechanisms of relaxation can be identified, corresponding to five different time scales. In order of increasing relaxation time, they are: (1) short-range Fickian diffusion of individual chain segments, (2) Rouse relaxation between chain entanglements, (3) Rouse relaxation of the whole chain, (4) reptation diffusion, and (5) Fickian long-range diffusion. It is the reptation region (4) of diffusion of chain segments to a depth of the order of the radius of gyration that is considered to be the most important in discussing the development of strength of interfaces. This is illustrated for the autohesion of polystyrene at 118  C by the data of Table 4 (Wool 2005). It can be seen that reptation leads to considerable interpenetration of the chains and, as discussed below, to significant strength. The reptation model envisages the polymer chain to be enclosed within an initial tube out of which it gradually escapes by wriggling in a snake-like manner (“reptating”: reptare to creep). The reptation relaxation time Tr corresponds to the time when about 70% of the chain has escaped from the initial tube. During this time,

30

D. E. Packham

Table 4 Chain relaxation and diffusion mechanisms: relaxation time (t) and molecular weight (M ) relationships. Mc is molecular weight between chain entanglements, te is the Rouse relaxation time between chain entanglements, tRO is the Rouse relaxation time of the whole chain, Tr is the reptation relaxation time

Mechanism Rouse relaxation between chain entanglements Rouse relaxation of the whole chain Reptation

Relation between relaxation time and molecular weight τe  Mc2

Relaxation timea 10 s

Average diffusion distancea at τ 30 Å

τRO  M2

21 min

60 Å

Tr  M3

1,860 min

110 Å

Numerical data refer specifically to polystyrene (M = 245,000) welded to itself at 118  C a

the interfacial thickness increases with time t, in proportion to t¼, in contrast with the t½ relationship for Fickian diffusion. The relaxation time Tr is proportional to the cube of molecular weight.

2.5.3

Interfacial Strength and Interdiffusion

The fracture strength of a polymer depends on molecular weight, the chain length, and in a similar way, the fracture of a polymer–polymer adhesive bond depends on the length of the chain, which has diffused across the interface. Figure 9 shows the molecular weight dependence of the fracture energy for polystyrene. It can serve as a focus for consideration of the mechanisms involved in polymer–polymer adhesion. Two characteristic molecular weights are involved in this discussion: Mc and M*. Mc is the average molecular weight between chain entanglements, the critical entanglement molecular weight. It represents the onset of the well-known zero shear viscosity law η  M3:4

(9)

and the molecular weight at which the dynamics change from Rouse to reptation. In contrast, M* is the value above which the molecules are too long for chain disentanglement (or pullout) to occur. They are related approximately by: M  8Mc

(10)

Low molecular weight When M < Mc (region A, Fig. 9) only short lengths of polymer chain are involved in the fracture process, which occurs by chain pullout. The fracture energy is low, up to about 1 J/m2. Fracture behavior in this region can be treated using the “nail solution” where the weak interface is modeled as if the two sides were nailed together with Σ nails per unit area, each of length L. The analysis shows that the fracture energy associated

Theories of Fundamental Adhesion

Fig. 9 Fracture energy versus molecular weight for polystyrene in the virgin state (After Wool 2005). Mc is the average molecular weight between chain entanglements; M* is the average molecular weight above which the molecules are too long for chain disentanglement (or pull-out) to occur

31

10,000 M∗ 1,000 C

GIc, J/m2

2

100 B 10

1

A Mc

–1 1,000

100,000 10,000 Molecular weight

1,000,000

with chain pullout, G, increases with L2 and with Σ. To this “friction” term must be added a surface energy term, so that G1c  2Sγ þ kL2 Σ

(11)

where S depends on surface roughness (S = 1 for a perfectly flat fracture surface) and k is a constant related to the “friction” involved in chain pullout. The term “friction,” of course, refers to inter- and intramolecular attractions, such as van der Waals forces. Intermediate molecular weight When Mc < M < M* (region B, Fig. 9), the chains are long enough for entanglement to occur and the fracture energy rises progressively. Fracture in this region occurs by disentanglement and chain pullout: bond rupture is not considered to be significant. Fracture in this region can be treated by the vector percolation theory, which models the polymer as a three-dimensional network, the links of which are progressively broken (here by chain disentanglement) until a critical value, the percolation threshold, is reached, and the whole net fractures (Wool 2005). High-molecular weight Here, M > M* (region C, Fig. 9) and the dominant facture mechanism becomes bond rupture. Vector percolation theory predicts extensive bond rupture throughout a significant volume of the polymer before failure occurs. With glassy polymers, this is often associated with crazing. The fracture energy is high tending to a limit, according to the Flory equation for molecular weight dependence: G1c =G1c  ¼ ½1  Mc =M where G1c* represents the fracture energy at high molecular weight.

(12)

32

D. E. Packham

To summarize, for autohesion (polymer–polymer welding), there will be rapid (Rouse) interdiffusion, which occurs to distances comparable to the radius of gyration of the entanglement molecular weight, Mc, say 3 nm (Table 4). The interface is very weak and could be described in terms of the nail solution, with chain pullout being the dominant fracture mechanism. As welding proceeds, chains diffuse by reptation diffusion, and the failure mechanism then involves chain disentanglement and, with more extensive diffusion at longer times, bond rupture dominates. When bond rupture begins to dominate, the weld will appear fully healed, even in the absence of full interpenetration. There is a rate effect here. At high test rates, chains may not have time to disentangle and may rupture instead. This will give a high fracture toughness, G1c, but at much lower deformation rates, weld weakness may be apparent. Wool (2002) cites, as an example, welding of a particular polystyrene at 125  C, where essentially full (short time) weld strength would be obtained in 156 min, but the fatigue strength would be about one-fifth of its virgin value. For complete welding, a reptation time of 435 min, the relaxation time for reptation, would be needed. The importance of this observation for the molding of thermoplastics is obvious.

2.5.4

Interfaces Between Incompatible Polymers

Despite macroscopic thermodynamic incompatibility, Sect. 5.2, an atomically sharp interface between two such polymers will not be stable. Although there is an enthalpy debt to be paid if a chain of polymer 1 starts to diffuse into polymer 2, there is an entropy gain. Helfand and Tagami 1972, Helfand 1992) introduced a model, which considered the probability that a chain of polymer 1 has diffused a given distance into polymer 2 when the interactions are characterized by the parameter χ. They predicted that at equilibrium, the “thickness,” d1, of the interface would depend upon the interaction parameter and the mean statistical segment length, b, as follows 1

d 1 ¼ 2b=ð6χ Þ2

(13)

Thus, despite the argument in Sect. 5.2, some interdiffusion can be expected between formally incompatible polymers. The extent will depend on the extent of incompatibility, e.g., on the value of the interaction parameter: the smaller the χ, the greater the possibility of interdiffusion. Different mechanisms were discussed above (Sect. 5.3) in the context of autohesion. Depending on the extent of interdiffusion, fracture may occur at low energy by chain pullout or by chain disentanglement with bond rupture accompanied by crazing giving the highest fracture energies corresponding to the greatest extent of interdiffusion. In principle, the same considerations should apply to adhesion between different polymers. The toughness of an interface will be expected to be related to the depth of interpenetration of the chains. Wool (1995) has argued that the fracture energy, G, for

2

Theories of Fundamental Adhesion

33

Fracture energy GIc (J/m2)

35 30 25 20 15 10 5 0 0

0.5

1

1.5

2

2.5

3

Number of entanglements (Nent)

Fig. 10 Fracture energy G1c of bonds between a range of immiscible polymers as a function of number of entanglements (After Wool 2005)

chain pullout, is proportional to the square of the interface thickness, which, via Eq. 6, gives G / d 1 2 / 1=χ

(14)

cf. Eq. 13 above. For higher fracture energies, a linear relationship with the number of chain entanglements, Nent, has been found for a range of polymer pairs, Fig. 10.

2.5.5

Copolymer Compatibilizers

Much higher adhesion may be obtained between incompatible polymers through the use of copolymer compatibilizers. If an AB diblock copolymer is introduced at the interface between two incompatible polymers A and B, the interface may be toughened by diffusion of the copolymer ends into the respective homopolymer. Much ingenuity has been employed in developing this type of compatibilizer. A range of fracture toughness values (Gc) may be obtained. The general principles are similar to those discussed in Sect. 5.3. For diblock copolymers, both surface density (Σ) and degree of polymerization (N ) of the blocks are important. For example, if the degree of polymerization of the A block were shorter than the entanglement length, but that of the B block longer, interdiffusion could occur at the interface, but the bond would be weak with chain pullout of the A part of the copolymer from polymer A. Typical results are illustrated schematically in Fig. 11 (cf. Creton et al. 1994a): strength of the bond increases both with surface density of copolymer and with segment chain length (here segment A). However, as all the lengths of segment A are below the entanglement length, the fracture energy corresponds to chain pullout and is low.

34

D. E. Packham Gc

Fig. 11 Adhesion between incompatible polymers A and B. Interface reinforced by diblock copolymer AB, with A-block length below entanglement length

Increasing

~1J/m2

A-block length

Surface density of copolymer (S )

Gc

~100 J/m2

Increasing plastic deformation

Weak layer at interface

Surface density of copolymer (S )

Scission at diblock junction

Fig. 12 Adhesion between incompatible polymers A and B. Interface reinforced by long chain length diblock copolymer

With greater degrees of polymerization, chain entanglement can take place, Fig. 12. At low surface density (Σ), chain scission will occur close to the junction of the diblock, with each fragment remaining on the “correct” side of the interface. The fracture energy is low. As the copolymer surface density increases, fracture energy rises steeply with plastic deformation, e.g., crazing, occurring in the polymer followed by chain scission or pullout. If an increase in copolymer surface density continues, eventually the surface becomes saturated and a weak layer forms at the interface with fracture energy falling toward a limiting value (cf. Creton et al. 1994b).

2

Theories of Fundamental Adhesion

35

Fig. 13 Schematic representation of a random copolymer at the interface between two incompatible homopolymers. Incompatibility increases in the order (a), (b), and (c)

Toughening of a polymer–polymer interface with random copolymers can sometimes be more effective than with diblocks, provided the polymers are not too incompatible. This is of industrial, as well as of scientific, interest as random copolymers are usually cheaper to produce. Each molecule in diblock copolymers will form a single, strong chemical linkage across the interface. However if the incompatibility between the homopolymers is not too large, a molecule of a random copolymer will form coils wandering many times across interface, forming many “stiches.” At larger incompatibilities, the random copolymer will no longer describe coils, but the homopolymers will still be linked by “loops.” If the incompatibility is too large, the copolymer will simply form collapsed globules at the interface, a weak boundary layer giving no enhancement of adhesion, cf. Fig. 13. As with block copolymers, the important parameters are the surface density and length of the copolymer chains. Toughening of the interface may occur as a result of pullout or scission of the connector chains, or of fibril or craze formation in the matrix. This last mechanism gives the highest fracture toughness, G, and tends to occur at high surface density of chains. Needless to say, implicit in the general discussion of Sects. 5.4 and 5.5 is that the time and temperature of contact are such that the degree of interdiffusion discussed can occur. The lowering of temperature, and the shortening times will lead to kinetic inhibition.

2.5.6

Diffusion Theory: Conclusion

The ideas advanced by Voyutskii and his colleagues over half a century ago have now been developed far beyond what was possible at the time. It is now possible to relate with some clarity the circumstances where interdiffusion plays a significant part in adhesion, and plausible molecular models have been developed, which go a long way to describe it, relating joint strength to interfacial structure. These scientific developments have important technological implications for the development of new multicomponent polymer materials and for the recycling of mixed plastics waste.

36

D. E. Packham

2.6

Weak-Boundary-Layer Theory

2.6.1

Practical Adhesion

This article is concerned with theories of fundamental adhesion, i.e., with theories concerned with the reasons for two materials holding together in an adhesive bond, however weak. In contrast, a weak boundary is a cohesively weak layer in the interfacial region (some authors use the term “interphase”) of an adhesive joint, which may cause the joint itself to be weak, i.e., to fail at a low stress or with low fracture energy. So the question of weak boundary layers occurs when the level of practical adhesion is under consideration, and may form part of an answer to a question such as “why did this joint fail at such and such a stress?” However, 50 years ago, when there was considerably less basic understanding, and more confusion, in the study of adhesion than there is now, “weak boundary layer theory” would often be discussed alongside theories such as the adsorption theory and the electrostatic theory. The tendency to group theories of adhesion in this way has persisted, and so it may now be useful to discuss the significance of weak boundary layers.

Need for Pretreatment At one level, the existence of weak boundary layers and their deleterious effect on adhesion is obvious. Surface pretreatments, prior to adhesive bonding, are almost always a prerequisite for success. That is, because almost all surfaces will be contaminated to some degree from the environment to which they have been exposed. Generally speaking, extensive cleaning under conditions of ultrahigh vacuum is required to remove such contamination and to maintain a molecularly “clean” surface. Good adhesion is aided by high-substrate surface energy, partly to aid wetting (see ▶ Chap. 4, “Wetting of Solids,” ▶ 5, “Spreading of Liquids on Substrates,” and ▶ 6, “Thermodynamics of Adhesion”) and partly to increase fracture energy G, see Eqs. 1 and 3. The universal tendency for high-energy states to move toward lower-energy states is manifested here by adsorption onto a highenergy substrate surface of extraneous material. This can come from other solids or liquids with which the surface has been in contact, as well as from the atmosphere. Thus the surface of a metal, which, outside the realms of surface science might be described as “clean,” will be covered by complex layer including an oxide, probably partly hydrated, perhaps carbonated, adsorbed water, residual lubricant, and processing aids. Such a layer, in molecular terms, will be thick. It is unlikely to have good wetting properties and is likely to be cohesively weak. Unless removed by pretreatment, it is likely to form a weak boundary layer and to give poor practical adhesion. A successful pretreatment is very unlikely to produce a molecularly “clean” surface, but the surface produced will not suffer from the same cohesive weakness and lyophobicity of the original. Polymers are also likely to require pretreatment, because their surfaces usually carry complex layers, chemically different from the bulk material. The formation of these layers is now considered.

2

Theories of Fundamental Adhesion

2.6.2

37

Polymer Surfaces

Few, if any, polymers in commercial practice are used without the incorporation of a range of additives. These may include antioxidants, processing aids, perhaps plasticizers and lubricants. Elastomers are likely to be based on even more extensive formulations including a complex vulcanization mixture. In Sect. 5.2, the question of polymer–polymer compatibility was considered, and it was argued that the majority of binary polymer blends will be two phased because the component polymers are incompatible, the free energy of mixing, Δ G m, will be positive. Essentially, the same argument applies when polymer additive solutions are considered. The physical chemistry is similar: the entropy term will be small compared with that for a low-molecular-weight solvent–solute combination, and the enthalpy term is likely to be positive (cf. Eqs. 6–8). Most polymer-additive systems are thermodynamically unstable: the additive will tend to migrate out of the polymer. Loss of plasticizer from PVC (polyvinyl chloride) is a widely experienced phenomenon. The mode of action of antistatic agents depends on their progressive migration to the polymer surface. Most adhesives are polymeric and set during the bonding process by loss of solvent, by solidifying from the melt, or by some form of cross-linking or polymerization reaction. In all of these solidification processes, there is the potential for lowmolecular-weight additives to migrate to the substrate–adhesive interface and to form a weak boundary layer. Moreover, almost all polymers have a distribution of molecular weights, and there will be a similar tendency for low-molecular-weight fractions of the polymer itself to be rejected to the interface during solidification, especially during cooling from the melt.

2.6.3

Formation of Weak Layers During Bonding

The need for pretreatment to remove weak surface layers is widely understood. More subtly, other weak layers can form during the bonding process by the migration of low-molecular-weight species to the interface: sometimes these can lead to poor joint performance. Many examples in the literature show that these effects occur. Classic experiments include removing low-molecular-weight fractions from a polymer, and observing increased adhesion, and complementarily, adding low-molecularweight fractions, and observing a reduction in adhesion (Bikerman 1961). In Sect. 5.5, examples were given of weak boundary layers formed by copolymer compatibilizers.

2.6.4

The Legacy of Bikerman

Much of the early work on weak boundary layers was done by Bikerman and his colleagues in the 1950s and 1960s. His frequent observation of cohesive failure close

38

D. E. Packham

to the interface, combined with direct demonstration of a weak boundary layer mechanism in specific examples, led him to take up an extreme position (Bikerman 1961). This insisted on (1) the widespread occurrence of weak boundary layers, (2) the impossibility of adhesive (i.e., interfacial) failure, and (3) the practical strength of an adhesive bond depended on the rheology of the joint, not upon the forces at the interface. Although Bikerman’s position would not be supported today, it stimulated developments in the understanding of theories of adhesion, which can, to some extent, be regarded as his legacy. The potential for the formation of weak boundary layers (1) is well established, and much of the technique of adhesive bond preparation is designed to avoid them. Locus of failure Where weak boundary layers exert an effect on adhesion behavior, failure typically occurs within the weak layer, close to the adhesive–substrate interface. The boundary layer will often be very thin, in macroscopic terms, and there are large numbers of examples where workers have reported the adhesive (i.e., interfacial) failure of a joint, when in fact it could be shown that failure was cohesive within a boundary layer. Bikerman’s insistence (2) that adhesive failure could not occur had the beneficial effect of concentrating attention on the locus of failure. Careful examination of locus of failure is an important technique – still sometimes neglected – in the study of adhesive bonds. Modern methods of surface analysis are appropriate here. They have established that, although adhesive failure may be rare, it does occur sometimes. In considering locus of failure, it must be emphasized that observation of cohesive failure close to the adhesive–substrate interface does not, itself, prove the presence of a weak boundary layer. The idea that an adhesive joint “fails at its weakest link” is naïve. The locus of failure of a joint depends in a complex way on the chemical and physical properties of the materials in the joint, and on the way the stress is applied to them. Cohesive failure close to an interface is entirely possible in the complete absence of weak interfacial layers (Good 1972). Dillard has demonstrated how failure can change between cohesive, interfacial, and oscillating for exactly the same joint, depending on how the stress is applied (Chen and Dillard 2002). Practical bond strength Bikerman was right (3) in insisting on the importance of joint rheology in determining the practical strength of an adhesive bond. It was an important point to make at the time. However, he was wrong to say that interfacial forces were irrelevant. The connection between the two, expressed in a simple form in Eq. 3, is discussed in detail elsewhere in this handbook (▶ Chap. 3, “Forces Involved in Adhesion”).

2.7

Conclusions

A theory in science is a model that aims to order the experimental results in a field and, for a fruitful theory, to suggest new experiments and their outcome. We do not look to theories for immutable truths: we expect them in time to be modified and discarded (Popper 1959; Khun 1970).

2

Theories of Fundamental Adhesion

39

In the light of our present understanding of adhesion, what can be said about the “classical” adhesion theories? On one level, it could be said that the adsorption, mechanical, and diffusion theories (and some would add electrostatic theory) maintain their importance undiminished. It is still acknowledged that adsorption plays a vital role, that there are examples where diffusion is crucial and so forth. However, the classic theories originated at a time when there were few, if any, experimental techniques available to study surfaces and interfaces at the nanometer level. We now benefit from several decades of results from scanning probe technologies and surface analysis and theoretical advances in contact mechanics and molecular chain dynamics. We have direct evidence for the presence of particular groups on a surface, for the effects of variations of topography and chemical nature over small distances on a surface. Consequently, in many cases, we have a much clearer understanding of what is happening in the interfacial region, and of how it affects adhesion. One consequence of this understanding is that we can often make plausible suggestions about adhesion mechanisms without having to hide behind broadbrush statements about “mechanical effects” or “chemical activation.” We should also recognize now that the mechanisms characteristic of classical theories are not completely distinct. This is our contemporary answer, no doubt provisional, to Newton’s challenge (Sect. 1.1). A good example of such an approach to theories of adhesion is provided by a recent extensive review of fracture and adhesion of soft materials by Creton and Ciccotti (2016). Nowhere in their review of adhesion mechanisms do Creton and Ciccotti find it necessary to refer to “classical” theories of adhesion, adsorption, mechanical, and so on. However, they discuss the use and range of validity of equations for fracture energy applicable to some of the soft materials which is their concern. These include equations which are much more specific versions of the general Eqs. 3 and 4 above. It is recognized that the practical surfaces used in adhesion are always to a degree rough: the only question is “how rough?” As the roughness of a surface changes, so must the chemical environment of its surface atoms and molecules. Thus, changing the roughness changes the local chemistry, which will affect the adsorption properties of the surface. The scale of roughness of surfaces encountered in adhesive joints varies from the macroscopic through to the nano- and molecular scale, from the classic realm of the mechanical theory to that of the diffusion theory. The energy expended in peeling strands of rubbery polymer from a macroporous surface has been successfully analyzed in essentially the same way as the pullout of individual polymer chains by the “nail” model in diffusion theory (Sect. 5.3) (Gent and Lin 1990). In principle, the range of types of forces that may act between adsorbate and adsorbed molecule on a nominally flat surface are the same as those that may act between diffusing molecules. Moreover, the way in which adhesion is enhanced is essentially the same whether we think in terms of adsorption, diffusion, or roughening. These provide mechanisms whereby the energy dissipation (ψ in Eq. 3) is enhanced. Whether the focus is on enhanced chemical interaction, surface roughening, or interdigitation of polymer chains, a common consequence is increased plastic or viscoelastic losses and crazing.

40

D. E. Packham

Each of the classical theories is best regarded as emphasizing a different aspect of a more comprehensive model, which, in principle, relates molecular dispositions in the region of the interface to macroscopic properties of an adhesive joint. It would be a mistake today to lay too much emphasis on the distinction between classical theories, although this has been valuable at various times during the last 80 years in stimulating the development of new concepts and in suggesting fruitful experiments.

References Berg JC (2002) Semi-empirical strategies for predicting adhesion. In: Chaudhury MK, Pocius AV (assoc ed) Adhesion science and engineering, vol II, Surfaces, chemistry and applications. Elsevier, Amsterdam, pp 1–73 Bikerman JJ (1961) The science of adhesive joints. Academic, London de Bruyne NA (1956) Aero Res Tech Notes Bull 168 Chen B, Dillard DA (2002) Crack path selection in adhesively bonded joints. In: Dillard DA, Pocius AV (eds) Adhesion science and engineering, vol I. Elsevier, Amsterdam, pp 389–442 Creton C, Ciccotti M (2016) Rep Prog Phys:79. https://doi.org/10.1088/0034-4885/79/4/046601 Creton C, Brown HR, Shull KR (1994a) Macromolecules 27:3174 Creton C, Brown HR, Deline VR (1994b) Macromolecules 27:1774 Deryagin BV, Krotova NA (1948) Doklady Akad Nauk SSSR 61:849 Deryagin BV, Krotova NA, Smilga VP (1973) “Adgezija tverdyck tel,” elIzdvo. Nauka, Moskva Deryagin BV, Krotova NA, Smilga VP (1978) Adhesion of solids. Consultants Bureau, New York/London, p ix. (Translation of original Russian version of 1973) Doi M, Edwards SF (1986) The theory of polymer dynamics. Clarendon, Oxford Evans JRG, Packham DE (1979) J Adhes Dent 10:177–191 de Gennes P-G (1998) Simple views on condensed matter, Series in modern condensed matter physics, vol 8. World Scientific, Singapore Gent AN, Lai SM (1995) Rubber Chem Technol 68:13 Gent AN, Lin CW (1990) J Adhes Dent 32:113 Good RJ (1972) J Adhes Dent 4:133 Helfand E (1992) Macromolecules 25:1676 Helfand E, Tagami Y (1972) J Chem Phys 56:3592 Hine PJ, El Muddarris S, Packham DE (1984) J Adhes Dent 17:207–229 Kendall K (2001) Molecular adhesion and its applications. Kluwer, London Khun TS (1970) The structure of scientific revolutions, 2nd edn. Chicago University Press, Chicago Kinloch AJ (1987) Adhesion and adhesives. Chapman and Hall, London Koh SK, Park SC, Kim SR, Choi WK, Jung HJ, Pae KD (1997) J Appl Polym Sci 64:1913 Leadley SR, Watts JF (1997) J Adhes Dent 60(10e):175–196 McBain JW, Hopkins DG (1925) J Phys Chem 29:188 Newton I (1730) Optics, 4th edn, bk 3, pt 1, William Innys, London. Packham DE (1998) The mechanical theory of adhesion – a seventy year perspective and its current status. In: van Ooij WJ, Anderson HR Jr (eds) First international congress on adhesion science and technology: invited papers. VSP, Utrecht, pp 81–108 Packham DE (2002) Surface roughness and adhesion. In: Chaudhury MK, Pocius AV (assoc ed) Adhesion science and engineering, vol II, Surfaces, chemistry and applications. Elsevier, Amsterdam, pp 317–349 Packham DE (2003) Mechanical theory of adhesion. In: Pizzi A, Mittal KL (eds) Handbook of adhesive technology, 2nd edn. Marcel Dekker, New York

2

Theories of Fundamental Adhesion

41

Pascoe MW (2005) Adhesives - historical perspective. In: Packham DE (ed) Handbook of adhesion, 2nd edn. Wiley, Chichester Popper KR (1959) The logic of scientific discovery. Hutchinson, London Possart W (1988) Int J Adhes Adhes 8(2):77–83 van der Putten AMT (1993) J Electrochem Soc 140:2376 Schneider B, Hennemann OD, Possart W (2002) J Adhes Dent 78(9):779–797 Voyutskii SS (1963) Autohesion and adhesion of high polymers. Interscience, New York Wadley L, Hodgskiss T, Grant M (2009) Proc Natl Acad Sci U S A. 106(24):9590–9594. https://doi. org/10.1073/pnas.0900957106 Wake WC (1982) Adhesion and the formulation of adhesives, 2nd edn. Applied Science, London Wool RP (1995) Polymer interfaces: structure and strength. Hanser, Munich Wool RP (2002) Diffusion and autohesion. In: Chaudhury MK, Pocius AV (assoc ed) Adhesion science and engineering, vol II, Surfaces, chemistry and applications. Elsevier, Amsterdam, pp 351–402 Wool RP (2005) Polymer diffusion: reptation and interdigitation/Polymer-polymer adhesion: incompatible interfaces/Polymer-polymer adhesion: models/Polymer-polymer adhesion: molecular weight dependence/Polymer-polymer adhesion: weld strength. In: Packham DE (ed) Handbook of adhesion, 2nd edn. Wiley, Chichester

3

Forces Involved in Adhesion Maurice Brogly

Contents 3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2 Physical Forces and Chemical Interactions: Basic Terms and Knowledge . . . . . . . . . . . . . . . 3.2.1 van der Waals Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.2 Acid-Base Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.2.3 Chemical Forces and Covalent Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3 Experimental Determination of Adhesion Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.3.1 Adhesion Forces Contribution to the Work of Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4 Case Study: In Situ Determination of the Work of Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.4.1 Force – Distance Experiments on Self-Assembly Monolayers (SAMs) . . . . . . . . . . 3.4.2 Influence of Capillary Forces on Adhesion Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

44 45 45 48 51 54 54 64 64 66 68 68

Abstract

The establishment of interfacial bonds through forces at the interface causes materials to attract one another. Therefore adhesion between materials and adhesive in assemblies are intimately related to the interatomic and intermolecular interactions at the interface of the two considered surfaces. Describing the mechanism responsible for adhesion in simple terms is difficult due to the complexity and evolving understanding of the subject. Nevertheless, when considering adhesion phenomena it is important to consider both the bulk and surface

M. Brogly (*) ENSCMu – LPIM – Equipe Chimie et Physico-Chimie des Polymères, Université de Haute Alsace, Mulhouse, France e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_3

43

44

M. Brogly

mechanical properties of the materials in contact and the type of interfacial forces established at the interface. High adhesion can only be obtained if the interface can sustain sufficient stress to induce dissipative forms of deformation, such as flow, yield, or crazing, in the polymer. Under most circumstances such dissipative processes can only be obtained when the interface is coupled with sufficient density of bonds. This chapter focuses on the description of the main forces responsible for adhesion, from strong covalent bonds to weak van der Waals forces, also considering some more specific interactions such as acid-base (like hydrogen bonding) or capillary forces (that could, as an example, influence adhesion of nanoparticles). The second part of this chapter concerns recent developments in experimental scanning probe techniques that may give assess to direct adhesion forces determination at the nanoscale. A case study is presented.

3.1

Introduction

The term of adhesion between two objects appears confusing or ambiguous, because generally it is employed to describe: first, the formation of the interface between a pair of materials, i.e., the establishment of interfacial bonds through forces at the interface which cause materials to attract one another and second, the breaking stress or energy required to break the assembly. One can easily understand that both interfacial forces and mechanical properties of adherends in the vicinity of the interface and in the bulk contribute to the global mechanical response of the assembly. Such a fundamental issue reflects a paradox that has stimulated intensive research for decades: if adhesion is only a matter of thermodynamics, the work needed for separation would be close to the surface energies of the adherends, which are typically 10
2–10
1 J/m2. Practical work of separation can be 10–106 greater! On the other hand, separation has often been viewed as involving processes at the molecular level. The force needed for separation should then be as large as 108 N/m2–109 N/m2. Therefore the first part of this chapter relates more precisely to the question: what, precisely, is the role of interfacial forces in determining the mechanical strength of bonded structures such as adhesive joints, composites, coatings, etc.? In other words, do interfacial forces determine, by their “weakness” or “strength,” the “weakness” or “strength” of joint systems? A related question is: how to access adhesion forces at a local scale? In recent years, atomic force microscopy (AFM) has become a powerful tool, sensitive enough, to detect small surface forces and to study adhesion at the nanoscale. Adhesion forces and surface mechanical properties of surfaces can be achieved with such a nanometer probe even if quantitative results need important calibration procedures. The purpose and scope of the second part of this chapter is to give highlight to the experimental methods that enables one to measure adhesion forces on the basis of AFM force-distance experiments. Relationships are proposed that give a complete understanding on how capillary and van der Waals forces contribute to adhesion forces at a local scale.

3

Forces Involved in Adhesion

45

3.2

Physical Forces and Chemical Interactions: Basic Terms and Knowledge

3.2.1

van der Waals Forces

The first forces subject to discussion are long-range van der Waals forces that include dispersion and polarization forces that arise from dipole moments induced in atoms and molecules by the electric fields of nearby charges and permanent dipoles.

Dispersion, Orientation, and Induction Forces Dispersion forces (London 1937) represent the most important contribution to van der Waals forces between atoms and molecules. They are always present, act between all atoms and molecules (including totally neutral ones), and generally exceed the dipole-dependent induction and orientation forces except for small highly polar molecules (Israelachvili 1991). Even if they are of week intensity, the perturbation responsible for dispersion forces is the existence of a finite dipole moment (given by the fluctuation of instantaneous positions of the electrons relative to the nuclear protons) that will generate an electric field that polarizes any nearby neutral atom, inducing a dipole moment in it. The same is for the second interacting atom. As a consequence, the resulting instantaneous dipole-dipole interaction gives rise to an instantaneous attractive force, called dispersion force. London has established using quantum mechanical perturbation theory the famous expression for the dispersion interaction energy (EDisp.) between two atoms or molecules (Eqs. 1 and 2) as follows: EDisp: ¼
CDisp: =r 6 ¼ ð
3hν1 ν2 α1 α2 Þ=

h
 2ð4πe0 Þ2 ðν1 þ ν2 Þr 6  ð
α1 α2 Þ=r 6

(1)

where r is the separation distance and CDisp. is a constant proportional to the product of the polarizabilities (α) of the considered atoms or molecules, h is Planck’s constant, ν is electronic absorption frequency, and e0 is the dielectric permittivity in vacuum. The main features of dispersion forces may be summarized as follows (Israelachvili 1991): – Dispersion forces are long-range forces and can be effective from large distances (>10 nm) down to interatomic distances; – Dispersion forces may be repulsive or attractive. Expression of the dispersion force does not follow a simple power law; – Dispersion forces can also align or orient molecules, though this effect is usually weak; – Dispersion interaction between two bodies is affected by the presence of other bodies nearby (nonadditivity interaction). For lager molecules, with diameters greater than about 0.5 nm, the London equation can no longer be used. As molecules grow in size, their center-to-center

46

M. Brogly

distance ceases to have any significance as regards the strength of cohesive energy. Likewise, the London equation cannot be applied to asymmetric (nonspherical) molecules such as polymers. Under such conditions, the exact dispersion interaction is difficult to compute. It is important then to consider that London theory must be modified and a complete description can be found in Spruch (1986) to account for the interaction between molecules at large distances (>10 nm). A retardation effect occurs because a finite time is required for an electromagnetic wave to be propagated through the medium the atoms or molecules are situated in. As a consequence, the dispersion energy drops-off with distance (EDisp. vary with 1/r7 rather than 1/r6) when the retardation effect comes into play. In the limit of very large separations, retarded dispersion interaction is given by the Casimir – Polder equation (Casimir and Polder 1948):   EDisp: ¼ ð
23hcα1 α2 Þ= 8π2 r 7  ð
α1 α2 Þ=r 7

(2)

where c is the speed of light in vacuum. Besides the most basic and predominant nonpolar interactions (dispersion forces), there are polarization or polar interactions between molecules of counter bodies, such as dipole-dipole interactions (Keesom 1922) and dipole-induced dipole interactions (Debye 1921). The essential difference between dispersion and polarization forces is that, while the former involve simultaneous excitation of both molecules, those for the latter involve only a passive partner. The Keesom orientation interaction energy between two molecules with permanent dipoles is temperature dependent and proportional to the dipole moments as follows:   
   EOrient: ¼
COrient: =r 6 ¼
μ21 μ21 = 3kT ð4πe0 Þ2 r 6 
μ21 μ21 =r 6

(3)

where r is the separation distance, and COrient. is a constant proportional to the temperature (T ), Boltzmann constant (k), and the square product of the dipole moments (μ) of the considered molecules. Debye argued that if the attraction energy was simply due to a Keesom effect then the interaction energy should be drastically reduced at high temperatures. Since experimental results were contrary to the prediction, he concluded that an additional attractive effect should be involved. He showed that an additional polar interaction should be induced between a permanent dipole and an induced dipole. The Debye induction interaction energy between two molecules with permanent dipoles is proportional to the square of the dipole moments and to the polarizabilities as follows: i   h EInduct: ðr Þ ¼
CInduct: =r 6 ¼
α1 μ21 þ α2 μ21 = ð4πe0 Þ2 r 6   
α1 μ22 þ α2 μ21 =r 6

(4)

where r is the separation distance, and CInduct. is a constant proportional to the sum of the square dipole moments (μ2) times the polarizabilities (α) of the considered

3

Forces Involved in Adhesion

47

molecules. If one of the interacting molecules has no permanent dipole (say the  second), the term α1 μ22 þ α2 μ21 is reduced to α2 μ21 . Finally the total van der Waals free energy of interaction has the following form:   Ewdw ðr Þ ¼
Cwdw =r 6 ¼
CDisp:þ COrient:þ CInduct: =r 6

(5)

An interesting property of van der Waals interactions must be considered that is the general dominance of dispersion forces in Eq. 5 except for highly polar molecules (i.e., water). Nevertheless rigorous computation of van der Waals force remains difficult.

van der Waals Forces Between Macroscopic Bodies To calculate the contribution of van der Waals forces in the interaction energy between macroscopic bodies one may sum the energies of all the atoms in one body with all the atoms in the other and thus obtain the “two-body” interaction energy. The resulting interaction laws are given in terms of a characteristic constant, namely, Hamaker constant (Hamaker 1937). The Hamaker constant is defined as: A ¼ π2 Cvdw ρ1 ρ2 and Cvdw  α1 α2

(6)

where Cwdw is the coefficient in the atom-atom pair van der Waals interaction energy, and ρ1 and ρ2 are the numbers of atoms per unit of volume in the considered macroscopic bodies. Typical values for the Hamaker constants of condensed phases (liquid or solid) are about 10
19 J for interactions in vacuum or air. Computation of Hamaker constants for polymers and various other materials could be found in Hough and White (1980). The nonretarded van der Waals interaction free energy between bodies of different geometries on the basis of pairwise additivity is given in Fig. 1. Direct applications are possible considering experiments such as surface force measurements with surface force apparatus (SFA) or adhesion force measurements with atomic force microscope (AFM).

R

b

a

D

D

D

EVdW = −

c

p.C VdW 6.D 3

EVdW = − A .( 1 + 1 + ln x ) 12 x 1 + x 1+x and x =

EVdW = −

A 12.p.D 2

D 2.R

Fig. 1 van der Waals free energy of interaction between interacting bodies ((a) atom – plane, (b) sphere – plane, and (c) plane – plane)

48

3.2.2

M. Brogly

Acid-Base Interactions

The Original Lewis Definition The Lewis definition of acid-base interactions are now over a half century old. Nevertheless they are always useful and have broadened their meaning and applications, covering concepts such as bond-formation, central atom-ligand interactions, electrophilic-nucleophilic reagents, cationoid-anionoid reagents, charge transfer complex formation, donor-acceptor reactions, etc. In 1923, Lewis reviewed and extensively elaborated the theory of the electron-pair bond (Lewis 1923). He states that a basic substance is one which has a lone pair of electrons which may be used to complete valence shell of another atom (the acid) and that an acid substance is one which can employ a lone pair from another molecule (the base) in completing the valence shell of one of its own atoms. Hence, the shared electron-pair bond model explains the existence of both nonpolar bond and polar link from the same premises. As the electrochemical natures of two atoms sharing an electron pair began to differ more and more, the pair should become more and more unequally shared, eventually becoming the sole property of the more electronegative atom and resulting in the formation of ions. Ionic and nonpolar bonds appear as logical extremes of a continuum of intermediate bond type. Differences in the continuum are only attributable to variations in the electron-pair donation, i.e., in the way in which the charges are localized within the molecule. As a consequence the distinctions between salts, acids, and bases; coordination compounds; and organic compounds are not of fundamental nature. Molecular Orbital (MO) Approach of Acid-Base Reactions Translated into the idiom of molecular orbital theory (Mulliken and Pearson 1969), the acid-base definitions should be read as follows: – A base is a species which employs a doubly occupied orbital in initiating a reaction. – An acid is a species which employs an empty orbital in initiating a reaction. The donor orbital is usually the highest occupied molecular orbital HOMO, and the acceptor orbital is usually the lowest unoccupied molecular orbital or LUMO. Thus it is important to understand that considering the interaction between two species, the acid is always defined relatively to the base and the base relatively to the acid. The molecular orbital definitions have a number of important consequences: First, it is not necessary that the donor and acceptor orbitals be localizable on a single atom or between two atoms, as implied by Lewis dot structures. That is, the orbitals may be multicentered even in a relatively localized representation. Thus donor-acceptor interactions involving delocalized electron systems (π-ring) (Hawthorne and Dunks 1972) are naturally assumed by the definitions. The same is for numerous organic or polymer molecules involving delocalized electrons. Second, the HOMO or donor orbital on a base is likely to be either bonding or nonbonding in character, the latter always being the case for monoatomic species. The LUMO or acceptor orbital of an acid is likely to be either antibonding or

3

Forces Involved in Adhesion

49

nonbonding in character, the latter always being the case for monoatomic species. Third, all degrees of electron donation are possible, ranging from zero to the complete transfer of one or more electrons from the donor to the acceptor. This continuity can be represented (Eq. 7) by wave functions (Klopman 1974) were the degree of donation increases as (a/b)2 does (a and b are weighting coefficients). ψ AB ¼ aψ A þ bψ B

(7)

where ψ AB is the wave function of the acid-base one-to-one adduct, ψ A is the ground state wave function of the acid, and ψ B the ground state wave function of the base. Table 1 reports the possible adducts classified in terms of the bonding properties of the donor and acceptor orbitals of the acid and base. Complete description of the mechanisms involved during the following adducts formation can be found in Jensen’s complete review (Jensen 1979). Even if the nature of an acid or a base is highly relative, a brief classification based on bonds properties and orbital symmetry can be proposed (where n represents the nonbonding orbital, σ represents the sigma orbital, and π the pi orbital. σ and π orbitals are also called bonding or antibonding orbitals if they address to a base or an acid, respectively). n donor: Lewis bases, complex and simple anions, carbanions, amines, oxides, sulfurs, phosphines, sulfoxides, ketones, ethers, alcohols, etc. π donor: Unsaturated and aromatic hydrocarbons having electron donor substituents, etc. σ donor: Saturated hydrocarbons, CO, single links like C-C, C-H, polar links like NaCl, BaO, silanes, etc. n acceptor: Lewis acids, simple cations, etc. π acceptor: N2, SO2, CO2, BF3, dienic and unsaturated and aromatic hydrocarbons having electron acceptor substituents, etc. σ acceptor: Brönsted acids, boranes, and alkanes having strong acceptor substituents like CHCl3, halogens, etc. Table 1 Possible acid-base adducts in terms of orbital properties

Acceptor Orbital of the Acid Non-bonding

Non-bonding Donor Orbital of the Base

Anti-bonding

n

σ*

π*

n

n.n

n.σ*

n.π*

σ

σ.n

σ.σ*

σ.π*

π

π.n

π.σ*

π.π*

Bonding

50

M. Brogly LUMO

LUMO

E LUMO

LUMO

HOMO LUMO

B HOMO HOMO A

LUMO C HOMO

HOMO F

D HOMO E

Fig. 2 Donor-acceptor molecular orbital interactions. HOMO represents the highest occupied molecular orbital and LUMO the lowest unoccupied molecular orbital

To summarize how acid-base reactions do work on the basis of molecular orbitals perturbation theory, the relative energies of the frontier orbitals HOMO and LUMO of a hypothetical species A and of the frontier orbitals of several hypothetical reaction partners B, C, D, E, and F are reported in Fig. 2. This figure is intended to represent possible variations of donor-acceptor properties in the broadest possible way. According to Fig. 2, with respect to B, complete electron transfer from B to A will be favorable and A will act as an oxidizing agent. With respect to C, the A(LUMO)C(HOMO) perturbation will be favorable and A will act as an acid. With respect to D, the A(HOMO)-D(LUMO) perturbation will be favorable and A will act as a base. Lastly, with respect to E, complete electron transfer from A to E will be favorable and A will act as a reducing agent. For F species, the frontiers orbitals energies are identical with those of A. Here neither species is clearly the donor or acceptor and species may display both behaviors simultaneously (case of multisite interactions encountered in concerted organic cycloaddition reactions).

The Case of Hydrogen Bonding In extension to the general discussion relative to Lewis acid-base interactions, it is important to understand that hydrogen bonding represents a typical example acidbase interaction. According to Pauling (1960), hydrogen bonding is partly covalent and partly ionic (polar). Nevertheless, it is obvious that electrostatic and charge transfer interactions are predominant for hydrogen bonds. In most cases, the principal charge transfer contribution is derived from the proton acceptor–proton donor charge transfer complex through the σ-type interactions.

3

Forces Involved in Adhesion

51

To conclude, the chemical phenomena subsumed by the category of acid-base reaction are the following: – Systems governed by the Arrhenius description, solvent system, Lux-Flood, and proton acid-base definitions. – Traditional coordination chemistry and “nonclassical” complexes. – Solvation and ionic dissociation phenomena in both aqueous and nonaqueous solutions. – Electrophilic and nucleophilic reactions in organic and organometallic chemistry. – Charge transfer complexes, molecular addition compounds, weak intermolecular forces, hydrogen bonds. – Salt melting and salt formation.

3.2.3

Chemical Forces and Covalent Bonding

The last case that should be considered in order to describe adhesion improvement is the possibility to form chemical bonds between adherends. Even if covalent bonds are just one variety in the continuum of bond types introduced before, it would be discussed in a special section as covalent bonding is in numerous cases required when high level of adhesive strength and durability of adhesive joints are required. Chemical bonds formed at an interface can greatly participate to the level of adhesion. The short-range forces involved in chemical bonding are quantum mechanical in nature and generally attractive. These bonds are generally considered as primary bonds in comparison with physical interactions such as van der Waals, which are called secondary interactions. The typical strength of a covalent bond is of the order of 100–1000 kJ/mol, whereas that of van der Waals interactions and hydrogen bonds does not exceed 50 kJ/mol. Chemical bonds depend strongly on the reactivity of both adhesive and substrate. To obtain strong adhesion, it is necessary that stress transfer could be achieved across the adhesive/substrate interface, so that cohesive failure is obtained. This stress transfer across the interface can be achieved by (i) chemical reaction to form coupling chains at the interface, (ii) the use of coupling chains placed at the interface, or (iii) chain interdiffusion, if the materials are sufficiently miscible.

Chemical Reaction to Form Coupling Chains at the Interface One of the most important adhesion fields involving interfacial chemical bonds is the use of adhesion promoter molecules, so-called coupling agents, to improve assembly joint strength. These species are able to react chemically on both ends (on one side with the adhesive, on the other side with the substrate). Silane-based promoter molecules are the most common type of coupling agent, widely employed in systems involving thermoset adhesives and glass, silicon oxide, silica, aluminum, or steel substrates. A complete review is available by Plueddemann (1982). Studies of the impact on epoxy fracture energy of model silanes (APS, GPS, MPS, AEAPS, PAPS) adsorbed on silicone oxide surface were done in Kramer’s group (Nakamura et al. 2007).

52

M. Brogly

The influence of chemical bonds on the joint strength (W), and more precisely on the intrinsic fracture energy (G0), of a glass/elastomer interface was investigated (Gent and Ahagon 1975). They have shown that G0 increases linearly with the surface concentration of vinylsilane coupling agent, between the glass substrate and the cross-linked elastomer. A second system that has been studied extensively because of its technical importance is the adhesion between epoxy resins and glass, particularly glass fibres. To improve the hydrolytic stability of the composite, the glass surfaces are normally modified by covering them with a thin layer of a silaneadhesion promoter. These silanes are well known to self-assemble into monomolecular and multi-molecular layers on surfaces (Flinn et al. 1994). Studies of the impact on epoxy fracture energy of model silanes (APS, GPS, MPS, AEAPS, PAPS) adsorbed on silicone oxide surface were done in Kramer’s group (Nakamura et al. 2007). One end of the silane molecule typically has di or tri methoxy or ethoxy functionality whilst the other end normally has amine or epoxy functionality. The ethoxy functionality is believed to condense with the hydroxyl functionality on the surface of the glass whilst the amine functionality can react with the epoxy. The amount of silane typically used is much too great to form a monolayer. Also, as the silane has multimethoxy or multiethoxy functionality, it can self-condense. The relatively thick layer of silane is believed to form a network and then the epoxy both mixes into and reacts with the network. Although it is clear that the silane causes covalent bonding between the glass and the epoxy, there is no simple way to estimate the actual density of coupling produced. Moreover, the silane condensed layer can be viewed as a rather porous layer (low degree of condensation) in which small molecules or additives have the possibility to diffuse. Similar silanes are also used as adhesion promoters on silicon dioxide. A second commonly used strategy to couple two bulk polymers is to introduce into one or both of the materials a small percentage of chemically modified chains (Lewis and Natarajan Gounder 1991; Tao et al. 2001) that can react with the other polymer to form coupling chains at the interface. A classic example (Laurens et al. 1994) is the introduction into polypropylene of some maleic anhydride grafted polypropylene chains to induce coupling with a polyamide such as Nylon 6. The maleic anhydride functionality can react with hydroxyl end groups on the polyamide chain to form a graft or block copolymer at the interface. The molecular mechanism of interfacial failure can be either pullout or scission of these coupling chains. Pullout at low force can occur if either one of the blocks in the copolymer formed at the interface is rather short or if there are too many coupling chains at the interface. In systems where multiple grafts are possible on a single chain, then the coupling chains can become so densely packed at the interface that they exclude other chains and so cannot entangle well with the bulk material, causing pullout failure. This situation has been observed when the chains themselves were long enough to entangle.

Coupling Chains Placed at Polymer-Polymer Interfaces Another case of interest concerns the adhesion between immiscible polymers. Adhesion is improved by the presence of diblock copolymers (Tirrel et al. 2001)

3

Forces Involved in Adhesion

53

or grafted chains (Brochart-Wyart et al. 1994) at the interface. The diblock is chosen so as that each block dissolves in the respective homopolymer. The forces involved are in most cases physical like van der Waals interactions which drive the mixing of each of the blocks across the polymer-polymer interface. Blocks should also have masses higher than the critical molecular weight between entanglements (Me) of the respective homopolymer. The form of coupling, that is to say the mechanism of interface failure, has been found to depend on the molecular weight of the coupling chains. Short chains can pull out of the bulk material at a force that increases with the length of the pulled-out section. As the length of the chains increases to somewhere between one and four times the length required to form an entanglement in the melt, the force required for pullout becomes greater than the force to break chains, so they fail by scission of covalent bonds. The extent of adhesion is strongly affected by the molecular failure mechanism as tough interfaces are normally only obtained when the failure is by scission. However, scission failure does not guarantee a tough interface. Nevertheless, Brown has proposed a molecular interpretation of the toughness of glassy polymers (Brown 1991) which can be applied to the failure of interfaces between reinforced immiscible polymers, for which a covalent scission (Brown et al. 1989) of the copolymer chains is observed at the interface. Assuming that entangled chains in the material are drawn into fibril, the fracture energy W is found proportional to: W  ν2 f 2 D=S

(8)

where ν is the surface density of entangled chains, f is the force required to break a polymer chain, D is the fibril diameter and S is the local stress applied at fibrilbulk interface. Moreover the chemist has the possibility to graft reactive functional groups on the blocks to promote chemical reactions between the blocks and the surrounding chains of the hosting polymer to improve adhesive strength by covalent bonding.

Coupling Polymers Chains by\Interdiffusion Coupling by chain interdiffusion (Wool 1995) can occur if two polymeric materials are miscible in each other, or at least sufficiently miscible to form an interphase. Welding is the most common form of interdiffusional coupling, but chain interdiffusion is also important in solvent bonding. Polymer diffusion normally occurs by the process of repetition in which the chain moves a long ‘tube’ formed by its entanglements with all the other chains. Hence interface coupling is formed by a chain end initially crossing the interface and then slowly more of the chain following it across. Failure after short joining times is thus expected to be by chain pullout; Even if polymer chain interdiffusion is based on physical interactions that give rise to the original differences of the chemical potentials which drive the mixing, chain scission failure occurs (i.e., chemical bonds rupture) as the diffusion distance increases. Even if it is not the main topic of this chapter, there has been a considerable amount of work on polymer chain repetition concept (de Gennes and Léger 1982) and on the kinetic aspects of chain coupling but the problem is not

54

M. Brogly

solved, as the location of chain end groups with respect to the interface is not easy to determine.

3.3

Experimental Determination of Adhesion Forces

3.3.1

Adhesion Forces Contribution to the Work of Adhesion

The magnitudes of forces responsible for adhesion (van der Waals, acid-base) can generally be related to fundamental thermodynamics quantities, such as surface free energies and work of adhesion. The work required to separate reversibly the interface between two bulk phases 1 and 2 from their equilibrium interacting distance to infinite one (Dupré 1869), represented on Fig. 3, is termed the “work of adhesion” and is equal to: W 12 ¼ γ 1 þ γ 2
γ 12

(9)

where γ 1 and γ 2 are the surface free energies of phases 1 and 2, respectively, and γ 12 is the interfacial free energy between phases 1 and 2. The work of adhesion is the decrease of Gibbs free energy per unit area when an interface is formed from two individual surfaces. Thus, the greater the interfacial attraction, the greater the work of adhesion will be and the smaller the interfacial free energy between phases 1 and 2 will be. Thereby, γ12 can be linked to the properties of the two individual phases. In the specific case of solid-liquid contact, phases 1 and 2 are commonly denoted phases S (solid) and L (liquid) and neglecting the spreading pressure of the vapor phase (V) of the liquid onto the solid, the solid-liquid work of adhesion (WSL) is: W SL ¼ γ S þ γ LV
γ SL ¼ γ LV ð1 þ cos θÞ

(10)

Fig. 3 Representation of Hertz and Johnson-KendallRoberts (JKR) contact shapes

1 1

2

2

3

Forces Involved in Adhesion

55

Liquid g LV

g LV Vapor

g SV q

Liquid g SL

gS

Solid

Solid

Fig. 4 Reversible separation of two interacting bodies. The corresponding work is called “work of adhesion”

where θ is the equilibrium contact angle at the three-phase contact point (Young 1805) as represented in Fig. 5. The first part of Eq. 10 is known as Dupré equation and its physical meaning is also represented in Fig. 4. More detailed consideration on wetting phenomena, surface energies, and thermodynamic of adhesion are presented in chapters relative to “Wetting” and “Work of adhesion” of this handbook. The subsections below will expose some experimental possibilities available to assess adhesion forces and interactions energies.

van der Waals Forces Contribution to the Work of Adhesion To reflect the contribution of the fundamental nature of the long-range interaction forces across the interface, it was suggested (Fowkes 1964) that surface free energies and work of adhesion may be expressed (Eq. 11) by the sum of two terms: the first one representative of London’s dispersion interactions (superscript D) and the second representative of non-dispersion forces (superscript ND), this latter include Debye induction forces, Keesom orientation forces, and acid-base interactions. γ ¼ γ D þ γ ND

and

ND W 12 ¼ W D 12 þ W 12

(11)

Fowkes proposes that the London dispersive part of the work of adhesion is quantified as twice the geometric mean of the dispersive component of the surface energy of solid 1 and 2. The total work of adhesion reduced to:  D 1=2 W 12 ¼ 2 γ D þ W ND 1 γ2 12

(12)

If only London’s forces are present across the interface, the work of adhesion is D fully determined because the nondispersive part is equal to zero. Indeed γ D 1 and γ 2 can be easily determined by wettability experiments (Wu 1982). The only existence of London interactions at the interface between two adherends leads to: W 12  W D 12

(13)

56

M. Brogly

Dispersive Work of Adhesion (WD) and Adhesive Contact Theories (Hertz, DMT, JKR) These theories concern the description of the contact between two homogeneous elastic bodies, usually two hemispheres or one hemisphere and one flat. Considered forces at the interface are only physical (van der Waals interactions). Hertz firstly describe such an elastic contact (Hertz 1896). He makes the hypothesis that solids are elastic, have smooth surfaces, and have small contact area with respect to solids dimensions, and finally have negligible adhesion and no friction. Outputs of Hertz theory are the shape of contact area, the load dependence of contact radius and stresses and strains in the contact zone. Johnson-Kendall-Roberts (JKR) extended the Hertz theory for the shape of two contacting but non-adhering elastic spheres to solids that do adhere (Johnson et al. 1971). Contact shapes are represented in Fig. 5 for Hertz and JKR approaches. The adhesion is due to the fact that real solids possess a finite surface tension. In the JKR theory, the attractive forces between the spheres are of infinitely short range and, hence, these forces manifest themselves only in the contact area. Derjaguin-Muller-Toporov (DMT) argue that since the attractive forces between the surfaces have a finite range, these forces also act in a region outside the contact zone where the surfaces are only a small distance apart (Derjaguin et al. 1975). The magnitude of these attractive forces outside the region of contact is responsible for deformation within the contact zone and determines the contact radius. Both theories show that the pull-off or adhesive force depends only on the interfacial tension and the radius of curvature of the undeformed surface and is independent of the contact radius and the elastic constants of the material such as the Young’s modulus and the Poisson’s ratio. From the JKR theory, the pull-off force or adhesion force of a sphere (radius R) from a flat is 3πRγ, where γ is the interfacial tension between the solid and the indenting sphere. The DMT theory predicts that the pull-off force is 4πRγ, i.e., greater by a factor of 4/3 than the adhesion force predicted by the JKR theory. The two

z

z

F

F

R

R r d

E n a Hertz contact

r d

E n a JKR contact

Fig. 5 Reversible separation of a liquid droplet from a surface. The corresponding work satisfies Dupré equation

3

Forces Involved in Adhesion

57

theories thus disagree in their predictions of interfacial tension from force measurements by this constant. However, there are two major differences between the theories that can be tested experimentally. One is the shape of the interface in contact under no load conditions and the second is the contact radius at which the surfaces separate. The JKR theory predicts that when the pull-off adhesion force equals 3πRγ, the surfaces separate from a finite area of contact and the radius at pull-off is 0.63 times the radius under no load conditions. On the other hand, the DMT theory predicts that at the instant of separation, the contact radius is 0, i.e., the surfaces separate only at the point where they have achieved their original undistorted shape. More interesting DMT and JKR theories give access to the thermodynamic work of adhesion. Indeed, based on different attractive stress distributions, established elastic theories for spherical contact, JKR and DMT, prescribe the adhesion force, FAdh, in two different ways as FAdh = 3/2πRW0 and FAdh = 2πRW0, respectively, with R is the sphere radius and W0 is the thermodynamic work of adhesion between bodies in contact. The debates on the validity of these two theories continued until Maugis showed that these two competing models are the limiting cases of his self-consistent transition model (Maugis 1992). This model adopts an adhesive potential between the surfaces exterior to the contact that exerts a constant stress, σ, until a critical separation distance, h, is reached (Dugdale approximation). Typically the value of h is chosen such that W0 = σh matches that of a potential describing van der Waals interactions between the surfaces. Nevertheless two important considerations must be outlined concerning the validity of these contact theories. The first one is that the considered solids in contact are purely elastic. Therefore the viscoelastic character of polymers still causes big difficulties in that context. The same is for the contact between a liquid adhesive and a solid substrate. As a consequence, experimental conditions have a major influence on the validity of the obtained results. Loading and unloading must be achieved in the linear elastic strain regime (first step of the loading regime for instance). The second point concerns the thermodynamic equilibrium of the contact. Experiments should be done as close as possible to equilibrium, i.e., close to zero strain rate. Making experiments at different strain rates and considering extrapolation to zero strain rate, one can assess the thermodynamic (reversible) work of adhesion (Amouroux and Léger 2003). If it is not the case, experiments will give the integral work of contact disruption (out of equilibrium).

Complications Arising from Capillary Forces The presence of liquid vapors in atmosphere can significantly affect the adhesion properties of bodies in contact. The major effect concerns the capillary condensation of water around surface contact sites. Liquids that easily wet surfaces (small contact angles) will spontaneously condense from vapor into surface pores and cracks as bulk liquid. As a consequence, a liquid meniscus can be formed between interacting bodies. This effect can strongly influence the adhesion of systems where the contact area is smaller than the capillary length (contact between a nanometer probe (AFM, nano-indentation) and a surface, adhesion of nanoparticules, etc.) At equilibrium, the meniscus curvature is related to the

58

M. Brogly

relative vapor pressure. For surfaces interacting via discrete contacts in ambient conditions, the validity of Dupré’s equation (Dupré 1869) is not obvious as the homogeneity of the medium is broken by the formation of water meniscus. Hence, neither the Dupré equation nor elastic continuum theories can be used to describe a meniscus-mediated contact. The equilibrium meniscus radius (r) is described by the Kelvin equation, r = γ LV V/kT ln( p/psat.), where γ LV is the surface tension of water, V is the molar volume of water, k is the Boltzmann constant, T is temperature, and p/psat. is the relative vapor pressure. The validity of the equation was experimentally verified by using a surface force apparatus (Fisher and Israelachvili 1979), where the effect of surface roughness is minimized by using smooth mica surfaces. The results imply that the Laplace pressure, PL = 2γ LV/r, acting inside a meniscus results in the capillary force between two contacting bodies. Thus, the maximum attractive adhesion force between a spherical body (radius R) and a flat surface due to the liquid bridge, at zero separation distance, and considering the direct solid-solid contact adhesion force inside the liquid annulus is (Israelachvili 1991): Fadh ¼ 4πRðγ LV cos θ þ γ SL Þ ¼ 4πRγ SV

(14)

Acid-Base Forces Contribution to the Work of Adhesion Acid-Base Work of Adhesion The knowledge of both the strength of acid-base interaction, i.e., the enthalpy of interaction, and the number of moles of acid or base functional groups interacting per unit area across the interface allow the determination of the reversible acid-base work of adhesion. In the particular case of adhesion between solids, it rapidly appears (Israelachvili 1991) that the contribution of the polar interactions (Keesom, Debye) to the thermodynamic work of adhesion could be neglected compared with both dispersive and acid-base contributions as experimentally confirmed (Fowkes et al. 1984). The acid-base component of the work of adhesion can be related to the variation of enthalpy per mole of acid-base interfacial adducts interaction, ΔH AB 12 , as follows:   AB AB AB W ND 12  W 12 ¼ f
ΔH 12 n12

(15)

where nAB 12 is the number of mole of acid-base bonds per unit interfacial area and f is equal to:    
1 f ¼ 1
d LnW AB 12 =dðLnT Þ

(16)

f is taken equal to unity in the case of the nondependence of W AB 12 with temperature. The quantitative determination of these quantities can be either theoretical or experimental.

3

Forces Involved in Adhesion

59

Theoretical determinations are mainly based on the assumption that the initial perturbation of the orbital of interacting species determines the course of a reaction or an interaction. In terms of equation of state, the perturbation theory was firstly developed. In terms of energy gap between HOMO and LUMO, acids and bases were defined by their hardness and softness. The Hard-Soft Acid-Base principle (Pearson 1963) describes some basic rules about kinetics and equilibrium of the acid-base interactions. The HSAB principle will be described as it has evolved in recent years on the basis of the density-functional theory (Parr and Yang 1989). For organic interactions, the following statements were proposed: – A hard acceptor has a high energy LUMO and usually a positive charge. – A soft acceptor has a low energy LUMO but does not necessarily have a positive charge. – A hard donor has a low energy HOMO and usually a negative charge. – A soft donor has a high energy HOMO but does not necessarily have a negative charge. The principle states that hard acid prefers to interact with a hard base, and vice versa, a soft acid with a soft base. A hard-hard interaction is fast because of large Coulombic attraction; a soft-soft interaction is fast because of large orbital overlap between HOMO and LUMO. However, the problem is what is the physical meaning of hardness? Berkowitz and Parr (1988) give a theoretical support to the absolute hardness, η. The absolute hardness determines the resistance of species to lose electrons. According to frontier orbital method, the relationship between η and the HOMO and LUMO energies is reduced to: η ¼
1=2½EHOMO
ELUMO 

(17)

where η is the absolute hardness, EHOMO is the energy level of the HOMO orbital, and ELUMO is the energy level of the LUMO orbital. Moreover, the absolute softness is defined as the reciprocal of the absolute hardness. The apparent success of the density-functional theory is to provide two parameters from which we can calculate the number of electrons transferred resulting mainly from the charge transfer between two molecules, i.e., from electrons flow until chemical potential reaches an equilibrium. As a first approximation, the number of electron transferred (NTrans) is given by (Shankar and Parr 1985): N Trans ¼ ½μB
μA Þ=2ðηA
ηB Þ

(18)

where μA is the chemical potential of the acid, μB is the chemical potential of the base, ηΑ is the absolute hardness of the acid, and ηΒ is the absolute hardness of the base. Calculation for various solid polymers could be found in a paper of Lee (1991).

60

M. Brogly

The subsections below will emphasize the experimental options for getting quantitative data on adhesion interaction enthalpy (ΔHAB). Effect of Ionocity and Covalency: Drago’s Concept Similar to the perturbation theory, Drago and Wayland proposed a four-parameter equation for predicting reaction enthalpies between acid and base species (Drago and Wayland 1965). Both species are each characterized by two independent parameters: an E value which measures their ability to participate in electrostatic bonding, and a C value which measures their ability to participate in covalent bond, leading to:
ΔH AB ¼ EA EB þ CA CB

(19)

where ΔHAB is the enthalpy of acid-base adduct formation, EA is the ability of the acid to participate in electrostatic bonding, EB is the ability of the base to participate in electrostatic bonding, CA is the ability of the acid to participate in covalent bonding, and CB is the ability of the base to participate in covalent bonding. A self-consistent set of E and C values is available (Drago 1973) for 33 acids and 48 bases, allowing ΔHAB prediction for over 1584 adducts. It is assumed that the conditions under which measurements are made (gas phase or poorly coordinating solvents) give rather constant entropy contribution and that most of the adducts are of one-to-one stoichiometry. The validity of Eq. 19 was clearly evidenced for polymer adsorption on various substrates. Even if Drago’s concept is an attempt to circumvent the quantum mechanical consideration of bond formation, the major importance of the above four parameters is their relationship with the HSAB principle. Actually, through the plot of EA versus CA for several liquids on a solid, one can obtain indirectly the values of the chemical softness 1/η from the slope of C/E, as represented by the following equation:   EA ¼
ΔHAB =EB
CA CB =EB

(20)

With the chemical hardness η and the chemical potential, one can easily apply the HSAB principle to acid-base interaction and calculate the number of electron transferred (Eq. 18). Effect of Amphotericity of Acid-Base Interaction: Gutmann’s Numbers Solvation, solvolysis, and ionic dissociation phenomena, in both aqueous and nonaqueous solutions are subsumed by the Lewis definitions. In addition to the previous discussion of the dual polarity character of Lewis acids and bases, it should be noted that many of them are amphoteric, by definition. Donor number, DN, were developed (Gutmann 1977) in order to correlate the behavior of a solute in a variety of donor solvents with a given basicity or donicity. A relative measurement of the basicity of a solvent is given by the enthalpy of its reaction with an arbitrarily chosen reference acid (SbCl5 in Gutmann’s scale). Latter, an acceptor number AN was introduced as the relative 31P Nuclear Magnetic Resonance (NMR) shift induced by triethylphosphine, and relative to acidic strength (AN = 0 for hexane and 100 for

3

Forces Involved in Adhesion

61

SbCl5). These AN numbers were modified (Riddle and Fowkes 1990) to express them in the correct enthalpic unit (kcal/mol). The most important assumption of Gutmann’s approach is that the order of base strengths established remains constant for all other acids (solutes), the value of the enthalpy of formation of a given adduct being linearly related to the donor number of the base (solvent) through Eq. 21:
ΔH AB ¼ aA DN B þ bA

(21)

where ΔHAB is the enthalpy of acid-base adduct formation, DNB is the donor number of the base, and aA and bA are constants characteristic of the acid. Graphically, this means that a plot of DNB for a series of donor solvents versus
ΔH of their adduct formation with a given acid gives a straight line, allowing the determination of aA and bA. By experimentally measuring the enthalpy of formation of only two adducts for a given acid, one can predict, through the resulting aA and bA values, the enthalpy of adduct formation of this acid with any other donor solvent for which DNB is known. Gutmann also proposes that the enthalpy of acid-base interaction could be approximated by a two-parameter equation of the form:
ΔH AB ¼ ðAN A DN B Þ=100

(22)

where ΔHAB is the enthalpy of acid-base adduct formation, DNB is the donor number of the base, and ANA is the acceptor number of the acid. The factor of 100 convert the ANA value from a percentage of the SbCl5 value to a decimal fraction. But one had to remind that on the 171 DNB values reported in the literature (Marcus 1984), only 50 were determined precisely, i.e., by flow microcalorimetric experiments. Nevertheless Gutmann’s approach has some limitation in the way it constitutes rather a prediction concept than an absolute way of determination of ΔHAB. Indeed this concept can be used to assess the determination of ΔHAB for probed liquids on a solid surface to assess the acid-base nature of the surface. IR Spectroscopic Tools to Assess Acid-Base Strength For specific functional groups involved in acid-base interaction, the enthalpy of acidbase adduct formation is related to the infrared wavenumber shift, Δν, of its absorption band (Fowkes et al. 1984) according to the following equation: ΔHAB ¼ kAB ΔνAB

(23)

Where ΔHAB is the molar enthalpy of acid-base adduct formation, ΔνAB is the infrared wavenumber shift, and kAB is a correlation constant between IR wavenumber shift and enthalpy. kAB is a characteristic constant of the functional group determined on the basis of compared infrared and microcalorimetric results of adduct formation. As an example, for the carbonyl functional group, C = O, kAB is equal to
0.99 kJ/cm/mol
1.

62

M. Brogly

The stretching wavenumber of the C = O vibration band is decreased by an amount ΔνΑΒ proportional to the enthalpy of acid-base bonding ΔHAB according to kAB. Such a methodology has been confirmed and case studies are available not only for polymer-solvent adduction but also for polymer/ polymer (Coleman et al. 1991) and polymer/metal adduction (Brogly et al. 1996, 1997). Moreover, infrared shifts can be used to determine Drago’s E and B constants of unknown polymers or substrates. If a test acid of known CA and EA gives a ΔνAB shift allowing the determination of ΔHAB then Eq. 19 is reduced to: E’B ¼
ΔHAB =EA
C’B ðCA =EA Þ

(24)

where E’B and C’B are “trial” values. The actual values of EB and CB occur at one point on the straight line obtained from Eq. 24. The use of multiple test acids shows that straight lines intersect at EB and CB. Concerning the limitation of this approach, one must consider that best results are obtained from infrared transmission spectra. This technique is well adapted for the study of solution (polymer/solvent acid-base interactions) or thin films (polymer/ polymer acid-base interactions). In the case of polymer/metal acid-base interaction, one can prefer infrared reflexion-absorption spectroscopy (IRAS) or polarization modulated infrared reflexion-absorption spectroscopy (PM-IRAS). Nevertheless this approach is a continuum-based (i.e., macroscopic) approach and it is important to understand that adhesion forces are not specified explicitly even if acid-base interactions contributed to ΔHAB. Adhesion Forces Measured by Atomic Force Microscopy Advances in nanotechnology and miniaturization of devices have led to the requirement of an increased understanding of the interactions and adhesion phenomena present in nanoscale contacts. The forces involved include van der Waals forces, hydrogen bonds, electrostatic forces, as well as chemical bond in some cases. Although van der Waals forces are only short range (nanometer range), they often rule the adhesion between two materials and are therefore responsible for contact interactions. In recent years, Atomic Force Microscopy (AFM) has become a powerful tool, sensitive enough, to detect small surface forces and to study adhesion at the nanoscale. Retraction profiles from Force-Distance AFM measurements provide quantitative, chemically sensitive, adhesion information at the nanoscale between probe tips and sample surfaces. Information about local adhesion, friction, or surface stiffness is readily obtainable. The purpose and scope of this section is to describe the experimental method that enables one to assess nanoadhesion forces. Indeed, the quantitative determination of the work of adhesion gives rise to numerous experimental and theoretical approaches. Nevertheless, all these approaches are mainly macroscopic and not made “in situ.” As a consequence, characterizations of local properties have stimulated widespread interest. Recent works have contributed to the understanding of adhesion mechanisms at a local scale. Burns et al. (1999) have demonstrated the influence of adhesion on sliding and on friction forces on chemically modified surfaces. Jones et al. (2002) have studied the effects of relative

3

Forces Involved in Adhesion

63

humidity, roughness, and surface treatment on adhesive properties of glass-glass contact. But few studies have been devoted to the contact between a local probe and a chemically modified surface or polymer surface. Dissipative phenomena in the adhesive contact by using dynamic AFM were nevertheless studied (Boisgard et al. 2002). Finally Brogly et al. (2009) have done experiments on chemically modified polymer surfaces in order to investigate the interplay between surface chemistry and mechanical properties at a local scale. Force Versus Distance Measurements with an AFM

Force measurements with AFM, in the contact mode, consist in detecting the deflection of a spring (or cantilever) bearing tip at its end, when interacting with the sample surface. The deflection of the cantilever is detected by an optical device (four quadrant photodiode) while the tip is vertically moved forward and backward thanks to a piezoelectric ceramic (or actuator). Thus, provided that the spring constant of the cantilever is known, one can obtain a Deflection-Distance (DD) curve and then a Force-Distance (FD) curve, by using Hooke’s law. The DD curves presented in this chapter were performed in air. A schematic representation of a DD curve obtained when probing a hard surface is reported in Fig. 6. In zone A, the cantilever is far from the surface and stays in a state of equilibrium (no interaction with the surface). The cantilever deflection is zero. During the approach toward (or withdrawal from) the surface, the tip interacts with the sample and a jump in (or jump-off) contact occurs (zones B (for loading) and E (for unloading)). These instabilities take place because the cantilever becomes mechanically unstable. Usually, for undeformable surfaces, because of mechanical instabilities, jump-in contact is not significant to determine attractive van der Waals forces. When in contact, the cantilever deflection is equal to the piezoelectric ceramic displacement provided no indentation of the substrate occurs (zones C (for loading) and D (for unloading)). An undeformable reference sample (cleaned silicon wafer) is used to scale the DD curve in deflection by fixing to unity the slope value of the A

B Cantilever deflection (i.e. Force)

C

C B

A

Djump-in Approach

Jump-in Deflection

Retraction D

D pull-off

FAdh = k tip . D pull-off

Jump-off Deflection

Piezo displacement z (nm)

Fig. 6 Schematic representation of a Force-Distance curve

64

M. Brogly

contact line. Due to adhesion forces during contact, the jump-off is more interesting than the jump in and occurs in position D. Considering the cantilever like a spring, knowing its spring constant, one can obtain the adhesion force (FAdh.) between the tip and the sample by using Hooke’s law: FAdh: ¼ ktip: Dpull-off

(25)

where, ktip represents the spring constant of the cantilever and Dpull-off represents the jump off deflection during retraction of the tip from the surface. The DMT theory also establishes a relationship between the adhesion force (Fadh), the tip radius (R), and the thermodynamic work of adhesion (W0): Fadh ¼ 2πRW 0

(26)

Strictly speaking, experimental measurement is made out of equilibrium leading determination of the integral work of disruption rather than reversible work of adhesion. On the basis of this relation, one can deduce W0 from experimental adhesion forces. In order to carry out a quantitative analysis, different experimental points should be taken into consideration such as: selection of tips that possess apex spherical shape (SEM imaging), cantilever spring constant, tip radius, linearity of the photodiodes, piezodriver hysteresis, cantilever and piezo thermal stabilities and tip contamination (Noel et al. 2004).

3.4

Case Study: In Situ Determination of the Work of Adhesion

3.4.1

Force – Distance Experiments on Self-Assembly Monolayers (SAMs)

In order to quantitatively and in situ measure the work of adhesion between an AFM tip and a given surface, one has to proceed according to a proper procedure that enables to control surface chemistry and thus surface forces. Organosilane molecules spontaneously organize to form self-assembly monolayers (SAMs). Such a selforganizing process gives rise to the elaboration of homogeneous model surfaces of controlled surface chemistry. Three different techniques are frequently used to obtain SAMs: Langmuir-Blodgett techniques, involving an air-water interface to transfer the assembled film to a solid substrate, solution adsorption of film molecules onto the substrate, and vapor–phase molecular self-assembling technique, which uses vapor deposition of the film onto the substrate. This last technique should be preferred, because the lack of solvent prevents the SAMs of a possible incorporation of small solvent molecules contamination and defects. Molecular films prepared with this method are homogeneous, stable, and resistant. Hydrophobic model surfaces could be prepared by using hexadecyltrichlorosilane (C16H42Cl3Si or SiCH3) and 1H,1H,2H,2H–perfluorodecylmethyldichlorosilane (C11H7Cl2F17Si or SiCF3) and

3

Forces Involved in Adhesion

Table 2 Water contact angles and surfaces energies of self-assembly monolayers (SAM) on silicon wafers substrates (From Noel et al. 2004)

65

Substrates Si–CF3 Si–CH3 Si–COOR Si–NH2 Si–OH

Contact angle of water ( ) 106  2 103  2 71  2 57  2 62

Surface energy (mJ/m2) 21  1 22  1 43  1 53  1 76  1

30

Cantilever deflection (nm)

10 −10

−5

−10 0

5

−30

SiCF3

−50

SiCH3

−70

Sias received

−90

SiEster

−110

SiNH2

−130

SiOH

−150 Piezo displacement (nm)

Fig. 7 Experimental Deflection-Distance curves (retraction) on functionalized silicon wafers

hydrophilic model surfaces by using (6-aminohexyl)-aminopropyltrimethoxysilane (C12H30N2O3Si or SiNH2) and 2(carbomethoxy)ethyltrichlorosilane (C4H7Cl3O2Si or Siester), for instance. Equilibrium contact angle measurements with water droplets and surface energy (determined by wettability using various probed liquids (water, tricresyl phosphate, α-bromonaphtalene, diiodomethane)) values obtained on silicon wafer grafted with such organosilanes are gathered in Table 2. It is important to note that surface roughness is extremely low ( 90
, and this represents “poor wetting.” In the extreme case of θ0 ! 0, the liquid becomes a thin film (e.g., oil deposited on a steel sheet). This is often called

gLV gSV

h

q0 gSL

r

gLV gSV

q0 gSL

Fig. 1 Schematic representation of low (left) and high (right) static contact angles, θ0. Interfacial tensions are represented by arrows and termed γ IJ (S solid, L liquid, V vapor/surrounding fluid phase)

4

Wetting of Solids

73

spreading. Although it will not be treated here, the thickness of such films depends on long-range molecular forces, which may span the entire liquid thickness of the film, leading to what is effectively “interference” between the two interfaces, liquid–solid and vapor–liquid (cf. chapter Spreading of Liquids on Substrates). However, this is beyond the scope of the present chapter as such spreading is fundamentally different from simple wetting. At the other extreme, the case of θ0 ! 180
, non-wetting, may be considered. This cannot, in fact, exist if the solid is truly smooth and if the environment is only gaseous but can in some cases when two liquids are involved (Shanahan et al. 1982).The importance of contact angle is that it determines the extent of (solid) surface coverage by a liquid. It is intuitively clear that the interfacial area covered by a sessile drop of given, fixed volume V of liquid increases as contact angle decreases, but this may be readily written formally in the case of a spherical, cap-shaped drop of contact radius r and height h. Writing t = tan (θ0/2) = h/r, and with a standard trigonometric formula for the volume, V, of a spherical segment: V = π h(3r2 + h2)/6, it is readily shown that the contact area, A, is given by
2=3 6V A ¼ π  r2 ¼ π 1=3 (1) t ð3 þ t 2 Þ In the range of possible contact angles, 0
θ0 180
, t is a monotonically increasing function of θ0, and therefore, A increases as θ0 decreases, at constant V. So, decreasing θ0 implies better coverage, or wetting, but what controls θ0? Between the solid and liquid and liquid and vapor phases, there are phase boundaries of the different substances. These can be imagined as 2D faces, or interfaces, but nature is never so clear-cut, and a vague transition zone, the interphase, typically of the order of a few nm (nanometers) or a few tens of nm thick constitutes the changeover. It has to be mentioned that an interphase and its width is not a uniquely defined thing as it has to be attributed to the macroscopic physical or chemical quantity of interest. In other words, each thermodynamic quantity of a heterogeneous system can possess its particular spatial function across its own interphase. The interphase will be considered in Sects. 2 and 3. However, accepting as a first approximation that 2D interfaces separate phases, it can be seen in the typical case exemplified in Fig. 1 that there are three interfaces: SL, LV, and SV (S, solid; L, liquid; V, vapor/surrounding medium), each of which has associated with it an interfacial tension γ IJ where I and J refer to S, L, and V (see Sect. 2 for definition). Note that the very concept of “force/distance” is meaningless in absolute terms, since the force will have no area to “pull” or “push” on: the idea of a “stress” becomes vague. In fact, it is Gibbs’ concept of “surface excess” (cf. Sect. 2) that allows one to see that the forces are, in fact, acting over a thin layer at the interface, and thus the concept of stress is still valid. The three phases, and therefore interfaces, meet at the drop periphery, at what is termed term the triple line or contact line. Again, this is not really a line, but a “tube,” but the concept of line is adequate in Gibbs’ treatment. A force balance between the various γ IJ exists at the triple line, and, at equilibrium, this is principally what governs the value of θ0 (see below for details).

74

M. E. R. Shanahan and W. Possart

Accepting the concept of contact angle, its importance may be appreciated in many sectors, biological and industrial. Without resorting to equations, it is intuitive (and correct) to realize that it is more difficult to remove an oil film from steel (contact angle ca. 0
) than a drop of mercury from glass (contact angle ca. 130
). Not only is adherence of the oil intrinsically better, but for a given volume, coverage is superior. The same, of course, has to apply to paints and inks. If an insecticide is applied to a leaf, it is clearly inefficient if, when the leaf slopes, the liquid simply rolls off. Therefore, a small, preferably zero, contact angle is required. In the case of adhesives, this significant coverage and propensity to wet the solid in contact is crucial in forming a good bond as it gives the molecules the opportunity to form adhesive interactions, even if the “true,” chemical, bond is produced later by hardening of the adhesive (drying, chemical reaction, cooling). If the surface is rough, the same wetting tendency will help ensure filling the interstices. This not only increases true contact area, but reduces stress concentrations, liable to be crack initiators. There are, of course, many cases where poor wetting is required. Nature masters this beautifully in the case of, for example, duck feathers and lotus leaves (Zhang et al. 2009). Mankind attempts to exploit it with PTFE and other polymers as cooking utensil coatings and raincoat treatments. The main objective of this chapter is to consider in more detail the phenomenology of wetting, specifically in the static case (kinematics are covered in chapter Spreading of Liquids on Substrates). In other words, the triple phase line will principally be considered to be at rest. The description will now be couched in terms of the needs of thermodynamic equilibrium.

4.2

Thermodynamics of a Phase Boundary: An Introduction

As mentioned above, the boundary between two thermodynamic phases must be a vague transition zone for any physical quantity because nature does not like sharp steps. Hence, in principle, any real phase boundary extends in three dimensions – the interphase (cf. Fig. 2a). This apparently innocent comment has the important result that all extensive thermodynamic quantities, such as concentrations, energies, entropy, etc., will depend on position across the interphase. This causes trouble when it comes to an exact description of a thermodynamic system with more than one phase, because there is virtually no experimental access to the spatial dependence of thermodynamic quantities across the phase boundary. Moreover, the beauty of classic thermodynamics is very much due to the assumption that parameters remain constant across a phase. The exact volume V of such a two-phase system is given by the total V ¼ V A þ V B þ V Interphase

(2)

where the terms on the right-hand side refer, respectively, to the two “bulk” phases and to the interphase. Hence, a method to circumvent the problem is needed. A pioneer in this subject (among many others!) was J.W. Gibbs, who introduced in a standardized way the notion of a “discontinuity” (Bumstead and Van Name 1961;

4

Wetting of Solids

75

Interphase

Discontinuity face

r r Phase A

Phase B

Phase A

Phase B

Fig. 2 Schematic drawing for an interphase in a radially configuration of spherical symmetry of two phases A and B. The interphase shown in (a) (left) is replaced by the 2D discontinuity face (b) (right) in interface thermodynamics

Gaydos et al. 1996). A brief version of this thermodynamic description for phase boundaries is now introduced.

4.2.1

Brief Description of Gibbs’ Model for a Phase Boundary

The “real” interphase (Bakker 1902) is replaced by the “artificial” discontinuity face (D-face) with a certain area AD AB – cf. Fig. 2b. (For example, see (Rusanov 1978; Rusanov and Prokhorov 1996) for more details.) The position of the D-face can be chosen arbitrarily, but only once, from the theoretical point of view or, in measurements, is fixed by the experiment itself. Indeed, the experimenter is unable to select the position of the D-face deliberately. Usually, he does not know that position exactly. Although the D-face and the interface are very similar in many cases, the term “D-face” will be used here because similarity does not mean identity from the theoretical viewpoint. As a first consequence, (2) must be replaced by V ¼ VA þ VB

(3)

for the two-phase system, i.e., the interphase volume VInterphase is “eliminated,” being arbitrarily attributed partially to each of phases A and B according to the given position of the D-face. Moreover, the balance of any extensive thermodynamic quantity, Y, in the two-phase system is no longer Y ¼ Y A þ Y B þ Y Interphase

(4)

Y ¼ YA þ YB þ YD AB

(5)

but is written as

76

M. E. R. Shanahan and W. Possart

Here, Y D AB denotes the excess of Y with respect to the D-face. It is a correction that guarantees the correct value for Y in the whole two-phase system after the interphase has been ignored. Note that any excess quantity can adopt positive and negative values as well. The extensive Y can be referred to the D-face area AD AB, thus providing the corresponding intensive quantity y y¼

Y ¼ yA þ yB þ yD AB AD AB

(6)

For example, the number of moles ni of a component i in the system ni ¼ ni, A þ ni, B þ nInterphase

(7a)

ni, A, ni, B, nInterphase = number of moles of component i in phases A and B and in the interphase is replaced by D ni ¼ ni , A þ ni , B þ nD i, AB ¼ ci, A  V A þ ci, B  V A þ ni, AB

(7b)

nD i, AB = excess number of moles of i for the given D-face ci, A, ci, B = concentrations of i (mole m3) in phases A and B A simple rearrangement of (7b) provides the definition of the absolute adsorption of component i in a thermodynamic system with N components ΓD i, AB

def

¼

nD i, AB AD AB

¼

ni  ci, A  V A  ci, B  V A AD AB

with

i ¼ 1...N

(8)

Indirectly, these N excess quantities replace the concentration profiles of all components in the system. Obviously, the value for any of the Γ D i, AB depends on and hence on the position for the D-face. As that position can be chosen only AD AB can be made zero, while all once for a given phase boundary, only one of the Γ D i, AB D other (N  1) Γ j, AB adopt nonzero values. (This naïve statement will be important for the coming discussions.) The interfacial tension is another important quantity for the thermodynamics of interfaces. The starting point stems from continuum mechanics for deformable solids ! where the balance of forces at point r in the solid is given by ! ! d F r

!

¼f

  ! ! ! r ¼ Div ! σ r

dV   ! f r = density of the external force that deforms a solid   ! ! ! σ r = tensor of internal stresses in the deformed solid !

(9)

4

Wetting of Solids

77

and the mechanical work of deformation ΔW of the volume V ΔW ¼

ðX V

ðσ ik  δeik Þ dV

(10)

i, k

! ! ! σ ik, δeik = components of the stress and the deformation tensors, ! σ r , ! e   ! r , respectively. In thermodynamics, pressure is preferred, but now it is a tensor which is defined as   ! ! ! ! ! σ r e r ¼ !

(11)

  !  ! ! ! ! ! p r , ! e r are symmetric, and hence they can be All three tensors ! σ r , ! transformed into diagonal form, provided a suitable system of coordinates is chosen. Then, the analysis of (10) for the boundary between the two phases will provide a definition for interfacial tension. As an illustration, the derivation is considered for the case of an isotropic horizontal planar phase boundary. The D-face spans the x,yplane of the Cartesian coordinate system, and the z-axis is parallel to, but in the opposite direction to, the gravitational force which is the external field here !

!

f ðzÞ ¼ g  ρðzÞ k

(12)

g = gravity acceleration ρ( z) = mass density !

k = unit vector for z-direction.

From (Eqs. 9, 11, and 12), it follows that 1 @pxx 1 1 B @x C 0 0 0 B @p C C B A 0 0 A ¼ B yy C ¼ @ B @y C g  ρðzÞ pzz A @ @pzz @z 0

0

pxx Div@ 0 0

0 pyy 0

and hence ðz pxx ðzÞ ¼ const;

pyy ðzÞ ¼ const;

pzz ðzÞ  p ¼ g 

ρðzÞ dz 1

p = isostatic external pressure.

(13a  c)

78

M. E. R. Shanahan and W. Possart

The two tangential pressure components pxx and pyy have to be equal for any z, by symmetry pxx ðzÞ ¼ pyy ðzÞ  pT ðzÞ

(13d)

pT = pressure in the x,y-plane at z Usually, the interphase is very thin (d  109–107 m) and (13c) can be approximated by d

ð2 ρðzÞ dz  0 ! pzz ðzÞ  pN  p ¼ const

g

(13e)

d2

pN = pressure component in z-direction. With (13), (10) is now ð ΔW ¼  ð ¼





  pT  δexx þ δeyy þ p  δezz dV

V

  pT  δexx þ δeyy  p  δexx þ δeyy þ p  δexx þ δeyy þ p  δezz dV

V





ð ¼





ð      ðp  pT Þ  δexx þ δeyy dV  p  δexx þ δeyy þ δezz dV

V

V

and for small deformations 1 ð

ΔW ¼

ΔAD AB



ðp  pT Þ dz  p  ΔV

(14a)

1

The second term is the well-known work of volume expansion, while the first term 1 ð

ΔW D AB

¼

ΔAD AB



ðp  pT Þ dz

(14b)

1

gives the work for increasing the area of the D-face by ΔAD AB. Equation (14) implies the definition of the interfacial tension for the chosen isotropic planar D-face: γD AB

def

1 ð

¼

ðp  pT Þ dz 1

(15a)

4

Wetting of Solids

79

With an assumed spherical shape for the isotropic interphase, a similar mathematical treatment provides the interfacial tension for the spherical D-face of radius rD γD AB

1 ¼ 2 rD

2r 3 1 ðD ð 4 ðpA  pT Þ r2 dr þ ðpB  pT Þ r2 dr5 0

(15b)

rD

pA, pB = isostatic pressures inside the phases A and B. They are different due to the curved interphase. These two examples illustrate that γ D AB is, in general, an excess quantity for a given D-face. As a consequence of this result, there is no physical difference between interfacial and surface tension in thermodynamic equilibrium. In the literature, the term surface tension is attributed to the D-face, or interface, between a condensed phase and the corresponding vapor phase. Some authors define the surface tension as the interfacial tension of a condensed phase in contact with vacuum. This is incorrect in the framework of equilibrium thermodynamics. A condensed phase will fill a vacuum with its vapor phase to be at thermodynamic equilibrium. Hence, any interfacial tension results from a defined pair of phases, never from a single phase, and they need to be marked by the subscript “AB” or similar. Otherwise, the interfacial tension would be ill-defined. γ D AB accounts for the real distribution of pressure or stress inside the phase boundary. It is called a tension because it possesses the dimension of a force per unit length [N m1] in the 2D space of the D-face and indeed acts somewhat like the tension in a membrane (except that there is no elastic proportionality between the tension and extension). Moreover, (15b) reveals that the value for γ D AB usually depends on is always positive. This is a the position, here rD, chosen for the given D-face. γ D AB condition of stability for the interphase and hence for the phase itself, and it is proven by experiment and also by statistical mechanics calculations. The value remains almost constant when p decreases. Therefore, pT must be negative, at least inside the dominant part of the interphase, and as a consequence, the interphase and the corresponding D-face will adopt the minimum volume and area, respectively. To apply thermodynamic energy balances (first law), knowledge of how mechanical work changes the state of the system must be available. As the phase boundary is included, the usual equations have to be revised. Figure 3 depicts the situation for the volume expansion δVof the phase A that is accompanied by a change of the principal radii of curvature from R1, R2 to (R1 + δz), (R1 + δz), respectively.Now, the mechanical work can be split into four parts. The first one comes from the work of volume expansion δW V ¼ pA  dV A  pB  dV B

(16a)

Part two results from the corresponding increase of area of the D-face according to (14b)

80

M. E. R. Shanahan and W. Possart

Fig. 3 Geometric scheme for the of the volume increase δV of a phase A with a curved D-face (D ! D0 )

z y+

δy

x+

y

δx

δz

D’

x

D R2 R1

O

D D δW D AB ¼ γ AB  dAAB

(16b)

The increase of D-face area can be accompanied by a change of the radii of curvature. This corresponds to the third contribution to the mechanical work:

D δW D curvature ¼ AAB  C 1  d

1 R1



þ AD AB  C 2  d

1 R2

 (16c)

C1, C2 = proportionality factors. The fourth and last part is related to the action of external forces such as gravity. The D-face possesses the total excess mass mD AB ¼

N X

mD i, AB ¼

i¼1

N X

D M i  nD i, AB ¼ AAB 

i¼1

N X

M i  ΓD i, AB

i¼1

which is raised by a height dh when the D-face shifts in the gravitational field. Hence, the corresponding work is D δW D gravitation ¼ AAB 

N  X i¼1

The total work is equal to

 M i  ΓD i, AB  g  dh

(16d)

4

Wetting of Solids

81

δW ¼ pA  dV A  pB  dV B þ



 1 1 D D D D þ γ AB  dAAB þ AAB  C 1  d þ AAB  C 2  d þ R1 R2 N   X þ AD  M i  ΓD AB i, AB  g  dh

(16f )

i¼1

In this equation, however, the work would seem to depend on AD AB and hence on the arbitrarily chosen position of the D-face. This is impossible from a physical point of view and terms of (16f) must clearly correct for this implicitly. The paradox is ! revealed from a virtual displacement δN of the D-face along its normal N . The corresponding virtual mechanical work is δW = 0. Another expression available is

δV A ¼ AD AB  δN ,

δV B ¼ AD AB  δN ,

D dAD AB ¼ AAB

 1 1 þ δN , R1 R1

δN ¼ δR1 ¼ δR2 and (16f) now provides the generalized Laplace equation

pA  pB ¼ γ D AB

1 1 þ R1 R2

 

N   C1 C2 X  þ M i  ΓD i, AB  g  cos φ 2 2 R1 R2 i¼1

(17a)

!

where ϕ = angle between N and vertical direction. The complete condition for mechanical equilibrium of a two-phase system with a curved phase boundary in a gravitational field is now available. It shows that the pressure difference between two curved phases depends on the properties of the D-face: interfacial tension, curvature, and adsorptions Γ D i, AB .Usually, the absolute D adsorptions Γ i, AB, as well as C1 and C2, are small. Then, the following approximation of (17a) can be used:

pA  pB ¼

γD AB

1 1 þ R1 R2

 (17b)

This is known as Laplace’s equation. For an anisotropic phase boundary such as in the case of oriented molecules, the interfacial tension becomes a diagonal tensor ! ! γD AB ¼



γ 11 0

0 γ 22

 (18a)

and the trace of the tensor provides the value for the interfacial tension 1 γ ¼ ðγ 11 þ γ 22 Þ 2

(18b)

82

M. E. R. Shanahan and W. Possart

Finally, the corresponding forms of Equation (17) become pA  pB ¼

N   γD γD C1 C2 X 11 þ 22  2  2 þ M i  ΓD i, AB  g  cos φ R1 R2 R1 R2 i¼1

pA  pB ¼

γD γD 11 þ 22 R1 R2

(19a)

(19b)

With these results, the energy balance of the two-phase system can be evaluated in detail.

4.2.2

Energy Balance for a Two-Phase System: Thermodynamic Potentials

The thermodynamic potential for the internal energy U in any thermodynamic system provides the starting point dU ¼ T  dS þ δW þ

X

μi  dmi

i

T = absolute temperature S = entropy μi, mi = chemical potential and mass of component i. Using (5 and 16f) for a system with the isotropic phases A and B, the following relation is obtained dU ¼ dU A þ dU B þ dU D AB   ¼ T  dS A þ dS B þ dS D AB  pA  dV A  pB  dV B



 1 1 D D D þ γD  dA þ A  C  d  C  d þ A 1 2 AB AB AB AB R1 R2 N     X X þ AD M i  ΓD μi  dmi, A þ dmi, B þ dmD AB  i, AB  g  dh þ i, AB

(20)

i

i¼1

The terms for the D-face are easily identified. For clarity and brevity, however, the following treatment is restricted to a plane interface and neglect gravity. Then, (20) reduces to   D dU ¼ dU A þ dU B þ dU D AB ¼ T  dS Aþ dS B þ dS AB  pA  dVA  pB  dV B þ X D μi  dmi, A þ dmi, B þ dmD þ γD AB  dAAB þ i, AB i

(21)

4

Wetting of Solids

83

Integration provides the full internal energy of the two-phase system   U ¼ T SA þ SB þ SD AB  pA V A  pB V B þ   X D μi  mi, A þ mi, B þ mD þ γD AB AAB þ i, AB

(22)

i

and the excess internal energy of the force-free plane D-face D D D UD AB ¼ T  S AB þ γ AB AAB þ

X

μi  mD i, AB

(23)

X UD D AB ¼ T  sD μi  Γ D AB þ γ AB þ i, AB D AAB i

(24)

i

Now, specific quantities are introduced uD AB ¼ SD

AB sD AB ¼ AD = specific excess entropy of the D-face AB

in order to get intensive thermodynamic quantities. Moreover, the excess Helmholtz free energy is D D D D FD AB ¼ U AB  T  S AB ¼ γ AB AAB þ

X

μi  mD i, AB

(25)

i

with the corresponding specific excess Helmholtz free energy fD AB ¼

X FD D D D AB μi  Γ D i, AB D ¼ uAB  T  sAB ¼ γ AB þ AAB i

(26)

These results give access to the interfacial tension as a thermodynamic excess quantity of the force-free plane D-face. Note that the interfacial tension and the absolute adsorptions of all components contribute to the interfacial energy, both the specific excess internal energy and the specific excess Helmholtz free energy. At a first glance, this statement is very natural, indeed. In the literature, however, interfacial tension and interfacial energy are very often considered as synonymous. This is incorrect. Finally, the total differential for (23) is derived and compared with the corresponding terms in (21). This yields the Gibbs adsorption equation D dγ D AB ¼ sAB dT 

X

ΓD i, AB dμi

(27)

i

Equation (27) depicts the interfacial tension as a function of temperature and, via the chemical potentials, of pressure and composition of the two-phase system.

84

4.3

M. E. R. Shanahan and W. Possart

Thermodynamic Equilibrium of Three Phases: Basics

From Sect. 1, it is clear that wetting and the contact angle depend on all three contacting phases. Hence, the Gibbs concept of excess quantities as thermodynamic quantities for phase boundaries has to be extended in an appropriate way. For a sessile drop on a flat, homogeneous solid, the situation is illustrated in Fig. 4. The three phases create a new microscopic boundary region at their “line” of common contact. For the sessile drop, this region forms a curved tube along the rim of the drop on the solid. Inside the tube, the three D-faces for the three two-phase boundaries respond by deforming, shown very much exaggerated in the drawing. The real three-phase region (tube) is replaced by a triple line (TL) as the new model element for thermodynamic considerations. In theory, its position results from the extrapolation of the three undisturbed D-faces to their common intersection. Accordingly, the tangent lines on these extrapolated D-faces DSL and DLV to the point of intersection lead by definition to the thermodynamic contact angle θe. For reasons of exactness, θe is introduced for the thermodynamic equilibrium of the three-phase system. The static contact angle θ0 in Sect. 1 refers to the triple line at rest, i.e., to mechanical equilibrium which is a necessary but insufficient condition for thermodynamic equilibrium. θe is the first excess quantity for the TL. As the position of any of the three D-faces can be chosen arbitrarily, the position of the TL and the value for θe are arbitrary too. Fortunately, this vagueness is of no serious consequence as long as the two-phase and three-phase boundaries possess microscopic thicknesses. Note that θe is a local property of the given TL in thermodynamic equilibrium. Hence, for isotropic phases, θe possesses one and only one value!

DLV Vapour Liquid qe DSV DSV

T Solid

Fig. 4 Sketch of the cross section through a sessile drop of liquid L on a flat solid S, both immersed in their common vapor atmosphere V. The D-faces DLV, DSV, and DSL describe the corresponding phase boundaries. T marks the triple line as the new model element for the three-phase boundary ) between the phases S, L, and V. θe = contact angle in thermodynamic equilibrium region (

4

Wetting of Solids

85

! The pressure tensor ! p will also vary in the three-phase region as compared with the two-phase boundary. Hence, a line tension κSLV has to be introduced as the corresponding excess quantity for the TL. Although the mathematical derivation is omitted here, κSLV follows from the same physical arguments that resulted in the definition of the interfacial tension in Sect. 2. It acts along the tangent of any point of the TL and has the dimension of a force as it refers to a tension in a 1D space. By analogy with a D-face, new excess quantities have to be introduced for all extensive thermodynamic quantities with respect to the TL. Details are not given here. For brevity, only a result of consideration of the Helmholtz free energy for a three-phase system in thermodynamic equilibrium under the influence of gravity is mentioned. It leads to a condition for the mechanical equilibrium at the TL on a solid with an arbitrary surface profile (roughness):

γ SV ¼ γ SL þ γ LV  cos θe þ

κSLV dκSLV þ R0 dR0

  j cos φS j

(28)

γ IJ = interfacial tension of the DIJ -face far from the TL R0 = radius of curvature in the considered point of the TL φs = angle of inclination of the extrapolated D SV-face in the considered point of the TL. Accordingly, θe does not only depend on the three interfacial tensions but also on the line tension and on the curvature of the TL. κSLV is of the order of 1010 N for solids in general. Nevertheless, it deserves attention when the phases are soft and one of the contacting phases becomes very small. Figure 5 illustrates the situation for a sessile liquid drop on a rough homogeneous solid surface. θe is defined as before. However, the true value of θe cannot be obtained from the plane optical projection of the sessile drop anymore as the experimenter cannot look into the local surface profile in order to get the local inclination of the tangent on the DSL-face. Fig. 5 Schematic cross section through a sessile liquid drop on a rough solid, both immersed in their common vapor. Now, the tangent on the extrapolated DSL-face at a selected point of the TL is inclined according to the solid surface profile. θe is defined as before

DLV Vapour Liquid qe

js

Ds L

Solid Tangent at DSL

DsV

86

M. E. R. Shanahan and W. Possart

Fig. 6 Schematic cross section through a sharp solid edge with the constant local θe of a moving TL (liquid meniscus not shown)

q

q

q

q Solid

Experimentally, it seems as if the contact angle varies substantially whenever the TL moves over a sharp edge. Figure 6 shows that this is just an optical illusion because the experimenter cannot resolve the local microscopic curvature of the edge. At a very sharp, angle on the solid, the contact angle may be considered to become indeterminate, an effect known as canthotaxis. However, this proposition assumes continuum mechanics, which clearly breaks down at this scale.

4.4

Experimental Aspects of Static Wetting

In Sects. 2 and 3 the concept of interfacial tension and interfacial free energy was defined, and their importance in wetting was noted, without going into details. For TLs with large radii of curvature (R0 >106 m), (28) can be approximated by γ SV ¼ γ SL þ γ LV  cos θe

(29)

The thermodynamic contact angle is still a local excess property of the TL. Treating the γ IJ as tensions (i.e., with dimensions of force/distance or mass/time 2 in some old literature!), as opposed to energies (i.e., dimensions of energy/area), in Fig. 1a balance of tensions at the triple line may be considered. This results directly in (29). This classic equation is usually referred to as Young’s equation, although Thomas Young never, in fact, wrote the expression in his often-cited paper (Young 1805)! If it is preferred to treat the γ IJ as interfacial free energies, and apply a simple argument of “virtual work,” the same relation is obtained. Having applied a force balance horizontally, this must be possible vertically, so what “resists” γ LV sin θe? Strictly, the action of tensions at the TL, specifically γ LV, deforms the solid surface, but it is only under conditions in which the solid is very soft that the effect is noticeable. The slight “bump” on the solid in Fig. 4 has a height given approximately by γ LV  sin θe/G, where G represents the (elastic) shear modulus of the solid. Simple insertion of typical values shows that the effect becomes noticeable only for G < 1 MPa (Shanahan and de Gennes 1986). As an example, aluminum has a value of G of ca. 25 GPa, and for water γ LV is ca. 70 mN m1. Taking sinθe as 1, it is found that the “wetting ridge” height is

4

Wetting of Solids

87

ca. 31012 m: well below atomic dimensions and thus meaningless in a continuum approach! In general, the effect perpendicular to the solid is negligible. However, the wetting ridge can be significant on soft solids and even modify wetting kinetics (Shanahan and Carré 1995). In (29), it may be noted that each surface tension has two suffixes representing two of the three, S, L, and V, although the use of γ LV = γ L or γ SV = γ S often found in the literature. In the case of the solid, γ SV changes a lot when it shares a common vapor phase with a liquid because the significant part of the surrounding vapor is due to the liquid. In the above thermodynamic treatment, allowance is already made for this effect. However, many workers prefer to adopt the alternative notion of equilibrium spreading pressure, π e = γ S  γ SV, where γ SV means the solid surface tension in presence of a liquid now. π e is often considered negligible for low activity solids, such as polymers. (Figure 15 demonstrates that this assumption should be checked!) The importance of (29) is that it may, in principle, allow one to estimate the surface tension of a solid, γ SV, albeit indirectly, from knowledge of the surface tensions, γ LV, of “probe” liquids and their contact angles on the solid surface in question. Knowledge of γ SL is required too, and the problem is considered in Sect. 5 and chapter Thermodynamics of Adhesion. Whereas liquid surface tension is readily measured (see below), that of a solid is inaccessible directly from experiment due to the intrinsic lack of molecular mobility. Firstly, liquid tension will be considered.

4.4.1

Liquid Surface Tension

Maximum Bubble Pressure If a tube of radius r is immersed vertically in a liquid to a depth h, the local pressure will be ρgh, where ρ is liquid density (more precisely, the difference between liquid density and that of the local vapor environment, but the difference is generally negligible) and g is gravitational acceleration. When air is blown through the tube, it will cause bubbles to form at the immersed end of the tube. Laplace’s equation (17b) relating the excess pressure on the curved side of the meniscus, Δp, to the principal radii of curvature R1 and R2 (both equal to R for a spherical surface) is given by

Δp ¼ γ LV

1 1 þ R1 R2

 ¼

2γ LV R

(30)

The internal bubble pressure is therefore Δp + ρgh. When the bubble first starts to grow, it will have a large value of R, which will then decrease until the bubble becomes a hemisphere. Thereafter, the radius grows again. Thus, when the bubble is a hemisphere, of radius r, the pressure is maximal. Thus maximal bubble pressure, pmax, is given by pmax ¼

2γ þ ρgh R

(31)

88

M. E. R. Shanahan and W. Possart

Fig. 7 Schematic representation of maximum bubble pressure method

Δp

h R

If the air line is fitted to a pressure gauge, the value of pmax may be measured and the value of liquid surface tension, γ LV, then readily calculated (Fig. 7).

Drop Weight An appealingly simple, but rather inaccurate, method to obtain liquid surface tension values is that of letting a drop accumulate at the end of a circular orifice of radius R, mounted horizontally like a tap. When sufficiently large, the drop will drop. On weighing the fallen liquid, a pseudo-equilibrium may be drawn between drop weight at separation, mg, and the hitherto surface tension-based retaining force, 2πRγ LV. Since some liquid is generally left behind, and since it is not entirely clear that the correct surface tension term to be used is that given above, generally a correction factor, φ, is invoked where typically 1 < φ e), it may be expected that the upward meniscus force will be given by F = 4πRγ LV (see Wilhelmy Plate). However, spurious effects occur (interference across the ring diameter?) leading to the adoption of an empirical correction factor, β. Liquid surface tension is then given by γ LV ¼

βF 4πR

(34)

where β is a function of R (and e). Table 1 gives a few examples of β as a function of R and e.

4.4.2

Contact Angle

Contact angle is the angle subtended, in the liquid, between the tangent to the liquid/ vapor interface and the solid/liquid interface at the triple, or three-phase, line, in the plane perpendicular to this line. In Sect. 3, it was shown from thermodynamics that the contact angle at thermodynamic equilibrium, θe, is a function of many quantities: the interfacial tensions of the three D-faces, line tension, the curvature of the triple line, and surface profile (cf. 28). For real solids in wetting experiments, the list of influences can be extended by surface heterogeneity in terms of locally varying interfacial tension, by adsorption layers, and by other “imperfections.” As a result, measured static contact angles always show some hysteresis Δθ. The so-called static advancing contact angle θe,a measured after the triple line has advanced onto a previously unwetted surface area is always larger than the so-called static receding angle, θe,r, found after the triple line has receded onto a solid surface region that was previously in contact with the liquid Δθ ¼ θe, a  θe, r > 0

(35)

Static contact angle hysteresis summarizes the difference in wetting of real surfaces and of ideally smooth homogeneous solids. The hysteresis can be even larger under dynamic measuring conditions. The static contact angle hysteresis does not depend on the experimental setup used for contact angle measurement provided the same solid and the same liquid are used and equilibrium is attained. As a consequence, the simplified equation (29) cannot be applied before the thermodynamic contact angle for the corresponding smooth solid has been deduced from θe,a and θe,r. It should be mentioned that the contact angle hysteresis is not a consequence of (28) only. Moreover, gravity causes some kind of stick–slip behavior that depends Table 1 Two examples of β for different values of R and e for the du Noüy ring

R (cm) 1 1.85

e (cm) 0.02 0.03

β 0.93 0.88

92

M. E. R. Shanahan and W. Possart

on the slope of the roughness asperities when the three-phase boundary moves over the rough solid surface. It can be shown that static contact angle hysteresis depends upon this effect (e.g., Bartell and Shepard 1953a, b; Johnson and Dettre 1964; Busscher et al. 1984; Bischof et al. 1988). Various propositions, such as cos θmeas ¼ r  cos θe ;

r¼

Atrue Ageom

(36)

(Wenzel 1936) Ageom = macroscopic area of the solid surface, i.e., width by length Atrue = microscopic area due to surface roughness within Ageom. θ e ¼ θ e, a

(37)

 1 θ e, a þ θ e, r 2

(38)

(Thiessen and Schoon 1940) θe ¼

(Adam and Elliott 1962) have been proposed in the literature to tackle this problem. However, inspection of experimental data reveals that none of these formulae solves the problem totally and satisfactorily – cf. Fig. 12 as an example. Instead, it is clear from experimental results that the functions θe, a(Δθ) and θe, r(Δθ) follow straight lines θe, a ¼ θe þ a  Δθ

(39)

θe, r ¼ θe  b  Δθ

(40)

As roughness is a major cause of Δθ on homogeneous solids, the intersection of these lines with the ordinate is the value θe being sought (Schulze et al. 1989). Hence, the equilibrium contact angle can, in principle, be determined from a suitable experiment. In the following sections on experimental methods for contact angle measurement, we shall summarize θe,a and θe,r as θ0, for brevity. Of course, both advancing and receding angles must be measured for a complete appraisal.

Goniometry Contact angles may be assessed by various techniques, of which a few of the better ones will be mentioned. The simplest involves a plane optical projection of the three interfaces and the triple line. The tangent to the projection of the liquid/vapor interface at the triple line is estimated and the angles measured relative to the flat

4

Wetting of Solids

93

qe,a

150 PTFE - water

0.5.(qe,a + qe,r)

100 q[°]

qe,r PMMA - formamide qe,a 50

qe,r 0

0

20

40

60

Δq[°]

Fig. 12 θe,a and θe,r data measured with various liquids on homogeneous polymer surfaces after stepwise increases of roughness following grinding with abrasive cloth of increasing grade

projection of the possibly rough solid surface. There is equipment available on the market to do this, using in the simpler cases an eyepiece with two, independent goniometers, one to align with the projection of the solid surface and one to form the tangent to the projection of the liquid/vapor interface. It is relatively easy to set up an optical bench arrangement, either to observe the drop directly or to project its magnified image onto a screen, and then to draw the requisite tangents. However, placing a tangent to a curve is always somewhat subjective. An alternative method is to use the following procedure. Firstly, when posing a drop on the flat, horizontal surface, ensure from the top that its contact area is circular. (This may be done with a transparent slide, onto which are printed circles of various diameters. By adjustment of drop-to-transparent distance, it is possible to find a position at which a circle should coincide with drop circumference.) Any deviation of the triple line from a circle is a clear proof that the solid surface is either rough or inhomogeneous. The surface tension, γ LV, of the liquid (necessary anyway, to exploit (29)) and also its density, ρ are required. Large drops tend to flatten (“puddles”), whereas small drops stay as spherical caps (“dew drops”). The approximate contact radius, r  a, below which the drop is sufficiently close to sphericity, is given by the liquid’s capillary length: a = (γ LV/ρ g)1/2 where g is gravitational acceleration. For water, for example, a is ca. 2.8 mm. Having checked circularity of the contact area and droplet size, θ0 can be readily estimated from the simple formula for a spherical cap:

94

M. E. R. Shanahan and W. Possart

 θ0 h tan ¼ r 2

(41)

where height, h, and radius are as shown in Fig. 1.

“Puddle” At the other extreme from “dew drops,” a “puddle” may form (see Fig. 13). If the drop radius, r, considerably exceeds a, gravity flattens the liquid bulk. If γ LV is known, e.g., by one of the above methods, then the contact angle may be calculated from a measurement of drop height (in the flat part): cos θ0 ¼ 1 

ρ g h2 2 γ LV

(42)

As for the pendent drop (above), there exists an intermediate size range between the “dew drop” and the “puddle” for which γ LV may be estimated from drop shape (slightly deformed from sphericity), but its treatment is beyond the scope of this chapter (Padday 1971; Hartland and Ramakrishnan 1975). Modern equipment on the market comes with software that approximates the projection of the liquid meniscus to a solution of the corresponding differential equations.

Capillary Rise and Wilhelmy Plate The principle here is the same as discussed above, the only differences being that the liquid surface tension is assumed to be known and that the contact angle is nonzero. If the θ0 are greater than zero, capillary rise in a tube is less marked, and may even become capillary descent, if θ0 > π/2. The liquid level in the tube is lower than the surrounding liquid bath. When the θ0 are nonzero, (33) includes a term in cosθ0, and, after rearrangement, the following expression is obtained: θ0 ¼ cos 1

R ρ g h 2 γ LV

(43)

In much the same way, the Wilhelmy plate technique may be used, where force, F, becomes 2 l γ LV cosθ0. The disadvantage for both of these is that the solid of interest must be made into a suitable geometric form, which is not always convenient or even possible!

Fibers The special case of fibers will be given, since knowledge of their surface properties is essential in composite technology. In principle, the method for obtaining contact

q0

h

Fig. 13 Schematic representation of “puddle” of height h and contact angle θ0

4

Wetting of Solids

95

angle on a fiber is the same as for the Wilhelmy plate: A liquid meniscus forms round the fiber immersed in a liquid bath, and its weight, detected by a microbalance, can be used to obtain θ0. It is assumed, for simplicity, that the fiber is of circular section of radius r. With F0 being the force on the balance corresponding to the weight of the fiber before immersion, and F that measured with the fiber partially immersed, vertically, axis perpendicular to free liquid surface and to depth, l, an expression for the difference, ΔF, is obtained given by ΔF ¼ F  F 0 ¼ 2 π r γ LV cos θ0  π r2 l ρL g

(44)

In this equation, ρL is liquid density (or more precisely, the difference between density of liquid and air), and thus the right-hand term is a buoyancy correction. Since the surface tension term scales with r, and the buoyancy term with r2, clearly the correction becomes very small for thin fibers. (The same basic argument is used when employing a thin Wilhelmy plate, above.)

4.5

Free Energy Balance of Wetting

With the thermodynamic background (cf. Sects. 2 and 3) and with the experimental steps for getting θe,a, θe,r and the thermodynamic contact angle for the smooth homogeneous solid, θe, (Sect. 4), the picture can now be completed by discussing the excess Helmholtz free energy of wetting which is often called the Helmholtz free energy of adhesion. Wetting is indeed the adhesion of a liquid on a solid. Figure 14 illustrates the situation before and after wetting at thermodynamic equilibrium. For simplicity, consider a solid with at least one smooth and flat part of its surface. In the initial state (superscript 1), the solid has a phase boundary with the common vapor only. Hence, there are two D-faces with the areas A1SV , A1LV in the initial state. The solid is lowered until the flat part of its surface is just isothermally wetted at constant pressure by the flat liquid meniscus. This is the final state (superscript 2) with three D-faces (areas A2SV , A2LV , A2SL and one triple line. (Any immersion of the solid into the liquid would simply create new area of the D-faces.

Vapor Solid

2

Vapor

ASV

1

ASV

Solid

1

ALV

2

ALV

2

ASL Liquid

Liquid

Fig. 14 Schematic representation of a solid in vapor (left) being led to contact a liquid surface (right). A1IJ , A2IJ denote the areas of the corresponding D-faces

96

M. E. R. Shanahan and W. Possart

Such motion only complicates the thermodynamic free energy balance without giving any additional insight.) Note that the constitution of the system, i.e., the components, their number N, their concentrations, and chemical potentials μi as well as pressure and temperature is the same in the initial and the final state. Hence, the state of the bulk phases does not change. Moreover, the interfacial tensions γ SV and γ LV and the 2 N values for the absolute adsorptions Γ i, SV, Γ i, LV will not change either. Now (26) is employed and the excess Helmholtz free energies of all five D-faces may be written F 1SV ¼ γ 1SV A1SV þ

N X

μi Γ i, SV A1SV

(45a)

μi Γ i, LV A1LV

(45b)

μi Γ i, SV A2SV

(45c)

μi Γ i, LV A2LV

(45d)

μi Γ i, SL ASL

(45e)

i

F 1LV ¼ γ LV A1LV þ

N X i

F 2SV ¼ γ SV A2SV þ

N X i

F 2LV ¼ γ LV A2LV þ

N X i

F 2SL ¼ γ SL ASL þ

N X i

where Γ i, IJ = absolute adsorption of component i on the D-face IJ. A relationship between the areas of the D-faces exists ASL ¼ A1LV  A2LV ¼ A1SV  A2SV

(46)

Therefore, wetting is considered to be an isothermal, isochoric, and isobaric change of state, and the contribution from the triple line is neglected. As a result, the change of Helmholtz free energy of the whole system is given by the change of the excess Helmholtz free energy at the D-faces F 2  F 1 ¼ F 2SV  F 1SV þ F 2LV  F 1LV þ F 2SL  F wett

(47)

That is the excess Helmholtz free energy of wetting. Inserting (Eqs. 45 and 46), the following equation is obtained F wett ¼ ðγ SV þ γ LV  γ SL Þ ASL  ASL 

N X i

  μi  Γ i, SV þ Γ i, LV  Γ i, SL

(48)

4

Wetting of Solids

97

and for the specific value (per unit area) f wett ¼

N X   F wett ¼ ðγ SV þ γ LV  γ SL Þ  μi  Γ i, SV þ Γ i, LV  Γ i, SL ASL i

(49)

which is usually called the adhesion energy for the liquid–solid contact. The first term on the right-hand side corresponds to the so-called reversible specific or thermodynamic, work of adhesion (to within a sign change) W adh ¼ γ LV  γ SL Dupre0 s rule

(50)

as introduced by Dupré (1869). Note that Dupré formulated his rule on an empirical basis long before Gibbs developed the thermodynamics for interfaces in the form used here. However, there is a second term that covers the energetic contribution from the composition of the three-phase boundaries (via the absolute adsorptions), and hence the specific Helmholtz free energy of wetting may not be reduced to a balance of the interfacial tensions involved. A simple experimental test can be used to demonstrate that this is more than just a theoretical statement. Figure 15 depicts the change of the wetting equilibrium of polymer samples at a Wilhelmy balance (cf. Fig. 10) owing to a slight compositional change of the common vapor phase. The sample plate at the fine balance consists of either polymer A or polymer B, the test liquid is water, and there is the common gas phase in the compartment (Fig. 15, left scheme). At the starting point t = 0 min (Fig. 15, plot on the right), the plate just touches the water surface for wetting. The capillary in the setup is still empty so that the common gas phase consists of air and water vapor only. For that equilibrium, the wetting causes a certain force change Fmeas(t = 0) as compared to the dry polymer sample A or B. As the next step (t > 0), only the vapor phase is contaminated by a small amount of n-octane evaporating from a pendent drop hanging from the glass capillary now (cf. Fig. 15, left). n-Octane is “immiscible” with water. (N.B. According to thermodynamics, all substances mix, at least to some extent. Hence, the word “immiscible” corresponds to common usage.) As a saturated hydrocarbon, n-octane forms only weak physical interactions with water and the polymers. Nevertheless, the wetting force Fmeas(t) at the Wilhelmy balance changes immediately and significantly whenever a small amount of n-octane vapor is added and then removed! This is due to a change of the wetting equilibrium which in turn is determined by the three D-faces LV, SV, and SL. In fact, this is the type of experiment that may be used to introduce the convenient “spreading pressure,” representing reduction in surface tension due to adsorption, as mentioned earlier. For wetting, the extra force acting on the Wilhelmy balance is F  F 0 ¼ L  ðγ SV  γ SL Þ  buoyancy of the sample

(51)

98

M. E. R. Shanahan and W. Possart

Fmeas [p]

Balance *

Capillary

0,281

0,238 n-octane added 0,237

0,28 Polymer A

Sample plate

Air C8H18

H2O

0,236

0,279

0,235

0,278 Polymer B

0,234 n-octane exhausted

0,233 0,232

0

5

10

0,277 0,276

20 15 t [min]

25

0,275 30

Fig. 15 Sketch (left) for the Wilhelmy balance with sample plate and a capillary for n-octane evaporation into the common gas phase (air) above the wetting liquid (water). Fmeas = load change in gram force p as measured at the balance. For two polymers A and B (different types of cellulose), the right-hand graph shows the quick response of the wetting equilibrium when a little amount of octane is evaporated into the gas phase at t > 0 min and when the atmosphere is replaced by fresh air after some time (return of Fmeas to the initial value)

F0 = force due to the weight of the dry polymer sample before wetting L = length of the TL on the wetted polymer sample. The sample immersion depth does not change when n-octane is added to the composition of the vapor phase, and hence the measured change of force is ΔF ¼ L  ðΔγ SV  Δγ SL Þ ¼ L  ½Δγ LV  cos θ0 þ γ LV  Δð cos θ0 Þ

(52)

Since Δγ SL > 0, according to (26), the change of the wetting equilibrium is accompanied by a change of the specific excess Helmholtz free energy of each D-face: D Δf D IJ ¼ Δγ IJ þ

X

D Δμi  Γ D i, IJ þ μi  ΔΓ i, IJ

 (53)

i

Therefore, the addition of n-octane to the vapor phase not only changes the interfacial tensions but also the chemical potential (via the compositional change) and the absolute adsorptions (mainly for the LV and SV interfaces). As the consequence, an error results from utilizing Dupré’s rule (50) instead of the full expression (49) for the specific Helmholtz free energy of wetting. The magnitude of the error cannot be deduced from wetting experiments since they do not provide information on adsorption at the D-faces in the system.

4

Wetting of Solids

4.6

99

Concluding Remarks

It may be seen from the preceding sections that a variety of methods are available to obtain both liquid surface tensions and static contact angles. Some are cheap, or easily constructed, others less so. They all have their various merits, depending on the (geometric) presentation of the solid to be characterized. They also have their respective disadvantages, e.g., the sessile drop technique is simple, but care must be exercised in assuring correct drop size range and circularity of the triple line; the maximum bubble pressure is precise and capable of measurement of transient surface tensions but generally requires relatively advanced apparatus. All techniques for contact angle measurement are prone to static wetting hysteresis effects. Wetting hysteresis has its origins in surface roughness, microheterogeneities, and other “imperfections” and can be of the order of 10 s of degrees. It is rarely less than 2
or 3
, so the commonly seen habit of specifying contact angles to one, sometimes two, decimal places is pointless. This leaves another question: If static advancing and receding contact angles are measured, what is the equilibrium value θe on a flat smooth homogeneous surface? It is θe that is needed for Young’s equation. The experimental procedure described in the context of (39 and 40) provides the answer. Bearing in mind the limitations and approximations made for Dupré’s rule (50) and Young’s equation (29), finally consider the common treatment of contact angle data which makes use of the combination of the two equations: W adh ¼ γ LV ð1 þ cos θe Þ  f wett

(54)

This is the estimated specific wetting energy, provided that the surface tension γ LV of the wetting liquid is measured using one of the techniques mentioned and that equilibrium contact angle for the flat surface θe is derived correctly from the contact angle hysteresis. Nevertheless, it is unclear whether it is justified to neglect the adsorption terms in the energy balance. Additional conditions needed for such interpretation of wetting data are as follows: Pure liquids that neither swell nor dissolve the solid must be used. Solids should be “ideal” (flat, homogeneous, etc.). Small curvature of the triple line must be ensured. Interfacial tension is an excess property of the phase boundary of two phases and responds remarkably to minute changes of the composition! The use of literature data for γ LV of liquids should be avoided. Conditions of use can vary. Finally, there is another price to be paid for using (54): The interfacial tensions γ SL and γ SV of the solid have been eliminated, and hence, they cannot be deduced from wetting experiments. This obstacle was the reason for the development of so-called thermodynamic adhesion theories which introduce various proposals for a

100

M. E. R. Shanahan and W. Possart

relationship between the liquid–solid interfacial tension γ SL and the surface tensions γ LV and γ SV. These theories are considered in chapter Thermodynamics of Adhesion.

References Adam NK, Elliott GEP (1962) Contact angles of water against saturated hydrocarbons. Theory of the capillary layer between the homogeneous phases of liquid and vapour II. J Chem Soc. 2206–2209. https://doi.org/10.1039/JR9620002206 Bakker G (1902) Theory of the capillary layer between the homogeneous phases of liquid and vapour II. Z Phys Chem Stoichiom Verwandtschafts 42:68 Bartell FE, Shepard JW (1953a) Surface roughness as related to hysteresis of contact angles I. The system paraffin-water-air. J Phys Chem 57:211 Bartell FE, Shepard JW (1953b) Surface roughness as related to hysteresis of contact angles II. The system of paraffin-4M-calcium chloride solution-air and paraffin-glyceral-air. J Phys Chem 57:455 Bischof C, Schulze RD, Possart W, Kamusewitz H (1988) The influence of the surface state of polymers on the determination of the contact angle. In: Allen KW (ed) Adhesion, vol 12. Elsevier Appl Sci, London/New York, pp 1–16 Bumstead HA, Van Name RG (eds) (1961) Scientific papers of J Willard Gibbs, vol 2. Dover, New York Busscher HJ, van Pelt AWJ, de Boer P, de Jong HP, Arends J (1984) The effect of surface roughening of polymers on measured contact angles of liquids. J Coll Surf 9:319 Dupré A (1869) Théorie mécanique de la chaleur. Gauthiers-Villars, Paris, p 369 Gaydos J, Rotenberg Y, Boruvka L, Chen P, Neumann AW (1996) The generalized theory of capillarity. In: Neumann AW, Spelt JK (eds) Applied surface thermodynamics. Marcel Dekker, New York, pp 1–51 Hartland S, Ramakrishnan S (1975) Determination of contact angles and interfacial-tension from shape of sessile interfaces. Chimia 29:314 Johnson RE, Dettre RH (1964) Contact angle hysteresis. I. Study of an idealized rough surface. Adv Chem Ser 43:112 Padday JF (1971) The profiles of axially symmetric menisci. Philos Trans R Soc A 269:265–293 Rusanov AI (1978) Phasengleichgewichte und Grenzflächenerscheinungen. Akademie-Verlag, Berlin Rusanov AI, Prokhorov VA (1996) Interfacial tensiometry. In: Möbius D, Miller R (eds) Studies in interface science, vol 3. Elsevier, Amsterdam/Lausanne/New York/Oxford/Shannon/Tokyo Schulze RD, Possart W, Kamusewitz H, Bischof C (1989) Young’s equilibrium contact angle on rough solid surfaces – I. An empirical determination. J Adhes Sci Technol 3:39–48 Shanahan MER, Carré A (1995) Viscoelastic dissipation in wetting and adhesion phenomena. Langmuir 11:1396 Shanahan MER, de Gennes PG (1986) The ridge produced by a liquid near the triple line solid/ liquid/fluid. C R Acad Sci Paris 302:517 Shanahan MER, Cazeneuve C, Carré A, Schultz J (1982) Wetting criteria in 3-phase solid/liquid/ liquid systems. J Chim Phys 79:241 Thiessen PA, Schoon E (1940) Besetzung und Adhäsionsarbeit von Oberflächen fester organischer Verbindungen. Z Elektrochem 46:170 Wenzel RN (1936) Resistance of solid surfaces to wetting by water. Ind Eng Chem 28:988–994 Young T (1805) Cohesion of fluids. Philos Trans R Soc A 95:65–87 Zhang J, Sheng X, Jiang L (2009) The dewetting properties of lotus leaves. Langmuir 25:1371–1376

5

Spreading of Liquids on Substrates Günter Reiter

Contents 5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.2 Equilibrium Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.3 The Spreading Coefficient . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.4 The Flow Profile and Energy Dissipation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.5 The Dynamic Contact Angle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.6 The Rate of Spreading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.7 Dewetting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5.8 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

101 102 103 104 106 108 110 111 112

Abstract

A short overview of relevant processes and parameters in spreading of liquids on substrates is presented. In a simplified view, the dynamics of these processes can be understood as being controlled by the balance of driving forces and resistance due to dissipative processes. Analogies between spreading and dewetting are discussed.

5.1

Introduction

Spreading of a liquid on a substrate represents an elementary step in many processes like adhesion, coating, or painting. Spreading is the dynamic route of covering a substrate with a liquid up to the point when an equilibrium state is reached (Dussan 1979; de Gennes 1985; Cazabat 1987; Good 1992; Leger and Joanny 1992; Blake 2006; De Coninck and Blake 2008; Bonn et al. 2009; Marmur 2009). Many practical G. Reiter (*) Physikalisches Institut, Fakultät für Mathematik und Physik, Albert-Ludwigs Universität Freiburg, Freiburg, Germany e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_5

101

102

G. Reiter

Dewetting Spreading

Fig. 1 Schematic side-view representation of spreading of a drop toward a thin film (left) and dewetting of an unstable thin film via nucleation and growth of a dry patch surrounded by a rim (cross-sectional view) collecting the removed liquid

applications require the knowledge of how fast a liquid droplet, once deposited on a substrate, can spread out and cover a given area of the substrate. Because the complementary process to spreading, that is, the retraction or dewetting of a liquid from a substrate (see Fig. 1), is determined by the same intermolecular forces and the corresponding dissipation routes, both processes can be treated in analogous ways (de Gennes 1985; Brochard-Wyart and de Gennes 1992). Of course, the properties of the substrate represent equally relevant parameters as can be demonstrated by modifying the surface chemistry by standard chemical means (e.g., silanization, plasma treatment) (Israelachvili 1995). The process of liquid spreading is determined by short-ranged surface forces mainly controlled by surface chemistry but also by forces acting over longer distances up to some tens of nanometers (long-ranged forces). Moreover, chemical and physical heterogeneities are responsible for hysteretic behavior (Dussan 1979), which can significantly affect spreading and dewetting up to the point of prohibiting it. Generally, all spreading phenomena are governed by the interplay of surface and interfacial interactions, in particular in the region where the liquid meets the yet uncovered solid, the region of the three-phase contact line (de Gennes 1985).

5.2

Equilibrium Conditions

As stated already in 1805 by Thomas Young (1805), at thermodynamic equilibrium (no temperature gradients, concentration gradients, i.e., no net forces acting) the three surfaces that meet at the three-phase contact line define an equilibrium contact angle (θeq). Using a vector description for the forces (per unit length of the contact line), which are balanced at their contact, it can be seen that the surface tension (γ SV) between substrate and the surrounding medium (which is often the vapor of the liquid in question) has to be equal to the interfacial tension (γ SL) between substrate and liquid plus the projection of the interfacial tension (γ LV) between liquid and the surrounding medium onto the γ SL-vector, resulting in the Young-equation: γ SV ¼ γ SL þ γ LV cos θeq

(1)

5

Spreading of Liquids on Substrates

103

g LV

gSV

qeq

gSL

Fig. 2 Schematic representation of the balance of interfacial forces (represented by the arrows for γ SV, γ LV, and γ SL, which meet at the contact line (indicated by a small circle)). This force balance determines the equilibrium contact angle θeq

Using Young’s Eq. 1 and knowing γ SV and γ LV, γ SL may be obtained by measuring the contact angle θeq (see Fig. 2) (Zisman 1964). It should be noted that the equilibrium contact angle is a direct consequence of the balance of molecular interactions among the three materials at the contact line. At equilibrium, the contact line will not be set in motion and a drop will never spread. On the contrary, spreading occurs for a closed nonequilibrated system that tries to reach equilibrium. Alternatively, a liquid may be forced to spread under the influence of an external force. From a mechanical point of view, the balance of forces within the plane of the substrate neglects the force component γ LV sin θeq in the direction normal to the substrate. Usually one assumes that the substrate is completely rigid and not deformable. Thus, this vertical force component is balanced by a reaction force of the solid, which is equal to γ LV sin θeq. Furthermore, it is assumed that the solid is not generating a force component parallel to the surface when getting in contact with a liquid. In such a situation, the forces acting at the contact line could be balanced by any value of the contact angle (Dussan 1979). On liquid or deformable (e.g., thin membranes or elastomers) substrates, however, this normal force component leads to a vertical displacement of the contact line, that is, the formation of a wetting ridge (Shanahan and Carré 1995; Carré et al. 1996). The Neumann triangle construction can account for this case (Chen et al. 1996).

5.3

The Spreading Coefficient

Based on these three interfacial tensions, a spreading coefficient Seq may be defined, which characterizes the tendency of a liquid to spread out (or not) on a substrate: Seq ¼ γ SV  ðγ SL þ γ LV Þ

(2)

Using Eq. 1, this relation may be rewritten as:
 Seq ¼ γ LV cos θeq  1

(3)

104

G. Reiter

As pointed out by de Gennes (1985), in equilibrium, γ SV can never be larger than γ SL + γ LV. If it were, the total free energy of the system could be lowered by removing the solid–vapor interface, that is, by coating the substrate with a (thick) layer of liquid. Then only a solid–liquid and a liquid–vapor interface would remain, that is, the solid is wetted (coated) by the liquid. In other words, the “true” equilibrium value of γ SV is identical to γ SL + γ LV. However, if the solid to be coated is initially not in equilibrium with the surrounding vapor, the initial spreading coefficient Sin may be larger than zero. In particular for nonvolatile liquids (these cannot reach equilibrium by evaporation and condensation but only by diffusion on the surface of the substrate) Sin can be defined by referring to the interfacial tension SSO between the dry substrate and vacuum. Sin ¼ SSO  ðγ SL þ γ LV Þ

(4)

In equilibrium, some molecules from the liquid will be adsorbed on the substrate, lowering γ SO to γ SV. Consequently, Sin is larger than Seq and accordingly, after deposing a drop of liquid onto a dry substrate, flow will start (the drop spreads) until the equilibrium value of the contact angle is reached. θeq may well be zero representing the situation where the liquid will cover the substrate by a film of homogenous thickness.)

5.4

The Flow Profile and Energy Dissipation

In spreading, the whole liquid wedge close to the contact line moves with the velocity U of the contact line. A standard boundary condition for liquid flow on solid substrates is typically the no-slip at the substrate. Liquid molecules at a solid substrate experience (significant) interactions responsible for adhesion. Thus, in contrast to liquid molecules in contact with other liquid molecules, they are not moving easily on this substrate. As a consequence, the velocity of these molecules is approximated by zero velocity (= no-slip boundary condition). However, applying this boundary condition strictly would not allow for any displacement of a liquid on a solid substrate. Thus, to account for the experimental fact that contact lines can and do move, various assumptions, partially based on molecular mechanisms, have been proposed (see, e.g., Bonn et al. 2009). In general, a cutoff length, typically as small as the size of the molecules, has to be introduced (de Gennes 1985; Huh and Scriven 1971). For small contact angles (allowing to approximate sinθeq  θeq [in radians]), the flow pattern close to the contact line (the wedge region) can be approximated by the so-called lubrication approximation, which compares the flow in the wedge with flow in a thin film: the velocity of fluid molecules is only determined by their distance z to the substrate, that is, the molecules move preferentially in the xdirection parallel to the substrate and less in the direction away from the substrate.

5

Spreading of Liquids on Substrates

105

Thus, at any point x from the contact line the velocity v(z) is only a function of z[vx(z) = v(z); see Fig. 3]. As interactions between liquid molecules, and thus their viscosity η, are assumed to be independent of the distance z to the substrate also the relative velocity difference between molecules should be constant. Thus, in a film of local thickness h(x) that moves at the mean velocity U of the contact line, a decrease in velocity is found when going from the film surface toward the substrate with a velocity gradient dv(z)/dz inside the film. Thus, U is the average along the normal to the surface, that is, the z-direction, of this velocity profile U¼h

1

ðh

dz vðzÞ

(5)

0

Accordingly, a parabolic velocity profile is obtained (de Gennes 1985) (with the boundary conditions that the velocity vanishes at the solid side and the derivative of the velocity at the liquid/gas interface is zero, that is, there is no tangential stress): vðzÞ ¼

 3U
2 z þ 2hðxÞz 2 2h ðxÞ

(6)

and the rate of viscous dissipation (within one slice of the film) due to viscous friction of the liquid molecules that are moving past each other, integrated over the film depth h(x): ð h ð xÞ 0

  dvx ðzÞ 2 3ηU 2 dz η ¼ dz hð x Þ

(7)

Integration now over the whole moving object, from its lowest thickness at the contact line having a minimum thickness of hmin(x) to its thickest part with the maximum thickness of hmax(x), using the approximation h(x)  x  θdyn (with θdyn Fig. 3 Schematic representation of the velocity profile of liquid flow within a thin film (lubrication approximation). For the no-slip boundary condition (broken line) the velocity at the substrate (z = 0) is zero. In the case of slippage (full line), the liquid velocity is not zero at the substrate surface and only extrapolates to zero at a distance b (= slippage length) within the substrate

z v(z)

x b

106

G. Reiter

being the dynamic contact angle at the contact line) yields the total viscous dissipation per unit time and unit length of the contact line during spreading (Bonn et al. 2009): Ddiss ffi

ð hmax hmin

3ηU2 3ηU 2 hmax dx ¼ ln hð x Þ θdyn hmin

(8)

Clearly, if hmin would be zero the dissipated energy would be infinite. However, such a value is not physical because the smallest lengthscale in the problem is related to the moving molecules. Thus, it is plausible to set the lower cutoff length proportional to the size of the molecules. Basically, the same procedure can be applied to the case where the liquid is slipping on the substrate by introducing an extrapolation length (slippage length) b within the substrate (see Fig. 3) where the flow velocity would be zero, that is, the nonslip boundary would be reached at a virtual interface inside the solid. Thus, one expects hmin ~ b (de Gennes 1985). Slippage is particularly relevant in the case of entangled polymer melts spreading or dewetting on a smooth surface (de Gennes 1985; Reiter and Khanna 2000a, b). A large number of experimental studies on slippage have been recently reviewed (Neto et al. 2005).

5.5

The Dynamic Contact Angle

While a well-defined equilibrium situation allows a clear characterization of θeq, the displacement of the contact line (which evidently implies a velocity of displacement) requires a deviation from this value, resulting in a dynamic contact angle θdyn (see Fig. 4). In fact, whenever on ideal surfaces the contact angle deviates from θeq (which represents the absolute minimum in free energy of the system) the contact line will show the tendency to move in order to try to reestablish equilibrium. In

qdyn qeq qeq Advancing liquid

qdyn

Retracting liquid

Fig. 4 Schematic representation of the difference between the equilibrium contact angle θeq and the dynamic contact angle (θdyn). In equilibrium, the force vectors for γ SV, γ LV, and γ SL add up to zero within the plane of the substrate. However, in the case of an advancing (left) or retracting (right) liquid front a net force remains, indicated by the big (red) arrows on top

5

Spreading of Liquids on Substrates

107

general, the stronger θdyn deviates from θeq, the faster will be the movement of the contact line (see Fig. 5). Thus, for θdyn 6¼ θeq, instead of a perfect balance of interfacial tensions at equilibrium, a net force F (per unit length of the contact line) remains: F ¼ γ SV þ γ SL þ γ LV cos θdyn

(9)

Using Eq. 1, (γ SV + γ SL) can be replaced by (γ LV cos θdyn) and finally the so-called unbalanced Young force is obtained (Brochard-Wyart and de Gennes 1992)
 F ¼ γ LV cos θdyn  cos θeq

(10)

For small contact angles the cos-term can be approximated by the first term of its series expansion, that is, cos θ  1  12 θ2 . This leads then to   1 F ¼ γ LV θ2eq  θ2dyn 2

(11)

If the viscous forces are significantly smaller than the surface tension (as expressed by a capillary number Ca much less than 1: Ca = U  η/γ LV, with U being the velocity of the contact line and η the viscosity of the liquid), the shape of

q

qadvancing qreceding

Retraction

Spreading

0 Velocity of contact line displacement

Fig. 5 Qualitative representation of the variation of the contact angle as a function of the contact line velocity (a negative velocity represents retraction, a positive velocity represents spreading). On heterogeneous substrates, a difference between the values of the contact angle at zero velocity will be observed, depending on how zero velocity is approached, in the advancing (θadvancing) or receding (θreceding) direction. Contact angle hysteresis is defined as: Δθ = θadvancing  θreceding. In some cases, for example, when heterogeneity is caused by substrate roughness, the equilibrium contact angle may be approximated by θeq = ½ (θadvancing + θreceding). However, this last relation is not generally applicable

108

G. Reiter

the drop can be reasonably well approximated by a spherical cape. Laplace pressure PL(PL = 2γ LV  C, for a spherical cap of radius R the curvature C = 1/R and thus PL = 2γ LV/R) will be able to assure the same curvature everywhere, except for the region very close to the contact line, where long-range molecular interactions, for example, van der Waals forces, become important. At a macroscopic level, a spreading drop, for example, on a wettable substrate, can be well described by an apparent contact angle θap = 4V/πR3 (valid for small values of this contact angle): as the drop spreads, the apparent contact angle decreases as 1/R3.

5.6

The Rate of Spreading

The speed of spreading is determined by the balance of driving forces (corresponding to surface energy, represented by the contact angle, and, for big drops, gravity) and velocity-dependent dissipative forces. In many cases, the major contribution to dissipation originates from the velocity gradient experienced by the liquid molecules moving in the wedge region close to the contact line. The diverging increase of dissipation as the contact line is approached (see Eq. 8) represents the most important mathematical problem in the context of moving a three-phase contact line, as first pointed out by Huh and Scriven (1971). For the complete wetting case (S  0), a thin film of dynamical origin (off-equilibrium) forms ahead of the apparent contact line of the drop. The formation of a precursor film avoids that viscous dissipation diverges near the (geometrically apparent) contact line. Thus, macroscopic spreading of a droplet of a nonvolatile liquid (with surface tension γ LV), which wets the substrate, proceeds via the formation of such a thin precursor film ahead of the “visible” (apparent) contact line (Checco 2009). The thickness of this precursor film is truncated at a value given by: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi h0 ¼ a 2γ LV =2Sin (12) pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi with the molecular length scale a ¼ 2A=γ LV (where A is the Hamaker constant for the liquid on the substrate). h0 is usually larger than the molecular size a, because typically γ  Sin. In the equilibrium situation (after spreading being completed), h0 represents the final thickness of the so-called final pancake, that is, the resulting wetting film, which can be easily derived from the free energy F of an initial droplet of volume Ω (Brochard-Wyart et al. 1991): F ¼ F 0  A  S in þ PðhÞ  A with the constraint that Ω ¼ A  h: F 0 is the free energy of the system without droplet, A is the area covered by the droplet, Sin is the initial spreading coefficient, and P(h) is the contribution to the free energy that arises as the thickness is less than infinity. Thus, the limiting values of P(h) are: P(h ! 1) = 0 and P(h ! 0) = Sin = γ SO  γ SL  γ LV. For cases where van der Waals forces dominate the long-range interaction, P(h) = A/12πh2 is obtained. Minimizapffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi tion of F with respect to h leads to Eq. 12 h0 ¼ A=4πSin ¼ a 3γ LV =2Sin .

5

Spreading of Liquids on Substrates

109

A typical spreading experiment starts by depositing at time t = 0 a spherical liquid droplet of volume V, characterized by the base radius R0, onto an ideal, planar, and horizontal substrate, either solid or liquid. Suppose that the liquid is nonvolatile and that the initial shape of the now sessile droplet is part of a sphere with a macroscopic contact angle θ0 (for simplicity, the case of θ0 π/2 is assumed). This droplet is surrounded by a nonreactive and nonmiscible fluid phase (either gas or liquid). When S  0, the droplet spreads spontaneously and tends to separate solid and surrounding fluid phase. For a finite volume of the initial droplet, spreading continues only until the liquid layer, covering the solid surface reaches a mesoscopic or microscopic thickness controlled by the disjoining pressure (originating from longrange forces like van der Waals forces), which in most cases tends to thicken this layer, and which effectively competes with the capillary forces. Thus, the final thickness of the extended liquid structure is determined by the interplay between the spreading power and the liquid/gas interfacial tension (see Eq. 12). For negative values of the spreading coefficient (S < 0), the liquid droplet only partially wets the solid. If the liquid was forced to cover the substrate, it will retract (dewet) from it until one or many (macroscopic) liquid droplets with the equilibrium shape of a spherical cap have formed. If the liquid is deposited as a spherical droplet, it will spread out until it reaches its equilibrium shape, characterized by θeq. In describing spreading and the local molecular processes directly at the contact line, two different types of descriptions of sessile droplet spreading exist: hydrodynamic and molecular kinetic, which differ from each other mostly in the consideration of the dominant dissipation channel (de Gennes et al. 1990; Cazabat et al. 1997; Blake 2006). In the molecular kinetic theory of Eyring, adapted to describe the kinetics of wetting phenomena (Blake and Haynes 1969; Blake 2006; de Coninck and Blake 2008), the focus is put on the dissipative processes occurring in the vicinity of advancing contact line, mainly due to the attachment/detachment processes of molecules from the liquid onto the solid. This approach yields for the dynamics of the base radius [R(t)] and contact angle [θdyn(t)] , R(t) ~ t1/7 and θdyn(t) ~ t3/7, respectively. When dissipation is due to viscous flows generated in the core of the spreading droplet, a relation (see Eq. 8) between spreading velocity U, viscosity η, surface tension γ LV of the liquid (all represented by the capillary number Ca = U  η/γ LV), and the value of the contact angle θdyn(t) as a function of spreading time t can be formulated. This relation yields simple scaling laws determining the time evolution of the contact angle and of the droplet’s base radius R(t). For the case of spreading of circular droplets, one finds R(t) ~ t1/10 and θ(t) ~ t3/10. These scaling laws have been examined experimentally and shown to agree fairly well with experimental data for many liquid/solid systems (Tanner 1979; de Gennes 1985; Cazabat and Cohen Stuart 1986; Brochard-Wyart et al. 1991; Marmur 1997). De Gennes (1985) suggested that during the spreading process, the energy resulting from the unbalanced Young force is balanced by the total energy dissipation in the core of the droplet, at the advancing contact line and in a precursor film. Balancing the corresponding capillary driving force (given by Eq. 11) with the viscous force acting at the contact line (related to Eq. 8: Ddiss = Fvisc  U

110

G. Reiter

(de Gennes 1985) with the viscous force Fvisc being proportional to viscosity η and the velocity U at which the liquid moves) leads then to:  
 hmax 1 γ LV θ2dyn  θ2eq ¼  η  U=θdyn ln 2 hmin

5.7

(13)

Dewetting

When a liquid is forced to cover a substrate, which it does not like to wet, the film is either metastable or completely unstable (de Gennes 1985; Brochard-Wyart and de Gennes 1992; Leger and Joanny 1992). Initiated by a defect or some other perturbation, a hole may be nucleated and grow in diameter. The liquid removed from the dry circular spot (the hole in the liquid film) is collected in a liquid ridge (the “rim”) (see Fig. 1). The shape of this rim may be described approximately as a portion of a cylinder that is characterized by two contact angles θdyn and φ (see Fig. 6). In analogy to the complete wetting (see Eq. 13), for the movement of the contact line (position F) in the partial wetting case (with dewetting velocity V ), the following relation may be obtained:   γ LV θ2dyn  θ2eq θdyn ¼ LF  ηV

(14)

evaluated at the position F. Here, LF is an abbreviation for ln hhmax min For hole diameters much larger than the width of the rim, both ends of the rim (denoted by F and R in Fig. 6) move at approximately the same velocity. Moreover, at the position R the equilibrium value of the contact angle is zero (rim in contact with the film of the same liquid) and thus yields: γ LV ϕ3 ¼ LR  ηV

(15)

Here, LR is an abbreviation for ln hhmax evaluated at the position R. min For large rims, small values of Ca = Uη/γ and assuming that Laplace pressure is the same everywhere within the rim (i.e., the same curvature is assumed to exist everywhere in the rim) requires that ϕ ffi θdyn. As can be seen from Eqs. 8 and 13, the values of LF and LR are representing the logarithmic factor ln hhmax . While the min

f

R

F qdyn w

Fig. 6 Schematic representation of the rim forming in the course of dewetting of a thin liquid film. At the front position F of the rim, the contact angle assumes its dynamic value θdyn and at the rear position R it takes the value φ. The width of the rim is given by w

5

Spreading of Liquids on Substrates

111

maximum size is represented by width of the rim and is the same for both values, the minimum length scale is different for the positions F (proportional to the size of the molecules) and R (proportional to the thickness of the film). Thus, for very thin films in the nanometer range, this difference between of LF and LR is small as the width of the rim is typically much larger than the film thickness or the size of the molecules. Consequently, a highly useful relation is obtained between dynamic and equilibrium contact angle for the case of viscous dewetting:   γ LV θ2dyn  θ2eq θdyn ¼ γ LV ϕ3 ϕ ffi θdyn pffiffiffi θdyn ¼ θeq = 2

(16) (17) (18)

and a constant dewetting velocity of γ LV 3  θeq η

(19)

V  V  θ3eq

(20)

V or with V = γ LV/γ

It should be noted that a constant dewetting velocity, as predicted by Eq. 20, will only be observed as long as the assumption of nonslip boundary conditions applies, that is, main dissipation in the liquid wedge close to the contact line that is independent of the actual size of the rim. However, in the case of slippage as it occurs for polymer melts on “ideal” substrates (Reiter and Khanna 2000a, b), dissipation occurs over the whole moving interface that is proportional to the width of the moving rim. Thus, as the size of the rim increases in time, dissipation increases also. The driving capillary force, however, is constant and consequently the dewetting velocity is decreasing in time.

5.8

Concluding Remarks

The above presented concepts of liquid spreading and dewetting apply only in situations where the substrates are “perfect,” that is, chemically homogeneous and physically smooth. For example, on rough substrates spreading (Cazabat and Cohen Stuart 1986; Rolley et al. 1998) and dewetting (Zhang et al. 2005) dynamics will be modified. Moreover, in contrast to spreading of pure and simple fluids, the process can also be quite different for mixtures or solutions. For example, polymer solutions (Fondecave and Brochard-Wyart 1998) or surfactant solutions (Troian et al. 1990; Nikolov et al. 2002) can exhibit quite complex spreading behavior. Spreading will

112

G. Reiter

become even more complex when a chemical reaction between liquid and substrate is involved like in reactive spreading (Eustathopoulos 1998). In addition to spreading and dewetting driven by capillary forces, these processes may be “forced” by applying an external electric field like in “electrowetting” (Vallet et al. 1999; Verheijen and Prins (1999), Mugele and Baret 2005) or a mechanical force by inclining or moving the substrates. The latter process is highly relevant in coating technologies, for example, in the paper industry. As a general summary, the dynamics of spreading and dewetting is controlled by the balance of driving forces (e.g., capillary forces, gravity, or electric forces) and resistance due to dissipative processes that may be increasingly more complicated when complex fluids rather than simple ones are used.

References Blake TD Haynes JM (1969) Kinetics of liquid-liquid displacement. J Coll Interf Sci 30:421 Blake TD (2006) The physics of moving wetting lines. J Coll Interf Sci 299:1 Bonn D, Egges J, Indekeu J, Meunier J, Rolley E (2009) Wetting and spreading. Rev Mod Phys 81:739 Brochard-Wyart F, di Meglio JM, Quéré D, de Gennes PG (1991) Spreading of nonvolatile liquids in a continuum picture. Langmuir 7:335 Brochard-Wyart F, de Gennes PG (1992) Dynamics of partial wetting. Adv Coll Interf Sci 39:1 Carré A, Gastel JC, Shanahan MER (1996) Viscoelastic effects in the spreading of liquids. Nature 379:432 Cazabat AM, Cohen Stuart MA (1986) Dynamics of wetting: effects of surface roughness. J Phys Chem 90:5845 Cazabat AM (1987) How does a droplet spread?. Contemp Phys 28:347 Cazabat AM, Gerdes S, Valignat MP, Villette S (1997) Dynamics of wetting: from theory to experiment. Interf Sci 5:129 Checco A (2009) Liquid spreading under nanoscale confinement. Phys Rev Lett 102:106103 Chen P, Gaydos L, Neumann AW (1996) Contact line quadrilateral relation. Generalization of the neumann triangle relation to include line tension. Langmuir 12:5956 De Coninck J, Blake TD (2008) Wetting and molecular dynamics simulations of simple liquids. Annu Rev Mater Res 38:1 De Gennes PG (1985) Wetting: statics and dynamics. Rev Mod Phys 57:827 De Gennes PG, Hua X, LevinsonP (1990) Dynamics of wetting: local contact angles. J Fluid Mech 212:55 Dussan EB (1979) On the spreading of liquids on solid surfaces: static and dynamic contact lines. Ann Rev Fluid Mech 11:371 Eustathopoulos N (1998) Dynamics of wetting in reactive metal/ceramic systems. Acta Mater 46:2319 Fondecave R, Brochard-Wyart F (1998) Polymers as dewetting agents. Macromolecules 31:9305 Good RJ (1992) Contact angle, wetting, and adhesion: a critical review. J Adhesion Sci Tech 6:1269 Huh C, Scriven LE (1971) Hydrodynamic model of steady movement of a solid/liquid/fluid contact line. J Coll Interf Sci 35:85 Israelachvili J (2011) Intermolecular and surface forces, 3rd edn. Academic Press Leger L, Joanny JF (1992) Liquid spreading. Rep Prog Phys 55:431 Marmur A (1997) Line Tension and the Intrinsic Contact Angle in Solid–Liquid–Fluid Systems. J Coll Interf Sci 186:462 Marmur A (2009) Solid-surface characterization by wetting. Annu Rev Mater Res 39:473

5

Spreading of Liquids on Substrates

113

Mugele F, Baret JC (2005) Electrowetting: from basics to applications. J Phys Condens Matter 17: R705 Neto C, Evans DR, Bonaccurso E, Butt H-J, Craig VSJ (2005) Boundary slip in Newtonian liquids: a review of experimental studies. Rep Prog Phys 68:2859 Nikolov AD, Wasan DT, Chengara A, Koczo K, Policello GA, Kolossvary I (2002) Superspreading driven by Marangoni flow. Adv Coll Interf Sci 96:325 Reiter G, Khanna R (2000a) Real-time determination of the slippage length in autophobic polymer dewetting. Phys Rev Lett 85:2753 Reiter G, Khanna R (2000b) Kinetics of autophobic dewetting of polymer films. Langmuir 16:6351 Rolley E, Guthmann C, Gombrowcicz R, Repain V (1998) Roughness of the contact line on a disordered substrate. Phys Rev Lett 80:2865 Shanahan MER, Carré A (1995) Viscoelastic dissipation in wetting and adhesion phenomena. Langmuir 11:1396 Tanner LH (1979) The spreading of silicone oil drops on horizontal surfaces. J Phys D 2:1473 Troian SM, Herbolzheimer E, Safran SA (1990) Model for the fingering instability of spreading surfactant drops. Phys Rev Lett 65:333 Vallet M, Vallade M, Berge B (1999) Limiting phenomena for the spreading of water on polymer films by electrowetting. Eur Phys J B 11:583 Verheijen HJJ, Prins MWJ (1999) Reversible electrowetting and trapping of charge: model and experiments. Langmuir 15:6616 Young T (1805) An essay on the cohesion of fluids. Philos Trans R Soc Lond 95:65 Zhang F, Barilia G, Boborodea A, Bailly C, Nysten B, Jonas AM (2005) Partial dewetting of polyethylene thin films on rough silicon dioxide surfaces. Langmuir 21:7427 Zisman WA (1964) Chapter 1: Relation of the equilibrium contact angle to liquid and solid constitution. In: Contact angle, wettability, and adhesion. Advances in chemistry series, vol 43. American Chemical Society, Washington, DC, pp 1–51

6

Thermodynamics of Adhesion Wulff Possart and Martin E. R. Shanahan

Contents 6.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.2 A Few Aspects of Intermolecular Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3 Thermodynamic Models of Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.1 The Zisman Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.2 The Concept of Girifalco and Good . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.3 The Fowkes Approach and Its Modifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.3.4 Neumann’s Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

116 117 119 120 121 123 126 127 128

Abstract

Although practical adhesion depends on more than just interfacial bonds, clearly the latter are fundamental: the stronger the links, the higher the adhesion. In many cases, interfacial bonding is of a physical nature and is closely akin to wetting, with the same types of physical phenomena occurring. These can be treated from a thermodynamic standpoint. Wetting data may be used to estimate the thermodynamic energy of adhesion, provided certain assumptions are made and suitable models constructed for, in particular, a relationship between interfacial tensions. Several models have been established with greater or lesser success. The main ones are reviewed here, underlining their essential assumptions. W. Possart (*) Lehrstuhl Adhäsion und Interphasen in Polymeren, Universität des Saarlandes, Saarbrücken, Germany e-mail: [email protected] M. E. R. Shanahan Institut de Mécanique et d’Ingénierie-Bordeaux (I2M), CNRS UMR 5295, Université de Bordeaux, Talence, France e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_6

115

116

6.1

W. Possart and M. E. R. Shanahan

Introduction

As measured by mechanical tests, adhesion strength is not a quantity of reversible thermodynamics. This is due to various factors but mainly related to the dissipation of energy connected with the irreversible process of separation. The act of separation induces local stresses, and if the intrinsic adhesion is sufficiently strong, these are often sufficient to lead to irreversible, or at least time-dependent reversible, strain in one or both members in contact. Anyway, the deformation beyond strictly elastic behavior leads to an input of energy that is not restituted mechanically at all as it dissipates in part as heat or in an entropy term. Another part may be returned on a longer time scale than the separation process. This means that the actual, pragmatic energy of adhesion, as measured in a macroscopic mechanical test, is often far greater than the energy required theoretically for reversible separation at the interface between adhesive and adherend. The effect was shown initially by Gent and Petrich (1969) and later found by many others. Moreover, it has to be pointed out that for the majority of mechanical adhesion tests, the fracture face is not located at that interface. Some residual adhesive is often found on the adherend (or vice versa) as can be shown by surface-sensitive experimental techniques such as infrared external reflection spectroscopy, X-ray photoelectron spectroscopy, thermo-desorption mass spectrometry, secondary ion mass spectroscopy, etc. Nevertheless, the stresses responsible for this behavior require an interfacial linkage or adhesive bonding mechanism (cf. ▶ Chaps. 2, “Theories of Fundamental Adhesion” and ▶ 3, “Forces Involved in Adhesion,” of this book). In cases of good adhesion, this may be due to chemical bonds, but in many cases, physical interactions may be adequate to withstand the applied stresses. It is only the energy for the reversible separation at the very interface that is considered within the framework of thermodynamics. As a starting point, the thermodynamic equations, considered in ▶ Chap. 4, “Wetting of Solids” for wetting (this book), can be easily adopted by replacing the liquid by a solid adhesive S1 and considering the former solid as the adherend S2 now. Hence, the specific energy of adhesion, f adh S1 S2 , for the solid–solid contact takes the form f adh S1 S2 ¼

N
 X
 Fadh S1 S2 ¼  γ S1 V þ γ S2 V  γ S1 S2  μi  Γi, S1 V þ Γi, S2 V  Γi, S1 S2 (1) AS1 S2 i

where Fadh S1 S2 is excess Helmholtz free energy of adhesion and AS1 S2 is D-face area between solid adhesive S1 and adherend S2. The D-face describes the true interphase (cf. ▶ Chap. 4, “Wetting of Solids”), γ IJ = interfacial tension for the D-face IJ, μi = chemical potential of component i in the system, and Γi,IJ = absolute adsorption of component i on the D-face IJ. Clearly, the interfacial tensions and the adsorption state of all three D-faces contribute to this specific excess Helmholtz free energy of adhesion. In comparison, Dupré’s rule (Dupré 1865) for the reversible specific work of adhesion adh Wadh S1 S2 ¼ γ S1 V þ γ S2 V  γ S1 S2  f S1 S2

(2)

6

Thermodynamics of Adhesion

117

considers the part from the interfacial tensions only. As shown in ▶ Chap. 4, “Wetting of Solids”, it is hard to verify this approximation, and hence a considerable portion of energy could be disregarded without notice. Notwithstanding, three unknown quantities still remain in Eq. 2: the solid interfacial tensions. In ▶ Chap. 4, “Wetting of Solids”, a way to estimate the specific wetting energy from contact angle measurement is given provided the same level of approximation is chosen and Young’s equation is employed. This was the origin of the idea to deduce the reversible work of adhesion with Dupré’s rule by exploiting wetting experiments in a clever way. A number of such so-called thermodynamic adhesion models can be found in the literature, and they will be discussed in Sect. 3 of this chapter. However, before doing this, it is helpful to have a closer look at the very nature of the problem. Assuming that some test liquid, L, is used to characterize the wetting of both the solid adhesive and adherend, two Young’s equations are obtained: γ S1 V ¼ γ S1 L þ γ LV cos θe, S1 L γ S2 V ¼ γ S2 L þ γ LV cos θe, S2 L where θe refers to equilibrium contact angle. Note that the vapor phases are not exactly the same as the solids vary. Moreover, it should be pointed out again that Young’s equation applies neither to advancing nor to receding contact angles. The equilibrium contact angle, θe, is needed (cf. Sects. 3 and 4 in ▶ Chap. 4, “Wetting of Solids” of this book). Equation 2 becomes Wadh S1 S2 ¼ γ S1 L þ γ LV cos θ e, S1 L þ γ S2 L þ γ LV cos θ e, S2 L  γ S1 S2 Obviously, the reversible work of adhesion cannot be estimated because the solid interfacial tensions γ S1 V and γ S2 V cannot be deduced explicitly from wetting measurements as outlined at the end of ▶ Chap. 4, “Wetting of Solids”. Even if they could, γ S1 S2 would remain as an unknown quantity. Obviously there are simply less physical equations than unknowns to be found. This has motivated researchers to propose model equations that are supposed to solve this dilemma. These model equations make use of some general features of physical interactions that are summarized in Sect. 2.

6.2

A Few Aspects of Intermolecular Interactions

Molecular interactions inside a phase boundary are the microscopic driving forces for the formation of the interphase as introduced in Sect. 2 of ▶ Chap. 4, “Wetting of Solids.” Energies are additive, and hence the contributions from any type of physical and chemical interaction sum up to the internal energy of the phase boundary. As outlined in ▶ Chap. 3, “Forces Involved in Adhesion”, the theory of physical

118

W. Possart and M. E. R. Shanahan

interactions (also known as van der Waals interactions due to their “prediction” in his well-known modification to the ideal gas equation) between two free molecules A and B provides the approximate equations for the interaction energy of dipolar, induction, and dispersion forces, respectively: Edipole ¼ AB

!

Cdipole AB R6

with

Cdipole ¼ AB

!

2 p 2A p 2B 3ð4πe0 Þ2 kT

(3)

!

p J = permanent dipole moment of molecule J, e0 = absolute dielectric constant, k = Boltzmann constant, T = absolute temperature Einduction AB

Cinduction ¼  AB 6 R

with

Cinduction AB

¼

! ! ! p A! α B pA

! ! ! þ pB! αA pB

ð4πe0 Þ2

(4)

! ! α J = electric polarizability tensor of molecule J Edispersion ¼ AB

Cdispersion AB R6

with

! !  Cdispersion ¼f ! α A, ! α B , other parameters (5) AB

They all possess the same dependence on the distance, R, between the two molecules, and hence they sum up to the energy of an attractive physical interaction: Ephys AB

interact

¼ Edipole þ Einduction þ Edispersion AB AB AB ¼

Cdipole þ Cinduction þ Cdispersion AB AB AB 6 R

(6)

In addition, theory shows that the interaction constants, CAB, can be estimated by CAB ¼

pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi CAA  CBB

(7)

from the interactions between equal molecules A-A and B-B. (In the literature, the harmonic mean is found instead of the geometric mean sometimes.) Starting from these microscopic pair interactions, the Hamaker concept (Hamaker 1937) provides the approximate interaction energy per contact area of the macroscopic phases A and B at distance, d uphys AB

interact

  dispersion induction ¼  Cdipole þ C þ C  f ðd Þ AB AB AB

(8)

Equation 8 gives the contribution of intermolecular physical interactions to the specific internal energy, uAB. (Things become more complicated when the intermolecular distance, d, is of the order of a few molecular diameters.) Correspondingly, the specific excess internal energy of a D-face owing to physical interaction can be written as

6

Thermodynamics of Adhesion

119

disp ind uAB ¼ udipol AB þ uAB þ uAB

(9)

with udipol AB  2

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi dipol udipol AA  uBB ,

uind AB  2

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ind uind AA  uBB ,

udisp AB  2

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi disp udisp AA  uBB

(10)

The interaction energies of chemical bonds are additive, too. It is important to note, however, that no such formulae as Eqs. 3, 4, 5, 6, and 7 apply in general. Hence, Eq. 8 must not simply be extended by another additive term for chemical bonds. Finally, an entropy balance is needed for a complete description of the thermodynamic equilibrium. The specific excess entropy, s, of the D-face between the two phases A and B cannot be deduced from the concept of summation of pair interactions described. Statistical mechanics concepts are needed, which have still not been developed to the necessary quantitative level. In conclusion, only the specific internal energy part, u, of the specific excess free Helmholtz energy for a given D-face f AB ¼ uAB  T  SAB ¼ γ AB þ

X

μi  Γi, AB

(11)

i

can be described, while the specific entropy part, Ts, is still missing.

6.3

Thermodynamic Models of Adhesion

All thermodynamic models of adhesion aim to exploit wetting experiments on solids for a determination of the surface tension of solids and, as far as possible, of the thermodynamic energy of adhesion. As mentioned briefly in Sect. 1, the wetting experiment provides only the difference γ SV  γ SL ¼ γ LV cos θe, SL

(12)

thus leaving us with the two unknown interfacial tensions γ SV and γ SL of the solid. Clearly, an additional equation of state γ SV ¼ f ðγ SL , γ LV Þ

(13)

is needed to solve the problem. The existence of such an equation can be shown by thermodynamics, but it cannot be given explicitly. That is the starting point for the thermodynamic adhesion models: An explicit expression for Eq. 13 is constructed that should possess general validity. It should nevertheless be noted that none of the models is universally accepted and research continues in this field.

120

W. Possart and M. E. R. Shanahan

This issue and the common thermodynamic adhesion models will be discussed below. But as a first illustration, the widely forgotten approach by Antonov (1907) will be considered here. Antonov’s empirical rule for low energy solids (γ LV>γ SV) γ SL  γ LV  γ SV

(14)

is such an assumption for the equation of state (13). Inserting Eq. 14 into Eq. 12 results in cos θ0 

2γ SV 1 γ LV

(15)

Accordingly, the solid surface tension γ SV may be estimated from the wetting experiment with a single liquid but advancing and receding contact angle are not distinguished.

6.3.1

The Zisman Approach

In the light of the inaccessibility of two γ terms in Eq. 12, Fox and Zisman (1952) (Zisman 1962) postulated the following approach. The idea is based on the concept of critical surface tension (of the solid), γ C. They took a homologous series of liquids, e.g., n-paraffins, CnH2n+2, in their liquid state at ambient temperature with n = 5–16 and measured their equilibrium contact angle, θ0, on a given solid. It is not reported if these were the advancing or the receding contact angles. Now, γ LV increases with n, and values are known. Plotting cos θ0 versus γ LV for the various liquids, they obtained a straight line with slope (k) and constant term a: cos θ0 ¼ k  γ LV þ a

(16a)

whose extrapolated value of γ LV for cosθ0 = 1 was termed the critical surface tension, γ C, of the solid. With that, Eq. 16a changes into cos θ0 ¼ k  γ LV þ k  γ C þ 1

(16b)

and Young’s Eq. 12 reduces to γ SV  γ SL ¼ γ LV  γ C

(17)

Zisman also mentioned, however, that it depends very much on the chosen set of liquids whether the plot of cosθ0 versus γ LV provides a straight line. Curves or even irregular scattering are also sometimes observed. Moreover, Eqs. 16 presume a straight line for cosθ0 versus γ LV, while Eq. 15 does so for cosθ0 versus γ 1 LV . Zisman was careful to avoid saying this was the true solid surface tension, although it should be close. In fact, the hypothetical liquid just spreading on the solid should indeed have a surface tension close to that of the solid, provided γ SL is

6

Thermodynamics of Adhesion

121

quite small (cf. Eq. 17). This last supposition is plausible if the liquids and solid are chemically similar. To give two examples, it was found that polyethylene had a γ C of ca. 31 mNm1 and poly(tetrafluoroethylene), ca. 18.5 mNm1. As a next step, Zisman combines Dupré’s rule with Eqs. 12 and 17 to give 2 Wadh SL ¼ γ LV ð1 þ cos θ0 Þ ¼ kγ LV þ ð2 þ kγ C Þγ LV

(18a)

This reversible work of wetting possesses a maximum at 1 γC γ max , LV ¼ þ k 2

, Wadh SL

max

¼

1 k þ γ C þ γ 2C k 4

(18b)

By way of a prediction, Zisman considered this particular liquid as the optimal adhesive provided it retains all its interfacial interactions when it solidifies.

6.3.2

The Concept of Girifalco and Good

Although Zisman’s approach had the merit of originality, it clearly was not general. Despite being rarely true, Antonov’s rule (14) inspired Girifalco and Good (1957, 1960) pffiffiffiffiffiffiffiffiffi to follow a suggestion by Berthelot (for molecular attractions) and write γ SL  γ S γ L. However, the interfacial interaction should surely include additive contributions from each contiguous phase, and thus they arrived at 2 pffiffiffiffiffiffiffiffiffi  1=2 1=2 γ SL ¼ γ S þ γ L  2 γ S γ L ¼ γ S  γ L

(19)

Indeed, Eq. 19 goes back to Rayleigh (1890). It predicts zero interfacial tension if S and L are identical, as would be expected. All this assumed equal molecular sizes for S and L. If different, allowance had to be made, and the final version became pffiffiffiffiffiffiffiffiffi γ SL ¼ γ S þ γ L  2ϕ γ S γ L

(20a)

The dimensionless interaction parameter, ϕ, expresses implicitly the effect of all interactions in the given interphase on the interfacial tension, γ SL. In the framework of the Girifalco–Good concept, Eq. 20a represents the form proposed for the equation of state (13). Hence, Eq. 20a has to refer to all phases involved and should be written more precisely as pffiffiffiffiffiffiffiffiffiffiffiffiffi γ SL ¼ γ SV þ γ LV  2ϕSL γ SV γ LV

(20b)

Good pointed out himself that he had reduced the subscripts only for convenience. Equation 20b clearly shows that the parameter ϕ refers to two well-defined phases in a wetting experiment. Experiments with two “immiscible” liquids serve to estimate ϕ because γ SV is replaced by the known surface tension γ L2 V for the second liquid. ϕAB >1 indicates that intermolecular interactions appear in the interphase

122

W. Possart and M. E. R. Shanahan

A–B, which are not present inside the phases. Specifically ϕL1 L2 is found to vary from 0.32 to 1.17 for 137 pairs of liquids. However, only 15% of these ϕL1 L2 values are in the interval 0.95 ϕ 1.05 (Girifalco and Good 1957). Moreover, it is obvious that Eqs. 20 do not solve the problem of the missing equation of state (13) because a new unknown is introduced, viz., ϕSL. This point will be considered more closely below. Here it is asked first how Eq. 20b relates to the balance of the excess energy for the interface and what the consequences for the wetting experiment are. These questions are important since the specific interactions in an interphase add their interaction energies to the balance, not the corresponding forces. For a liquid wetting a solid, the specific excess Helmholtz free energy of the SL interface is given by (cf. Sect. 2 of ▶ Chap. 4, “Wetting of Solids”) f SL ¼ γ SL þ

X

μi  Γi, SL

(21)

i

where μi is chemical potential of component i and Γi,SL is absolute adsorption of component i at the SL interface. Therefore, consideration of only the interfacial tensions with Eqs. 20 means that the effect of adsorption is neglected in the Girifalco–Good concept: D fD SL  γ SL

(22)

If this approximation is accepted, Young’s and Dupré’s equations may be again introduced in Eq. 20b yielding pffiffiffiffiffiffiffi 1 þ cos θe pffiffiffiffiffiffiffi ϕSL γ SV ¼ γ LV 2

(23a)

Again, Eq. 23a reveals the problem. The wetting experiment with one test liquid L does not provide γ SV since ϕSL cannot be derived independently from wetting data. To circumvent the problem, Good proposed to combine a series of test liquids and to plot cosθe versus γ 0:5 LV according to Eq. 23a rewritten in the form pffiffiffiffiffiffiffi
pffiffiffiffiffiffiffi1 1 cos θe ¼ 2ϕSL γ SV  γ LV

(23b)

A straight line passing through (1) on the ordinate should be obtained if ϕSL possesses the same value for all the liquids used. This cannot be true in a strict sense since, by definition, ϕSL cannot be independent of the given liquid. Moreover, the Good plot implies that the changes in the common vapor phase V can be ignored when the test liquid is changed in the experiment. It was shown in Sect. 4.5, however, that these changes may not be neglected. Also, it has to be noted that Good et al. ignore the difference between the measured static contact angles θa, θr and the equilibrium contact angle θe. If, in addition, ϕSL is assumed to be close to one, an approximate value for γ SV follows from the slope of the straight line.

6

Thermodynamics of Adhesion

123

In the literature many authors have followed Good’s advice, but there is no way reliably to find the resulting error for γ SV . An additional and significant source of error is the use of the advancing (or receding) contact angle instead of θe. Finally, Dupré’s reversible work of adhesion is considered in the framework of the Girifalco–Good concept. Assuming that the surface tensions γ S1 V and γ S2 V of the solid adhesive and of the solid adherend, respectively, were estimated from wetting experiments in the way described above, Eq. 20b is modified to pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi γ S1 S2 ¼ γ S1 V þ γ S2 V  2ϕS1 S2 γ S1 V γ S2 V

(23c)

It has to be noted in passing that the vapor phase has changed again. It does not contain the components of the test liquid anymore. Dupré’s Eq. 2 becomes pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Wadh S1 S2 ¼ γ S1 V þ γ S2 V  γ S1 S2 ¼ 2ϕS1 S2 γ S1 V γ S2 V

(24)

Obviously, the dilemma still exists: The new interaction parameter ϕS1 S2 is not available, and the specific reversible work of adhesion is inaccessible. In Good and Elbing (1970) and Good (1977), Good et al. derive an equation that can be written in a generalized form as Cdipole þ Cinductive þ Cdispersion AB AB AB ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ϕAB  r  ffi (25) dipole dispersion dipole dispersion inductive CAA þ CAA þ CAA þ C CBB þ Cinductive BB BB for the calculation of ϕAB for the interface between phases A and B using the basic formulae of intermolecular physical interactions (cf. Eqs. 3, 4, and 5). However, this is also another approximation.

6.3.3

The Fowkes Approach and Its Modifications

Fowkes (1962, 1963, 1964, 1967) provided another approach for the analysis of wetting experiments by proposing that a surface tension γ (solid or liquid) may be expressed as the sum of a number of various components: γ ¼ γd þ γp þ γi þ γC þ . . .

(26a)

where the superscripts refer to contributions due to the intermolecular interaction forces of various types: dispersion, dipoles, induction interactions, covalent contributions, etc. In practice, he stopped after polarity, assuming most wetting interactions will not involve significant induction or chemical contributions: γ ¼ γd þ γp

(26b)

124

W. Possart and M. E. R. Shanahan

At this point, it must be emphasized again that a surface or interfacial tension is a tensor by its physical origin (cf. ▶ Chap. 4, “Wetting of Solids”) and hence it may not be split into additive scalar components in general. Such additivity holds for the energy, and referring again to Eq. 11 for the specific excess free Helmholtz energy for a surface of phase A in equilibrium with its vapor phase f AV ¼ uAV  T  SAV ¼ γ AV þ

X

μi  Γi, AV

(27)

i

It is seen that one formally arrives at f AV  uAV  γ AV

(28)

only by neglecting the specific excess entropy sAV and all absolute adsorptions Γi,AV. Using Eq. 9, uAV is now written as uAV ¼ upAV þ uiAV þ udAV  γ pAV þ γ iAV þ γ dAV

(29)

and with Eqs. 28 and 26, one gets upAV ¼ γ pAV , uiAV ¼ γ iAV , udAV ¼ γ dAV

(30)

Hence, Eqs. 28, 29, and 30 provide justification for the Fowkes Eqs. 26, and they illustrate the implied approximations. As his explicit version for the general equation of state (13), Fowkes assumes that only dispersion interactions are present in the interphase, and he proposes (for the case of wetting) γ SL ¼ γ SV þ γ LV  2

qffiffiffiffiffiffiffiffiffiffiffiffiffi γ dSV γ dLV

(31)

We add the subscript V here for the sake of correctness. In the original work, no reference is made to the corresponding vapor phase that must exist for thermodynamic equilibrium at the surface of any condensed phase (cf. Sect. 4.2). Obviously, Eq. 31 refers to the dispersion components of the solid and the liquid surface tensions that are new unknown quantities. Fowkes deduces the polar components γ dLV of the surface tensions for a series of liquids from mutual liquid–liquid wetting experiments, and γ pLV is determined from Eq. 26b. It is noted that the vapor phase is specific for any pair of liquids. Hence, the values derived for γ dLV and γ pLV depend on the partner liquid too! The results reported by Fowkes indicate that this is proven in his liquid–liquid wetting experiments. However, if additional interactions (e.g., induction) contribute to the state of the liquid–liquid interphase, there are again more unknowns than equations! Following Fowkes proposition, various workers postulated, more or less successfully, extensions to Eq. 31 in order to include other types of interphase interactions. It

6

Thermodynamics of Adhesion

125

must be emphasized that the term “extended Fowkes equations” is really a misnomer, since Fowkes did not condone their use. However, the two best-known versions are attributable to Owens and Wendt (1969): γ SL ¼ γ S þ γ L  2

qffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffi γ dS γ dL  2 γ pS γ pL

(32)

and to Wu (1971): γ SL ¼ γ S þ γ L 


4γ pS γ pL 4γ dS γ dL 
  γ dS þ γ dL γ pS þ γ pL

(33)

These two variants, based on geometric and harmonic means for the surface tension components according to Fowkes, are often used, but their theoretical basis is questionable as will be shown below. A further variant of Fowkes early work is due to van Oss et al. (1988). It is based on the concept that weak acid–base interactions (of the Lewis type) may contribute to interfacial tensions. Couched in a formalism inspired by Fowkes’ earlier work, it is proposed that a (general) surface tension may be written as γ ¼ γ LW þ γ AB where γ AB ¼ 2

pffiffiffiffiffiffiffiffiffiffi γþγ

(34)

In Eq. 34, γ LW represents the dispersion and polar components of Fowkes’ expression (26b), all together, and γ AB is the acid–base, or electron acceptor/donor, contribution. The terms γ + and γ  represent electron acceptor and electron donor characteristics of the phase. After some algebra, it is shown that in this approach, the expression for γ SL becomes γ SL ¼

qffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffi2 LW   γþ γ LW  γ þ 2 γþ L L γL S γS þ S qffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffi þ  2 γþ γ S γL S γL þ

(35)

The second term on the right amounts to the internal acid–base interactions, and the last is due to interfacial acid–base interactions. Some doubt has been cast on the validity of this formulation, but the intrinsic physical chemistry of allowing for electron acceptor and electron donor characteristics seems sound. Let us now look at the implications of Eqs. 31, 32, and 33. Using Eqs. 9, 10, and 28, the specific excess Helmholtz free energy can be expressed as f SL  uSL ¼ upSL þ uiSL þ udSL ¼ 2

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi upSV  upLV þ 2 uiSV  uiLV þ 2 udSV  udLV (36)

Dupré’s reversible work of wetting is introduced as

126

W. Possart and M. E. R. Shanahan

Wadh SL ¼ γ SV þ γ LV  γ SL  f SL  uSL

(37)

and with Eq. 30 follows Wadh SL ¼ γ SV þ γ LV  γ SL  2

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi γ pSV  γ pLV  2 γ iSV  γ iLV  2 γ dSV  γ dLV

(38)

which is a generalized form of Eqs. 31 and 32. A corresponding form for Eq. 33 would follow by replacing geometric by harmonic means. As before, W adh SL could be obtained from a proper wetting experiment. All these concepts suffer from the same problem as the thermodynamic adhesion model proposed by Girifalco and Good. Provided that attention is confined to polar and dispersion interactions and the use of the values for γ dLV and γ pLV from the liquid–liquid wetting is made, one still has two unknowns in Eq. 38: γ dSV and γ pSV . In the literature, this problem is circumvented (but not solved!) by incorporating the wetting results from a second liquid (or even a series of test liquids) for the data analysis: With any additional liquid, Eq. 38 can be rewritten, thus expanding the starting point from a single equation into an unlimited system of simultaneous equations. Although mathematically correct, this procedure is inconsistent with several aspects of the fundamental thermodynamic concept of the interphase and its corresponding D-face or interface (cf. ▶ Chap. 4, “Wetting of Solids”). As a result, the reliability of data deduced for γ dSV , γ pSV , γ iSV , and hence γ SV cannot be assessed. Bearing this in mind, now a brief look at the work of adhesion between solid adhesive and adherend in the framework of Fowkes’ general approach is taken. Equation 38 is of the form Wadh S1 S2 ¼ γ S1 V þ γ S2 V  γ S1 S2 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  2 γ pS1 V  γ pS2 V  2 γ iS1 V  γ iS2 V  2 γ dS1 V  γ dS2 V

(39)

Obviously, Wadh S1 S2 is as much in doubt as the components of the solid and liquid surface tensions. These are the physical limits of all these concepts.

6.3.4

Neumann’s Approach

Neumann et al. (1974) have adopted the stance of considering the advancing contact angle to be related to an equation of state. He followed the evolution of γ L cos θa = f(γ L) for a series of liquids on eight fluoropolymer surfaces. Thus, Neumann follows the lines of his predecessors. Combining the results of wetting experiments with an arbitrary choice of test liquids on a small number of polymer surfaces, he considers the phenomenological relationships found for that data pool as generally applicable. The treatment ignores the fact that all wetting pairs represent independent thermodynamic systems.

6

Thermodynamics of Adhesion

127

Neglecting spreading pressure, π e, he observed a linear variation of ϕ of the Girifalco–Good equation (20a). Combining this linearity with Girifalco and Good’s relation, he arrived at the following expression for Eq. 13: γ SL

pffiffiffiffiffi pffiffiffiffiffi 2 γS  γL 1 ¼ pffiffiffiffiffiffiffiffiffi with A ¼ 15m  N 1  A  γS γL

(40)

This relation is purely empiric and thus valid for the cases where it has been established. In another attempt, Li and Neumann (1990) used arguments based on the theory of intermolecular physical interactions along the lines of Eqs. 3, 4, 5, 6, and 7. Following Fowkes’ concept, they derived the expression 2 pffiffiffiffiffiffiffiffiffiffiffiffiffi γ SL ¼ γ LV þ γ SV  2 γ LV γ SV  eβðγLV γSV Þ

(41)

as an alternative version for the equation of state (13). Here, the model parameter β is the newly introduced unknown. From a pool of wetting data, β was derived to be equal to 124.7 m2N2 (Li and Neumann 1992). Hence, this relation is purely empirical, too.

6.4

Conclusions

All thermodynamic adhesion models start with the idea that wetting experiments with suitably chosen test liquids could provide the surface tension, γ SV, of the wetted solid. As shown in the previous sections of this chapter, numerous approximations are made in each of these evaluations of the wetting data. Let us mention three of the more important simplifications: (a) the thermodynamic adhesion models blur the physical distinction between interfacial tension and thermodynamic excess energy of an interface; (b) they neglect the contribution of the interphase composition (i.e., the role of the absolute adsorptions Γi); and (c) they ignore the fact that an interfacial tension belongs to two contacting phases in the framework of thermodynamics. In many published papers, the problem of how the thermodynamic contact angle is derived from measured advancing or receding angles is overlooked. These and more approximations are needed to handle the key problem of the overall situation: How should an explicit form for a general equation of state be constructed that is only implicitly given by theory γ SV = f(γ SL, γ LV)? The models vary considerably as can be seen by comparing Eqs. 17, 20, 31, 32, 33, 35, 40, and 41. Moreover, it is not presently possible to assess the accuracy of any of these approximate equations. Nevertheless, thermodynamic adhesion models are widely used in the literature. In conclusion, it must be emphasized that wetting experiments do not provide enough information on interphases for deducing the interfacial tensions, γ SV, γ SL of a solid, or, even worse, for the reversible work of adhesion, W adh S1 S2 . As far as future work is concerned, it would be a challenging goal to find an experiment enabling direct measurement of the specific excess free Helmholtz energy of adhesion Δf adh.

128

W. Possart and M. E. R. Shanahan

References Antonov GN (1907) Sur la tension superficielle à la limite de deux couches. J Chim Phys 372:5 Dupré A (1869) Théorie Méchanique de la Chaleur. Gauthier-Villars, Paris, p 369 Fowkes FM (1962) Determination of interfacial tensions, contact angles, and dispersion forces in surfaces by assuming additivity of intermolecular interactions in surfaces. J Phys Chem 382:66 Fowkes FM (1963) Additivity of intermolecular forces at interfaces. I. Determination of the contribution to surface and interfacial tensions of dispersion forces in various liquids. J Phys Chem 2538:67 Fowkes FM (1964) Attractive forces at interfaces. Ind Eng Chem 40:56 Fowkes FM (1967) In: Patrick RL (ed) Treatise on adhesion and adhesives, vol 1. Marcel Dekker, New York, pp 325–449 Fox HW, Zisman WA (1952) The spreading of liquids on low-energy surfaces. III. Hydrocarbon surfaces. J Coll Interface Sci 428:7 Gent A, Petrich RP (1969) Adhesion of viscoelastic materials to rigid substrates. Proc Roy Soc A 433:310 Girifalco LA, Good RJ (1957) A theory for the estimation of surface and interfacial energies. I. Derivation and application to interfacial tensions, J Phys Chem 904:61 Good RJ (1977) Surface free energy of solids and liquids: thermodynamics, molecular forces and structure. J Coll Interface Sci 398:59 Good RJ, Elbing E (1970) Generalization of theory for estimation of interfacial energies. Ind Eng Chem 3 (54):62 Good RJ, Girifalco LA (1960) A theory for estimation of surface and interfacial energies. III. Estimation of surface energies of solids from contact angle data. J Phys Chem 561:64 Hamaker HC (1937) The London – Van Der Waals attraction between spherical particles. Physica 1058:4 Li D, Neumann AW (1990) A reformulation of the Equation of State for interfacial tensions. J Coll Interface Sci 304:137 Li D, Neumann AW (1992) Contact angles on hydrophobic solid surfaces and their interpretation. J Coll Interface Sci 190:148 Neumann AW, Good RJ, Hope CJ, Sejpal M (1974) An equation-of-state approach to determine surface tensions of low-energy solids from contact angles. J Coll Interface Sci 291:49 Owens DK, Wendt RC (1969) Estimation of the surface free energy of polymers. J Appl Polym Sci 1741:13 Rayleigh (Strutt JW) (1890) LII. On the theory of surface forces. Phil Mag Ser 5 456:30 van Oss CJ, Chaudhury MK, Good RJ (1988) Interfacial Lifshitz-van der Waals and polar interactions in macroscopic systems. Chem Rev 927:88 Wu S (1971) Calculation of interfacial tension in polymer systems. J Polym Sci C 19:34 Zisman WA (1963) Influence of constitution on adhesion. Ind Eng Chem 10(19):55

Part II Surface Treatments

7

General Introduction to Surface Treatments Gary Critchlow

Contents 7.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.2 The Problem with Surface Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3 Removal of Surface Contamination . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.1 Degreasing Metallic Adherends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.3.2 Polymer Treatment: Removal of Contamination and Surface Functionalization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.4 Mechanical Treatments and the Role of the Interphase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5 Chemical Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.1 Chemical Treatments for Metallic Adherends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.5.2 Chemical Treatments for Polymeric Adherends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6 Electrochemical Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.1 Introduction to Anodizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.2 Structure and Composition of Anodic Films . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.3 Physical Factors Affecting Anodic Oxides . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.4 Alternatives to Phosphoric and Chromic Acid Anodizing . . . . . . . . . . . . . . . . . . . . . . . 7.6.5 Alternating (AC) Current Anodizing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7.6.6 Alternating Current–Direct Current Anodizing (ACDC Anodizing) . . . . . . . . . . . . 7.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

132 133 137 137 141 142 145 145 148 149 149 151 153 153 155 156 159 160

Abstract

The use of surface treatments to optimize adhesion has been well established. In this chapter, the main treatment methods for metals and polymers will be considered in terms of how such processes are carried out and their influence on surface physical and chemical properties. Consideration has been given to a range of treatments from simple degrease options to the more highly complex multistage

G. Critchlow (*) Department of Materials, Loughborough University, Loughborough, Leicestershire, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_7

131

132

G. Critchlow

processes. An attempt has been made to relate the changes in physicochemical properties to adhesion performance, and, where relevant durability.

7.1

Introduction

It is usually recommended that some degree of modification or treatment is applied to all surfaces prior to adhesive bonding. This is particularly important if structural or semi-structural bonding is to be carried out. Some examples of specific surface treatments and their influence on initial adhesion and durability are given later in this section and, more extensively, in ▶ Chap. 8, “Surface Treatments of Selected Materials.” The particular surface treatment applied in any given situation will depend upon the requirements of the bond and the service conditions that it will experience and will generally be chosen on a “fit-for-purpose” basis. Additional factors such as the adherend type and geometry and production constraints, including cost considerations, may also be relevant factors when selecting a treatment process. In simple terms, treatments are required to make the surface receptive to the applied adhesive. A treatment may also be required to provide a mechanically sound surface to be bonded and one which will also minimize the potentially damaging influence of external factors such as environmental ageing, corrosion, or applied loading. From the relevant literature, it is clear that the main function of any treatment on any surface is to provide desirable physical and chemical properties to that surface which will achieve the above-mentioned objectives. To do this, a particular treatment will facilitate one or more of the following functions: • To remove, or prevent the formation of, any weak boundary layer on the adherend surface. Examples of weak boundary layers include machine oils, weak oxides or hydroxides, and airborne contamination on metals, or, internal segregants from within polymers. A simple degrease may fulfill this requirement. Alternatively, mechanical or deoxidizing treatments may be required if friable, inorganic layers are present. • To maximize the degree of intimate molecular contact between the adhesive or primer and the adherend during the bonding and curing processes. This usually involves the creation of a high-surface free energy and the introduction of specific functional groups. • To generate a specific structure or texture on the adherend surface. Even though mechanical interlocking may not be responsible for primary bonding, the complex structures present on, for example, conversion-coated or anodized surfaces may help increase joint fracture energies by redistributing the stresses away from any interphasial polymer. These structures also create an extended surface area, which may be utilized if the adhesive is able to flow into, for example, the pores on anodic oxides. • To passivate the surface of the adherend prior to adhesive bonding. This is paramount for high-energy adherends as these are not only receptive toward the adhesive but also to all forms of contamination. This contamination may be

7

General Introduction to Surface Treatments

133

organic or inorganic and can influence the adherend in a number of ways, including by a reduction in its surface free energy, gross contamination, or corrosion. Passivating processes offer protection of the surface after bonding by providing electrochemical and barrier corrosion resistance to the substrate material. The minimum preparation which is usually carried out on adherend surfaces might include a simple degrease to remove processing aids and contaminants. There does, however, exist a hierarchy of treatments. It is generally thought that mechanical processes such as abrasion or grit-blasting offer improved bonding surfaces compared with degreasing-only and, in some cases, offer adequate levels of adhesion either alone or with coupling agents applied. Further information on the role of coupling agents is given in ▶ Chap. 11, “Organosilanes: Adhesion Promoters and Primers.” Similarly, simple chemical treatments such as acid etching tend to offer further enhanced bond performance compared with mechanical treatments. However, it is recognized that the current state-of-the-art processes, particularly those for structural or semi-structural metal bonding and which are highly complex, multistage treatments, including chemical conversion coating and anodic oxidation of metals perform best of all. The more advanced processes provide the adherend surface with all of the required physical and chemical properties mentioned above to obtain optimized initial adhesion and, importantly, durability. This chapter will firstly consider the problems associated with bonding as-received surfaces and the problems caused by the presence of contamination. Following this, details will be given of different surface treatment methods in terms of how they modify the above-mentioned surface physical and chemical properties, and how these changes impact upon the measured adhesion levels. In some cases, consideration has been given to alternative environmentally friendly surface treatments which might be considered as drop-in replacements for the current industrial standards. Many of the currently used processes are detailed further in ▶ Chap. 8, “Surface Treatments of Selected Materials.” This chapter will focus on the modification or treatment of metallic and polymeric adherends. It is, however, recognized that adhesion to other materials such as wood, ceramics, and glasses are industrially significant, and mention of relevant treatments for these materials are given in ▶ Chap. 8, “Surface Treatments of Selected Materials.”

7.2

The Problem with Surface Contamination

The removal of preexisting contamination, which often comprises low energy and friable material, to facilitate bonding to a mechanically sound surface is a fundamental requirement in establishing good adhesion. The requirement to remove such material is often overlooked, but it is, in fact, an essential stage in the more complex processes such as anodizing also. It is important at this stage to note that when considering polymers, it is possible that internal migration of low-molecular-weight materials or processing aids can

134

G. Critchlow

occur to create a highly modified surface layer. This layer might be considered as comprising internal contaminants. A simple illustration of this has been given by Ansanfar et al. (2009) in their consideration of natural rubber (NR) and NR-sytrenebutadiene rubber blends. In this study, internal migration of paraffin wax, a processing aid, was directly observed, which led to poor self-adhesion between these polymers. In addition to influencing the surface thermodynamics, such internal contaminants may also form weak boundary layers; this topic has been extensively discussed following the early studies by Brewis (1982). The removal of such internal contaminants will be considered later in this section. Following is a consideration of external contaminants on adhesive joints. There are many possible forms of external contamination; a reasonably comprehensive list is given by Averill et al. (1998). For simplicity, surface contaminants can be considered to fall within two broad groups: organic and inorganic materials. The removal of the organic compounds can, generally, be achieved using simple degrease processes, whereas the inorganic compounds might require a combination of degrease, desmut and, in extreme cases, deoxidizing solutions. As previously mentioned, substrate contamination can have a major effect upon the resultant quality of the bond due to the influence of the contaminant reducing the surface free energy. Naturally, in any conscientious facility that undertakes adhesive bonding, contamination will be reduced to an absolute minimum. In one study, Smith and Crane (1980) looked at the influence of controlled levels of contamination from various sources and determined their effects upon adhesive bond strength using aluminum alloy adherends. Ellipsometry and water contact angle measurements were used to study a range of process variables and contaminants. Aluminum panels treated by phosphoric or chromic acid anodizing (PAA or CAA) were contaminated with materials likely to be introduced by human error, including finger prints, cigarette smoke, and mucous, along with drink residues. Contaminants originating from a factory environment were also considered, including lubricating oils and silicone grease. In all cases, ellipsometry and water contact angles were able to detect the presence of these materials. In this study, it was noted that the presence of a level of 10 nm, and above, of lubricating oils had no effect upon the lap shear strength of joints made with a single part epoxide adhesive, even though an increase in the water contact angle was observed. Increasing the levels of oil did result in an increase in water contact angle toward 100
. Silicone grease was also introduced and evaluated; in this case, a more dramatic effect was noted in that a significant decrease in lap shear strength was observed. In an assessment of the sensitivity of the two types of surface treatment, small amounts of silicone grease, from 3.5 to 20 nm, were found to increase the water contact angle value and, using lap-shear testing, this was accompanied by a significant reduction in the failure load. Almost total apparent interfacial failure was noted for the silicone-contaminated joints. Significant work carried out to evaluate the effects of contamination has also been conducted in the area of automotive bonding, where adhesion to lubricated surfaces is desirable. Research carried out principally on behalf of the automotive industry has studied the ability of adhesives to absorb or displace contaminants to facilitate

7

General Introduction to Surface Treatments

135

the formation of bonds with the underlying substrate. In one study carried out on electrogalvanized steel adherends, Hong and Boeno (1995) showed that the adhesive formulation can play a key role in successfully accommodating oils on such surfaces. In a related study, Minford (1981) looked at the effects of lubricants as contaminants on bonded and spot-welded aluminum substrates. For the substrates, he used 2036 T4 aluminum lap-shear joints bonded with a single part hot-curing epoxide. Surface contamination was controlled by immersing the substrates to be bonded in varying concentrations of an emulsified forming lubricant. Durability was tested by exposing joints to 52
C and 100% relative humidity and measuring residual failure loads. Joints were also subjected to a 3.5% salt fog for 16 h in a 24-h cycle. Both simulated service conditions were assessed against freshly prepared joints. Other than the joints that were deliberately contaminated, all adherends received no other treatment. It was found that the adhesive was able to accommodate up to 0.82 mg/cm2 of oil before any loss of strength was noticeable. At the 0.95 mg/cm2 level, there was a significant loss in strength from 10 to 6 MPa and a shift in the locus of failure dropping from over 90% cohesive to demonstrate a mixture of 60% cohesive and 40% apparent interfacial failure. For the environmental ageing of joints, specimens were exposed to either humidity or salt-spray conditions for up to 180 days. Joints were periodically removed and loaded to 50% of their original bond strength and then finally tested to destruction after 280 days of exposure. Joints retained their original strengths with up to 0.62 mg/cm2 level of contamination. At 0.71 mg/cm2, a reduction to 45% of the original joint strengths was observed and at 0.82 mg/cm2, the results became variable. Paradoxically, contaminated joints exposed to salt-spray conditions seem to perform better than the as-milled substrates which failed at 2% of their original joint strength after only 90 days exposure. The only contaminated joints that performed worse than this were the ones prepared at the 1.05 mg/cm2 level. Overall, it was this latter level of contamination that caused the most detrimental effect upon the fresh and aged specimens. Debski et al. (1986) have tried to explain the mechanisms by which the adhesive is able to function even through oily layers by investigating the ability of an epoxide adhesive to either displace or absorb such contamination. They determined the surface free energies for a diglycidyl ether of bisphenol A (DGEBA) based epoxide resin on ASTM 3 grade oil, plain steel, and a zinc-coated steel. Utilizing these surface free energy values, they investigated the thermodynamic possibility of the epoxide adhesive displacing a layer of an apolar oil on a steel surface. The model was that of a drop of resin spread over a layer of oil upon a metal substrate. Assuming that the resin remained geometrically constant they concluded that, theoretically, the oil could be displaced if the substrate was flat and determined that no oil could become entrapped within surface crevices. To deal with this entrapped oil, they assumed that any oil trapped beneath the adhesive was absorbed into the epoxide matrix. Using differential scanning calorimetry (DSC), they monitored the cross-linking kinetics of an adhesive containing 6% dicyandiamide and of the adhesive with the same formulation but with the addition of a CaCO3 filler. They found that the effect of the increase in oil content was to lower the glass transition temperature (Tg) during the cross-linking process. However, the Tg of the fully cross-

136

G. Critchlow

linked adhesive was unaffected even at the highest levels of oil contamination. Similar trends were also noted for the unfilled epoxide adhesive. Thermal gravimetric analysis (TGA) was also conducted to evaluate the levels of oil that were lost due to evaporation from the free surfaces of the adhesive samples. The filled and unfilled adhesives both containing 10% oil were monitored at 210
C for weight loss alongside control samples containing no oil. After 1 h of heating, no loss of weight was recorded for either control sample; however, the unfilled sample with oil lost approximately 3% the filled version lost 1% and a pure oil sample lost 36% of its original weight. This study demonstrated the capacity of an epoxide adhesive to accommodate oil from substrates in one of three ways; the displacement, absorption, and “digestion” by the filler and absorption followed by evaporation at free surfaces. However, the latter would be a second-order mechanism in real joints as the free surface to volume ratio would be significantly lower than that present in the experiments conducted. Anderson (1993) also discovered that contaminated bondlines can behave differently under different loading conditions. He studied the effects of HD-2 grease upon steel butt joints and tapered double cantilever beam test specimens that were bonded together with either Hysol EA946 or EA 913 epoxide adhesives. The substrates were prepared by vapor degreasing, grit-blasting, and a final vapor degrease prior to bonding. For the contaminated joints, varying solutions ranging from 0.5% to 2% of HD-2 grease were prepared by thinning down in 1,1,1 trichloroethane, which was then sprayed onto one of the substrates from a bonded pair. The test results from the tensile butt joints showed that there was a change of failure mode from that of cohesive to one of near interfacial even for the lowest levels of contamination. Joints prepared with the EA 913 adhesive lost approximately 50% of their overall tensile strength at a contamination level of 400 mg/m2, while EA946 was hardly affected. The results for the tapered double cantilever beam trials were somewhat different, and significant effects were noted for both adhesives. The EA946 results were significantly reduced from approximately 1,200 N/m to around 600 N/m at the 100 mg/m2 level. The EA913 results were also similarly affected and were significantly reduced from an initial value of 400 N/m for the uncontaminated specimens to approximately 20 N/m at the lowest levels of contamination, 50 mg/m2. Trials were repeated for the EA 946 adhesive but with the addition of a silane primer to the bondline. The silane was applied by spraying a solution of 5% A187 in 40% toluene, 40% absolute ethanol, 5% butoxy ethanol, 5% deionized water, and 5% n-butanol. The results of both the tensile butt joint and cantilever beam trial showed improvements in the resistance to contaminant levels. For the butt joint trials, only a small loss of strength was observed up to the 200 mg/m2 contaminant level. There was also a notable improvement in the fracture toughness values from a value of 1,200 to 4,000 N/m for the uncontaminated test specimens. Losses were again noted at the 100 mg/m2 level of contamination. In the above studies, the influence of various deliberately introduced chemical contaminants on the prebond condition of the adherends has been studied. Also of significance is the influence of prebond handing and the ability of some surface treatments to accommodate physical damage as well as contamination; this aspect has been investigated by McNamara et al. (1979).

7

General Introduction to Surface Treatments

137

It is clear that the role of surface contamination and its effects upon the resultant joint strength is complex; this point has been further highlighted in the work of Olsson-Jacques et al. (1996) who studied the influence of adsorbed aviation fuel on various differently treated surfaces. It is, however, important to note that the degree of degradation is a function of the specific adhesive-primer-treatment-adherend combination used. Therefore, to ensure good adhesion, in the general case, it is recommended that contamination is removed prior to adhesive application. Some methods to achieve this for metallic and polymeric adherends are discussed in the following sections.

7.3

Removal of Surface Contamination

7.3.1

Degreasing Metallic Adherends

For metals, an atomically clean surface will have a positive spreading coefficient, so wetting by a subsequently applied adhesive will occur; this topic has been dealt with elsewhere in this book and previously by Shanahan (2005). A large spreading coefficient is particularly important if complete wetting of highly textured surfaces is to be achieved. Note that the texture produced by various treatments may vary from the macro- or microscale created by grit-blasting to the nanoscale porosity present on anodized surfaces. Surface cleanliness, surface free energy, and wettability are intimately linked as has also been discussed, for example, by Rance (1985). Degreasing is the simplest possible method of achieving a clean surface and thereby increasing its surface free energy and spreading coefficient. The simplest form of degreasing, involving the use of organic solvents, is used as a stand-alone treatment and as the first stage in a wide variety of multistage processes. Degreasing methods can range from simple immersion or wiping techniques to vapor degreasing or ultrasonic cleaning. The solvents industrially employed include isopropyl alcohol, acetone, methyl ethyl ketone, perchloroethylene, trichloroethylene, or 1,1,1 trichloroethane. However, environmental as well as health and safety legislation is rapidly reducing the range of solvents that can be used for this method (Averill et al. 1998). Furthermore, some standards agencies no longer advocate the use of solvents for the degreasing and cleaning of substrates to be bonded, preferring instead detergent degreasing or alkaline cleaning. For example, Lesley and Davies screened a range of approximately 150 solvents and degreasing agents in the search for a cleaner to replace the preexisting chlorinated vapor degreasing of aluminum and steel components for adhesive bonding. They concluded that substrates treated with aqueous cleaners resulted in surfaces with less residual organic matter than with any of the organic solvents they reviewed. Extensive studies have been carried out by Thiokol on replacements for methyl chloroform (TCA) for the cleaning of large-scale parts with great success achieved by various aqueous-based cleaners. Alternatives to the commonly used degreasing processes are sought for many reasons, for example: established processes may not be adequate for difficult-to-

138

G. Critchlow

Table 1 A summary of cleaners studied code to treatments Code No. 1 2 3 4 5 6 7 8 9 10 11

Process description Untreated 20% Circamax 103 alkaline cleaning solution 2% Domestic detergent 1 solution 2% Domestic detergent 2 solution 2% Domestic detergent 3 solution 2% NMR Sea Power 100 solution 2% NMR Sea Power 100 solution +1% sodium hydrogen orthophosphate solution 2% NMR Sea Power 100 solution +1% sodium hydrogen orthophosphate +1% sodium metasilicate solution 0.5% NMR Sea Power 100 solution +1% sodium hydrogen orthophosphate solution 5% Domestic detergent 4 + 1% sodium hydrogen orthophosphate solution 1% Domestic detergent 4 + 1% sodium hydrogen orthophosphate solution 20% Circamax 103 alkaline cleaner solution followed by 20% Circamax 115 acid cleaner 2% NMR Sea Power 100 solution +1% sodium hydrogen orthophosphate solution followed by a 4% sulfuric acid solution

remove materials or the processes may use undesirable volatile organic compounds (VOCs); alternatively, they can be carcinogenic or ozone depleting. In a study by Smith et al. (2000), a range of aqueous-based cleaners was evaluated in terms of their effectiveness on stainless steel. The usefulness of contact angle analysis was also demonstrated in this work. It is recognized that a monolayer of organic material present on metal surfaces will, typically, give a water contact angle between 55
and 95
. If the organic layer is effectively removed, the contact angle will be reduced to zero. Contact angle measurements can provide a simple method to assess the effectiveness of various cleaning agents. Smith et al. considered a range of cleaners on as-received stainless steel; the cleaners are detailed in Table 1. It was shown that the detergents were moderately effective in removing organic material from the stainless steel as can be seen from the much reduced contact angles in Table 2, compared with the untreated control. However, they were not as effective as the proprietary metal cleaner. Of particular interest, a natural detergent based on seaweed proved very effective in removing organic material as was confirmed using Auger electron spectroscopy (AES), particularly when used in combination with sodium hydrogen orthophosphate. As would be expected, the degree of cleaning increased with treatment time. Ultrasonics also significantly increased the effectiveness of the proprietary cleaner and the rate of cleaning. The alkaline and acid-based cleaners were also shown by AES to modify the surface oxide by enrichment of chromium, and the thickness of the oxide layer was reduced by both the alkaline and the acidic etches. The latter point may be relevant in terms of the durability of subsequently produced bonds. For the removal of strongly adsorbed or chemisorbed contamination, simple solvent cleaning may not be effective. Litchfield et al. (2006) have detailed a

7

General Introduction to Surface Treatments

139

Table 2 Summary of contact angle data Pretreatment Code 1 2 2 3 4 5 6 7 8 9 10 11

Ultrasonics – √ – √ √ √ √ √ √ √ √ √

Contact angles after x minutes in cleaning solution 0 2 5 10 89 14 11 9 23 14 17 – 36 37 31 22 – 42 28 – 72 83 53 36 23 33 19 10 50 41 17 15 12 30 54 21 20

20 5 10 – – – – 16 13 20 30 15

range of processes for the removal of cross-linked epoxide resins from both metal and composite surfaces. Methods such as sodium hydride immersion, CO2cryoblasting, and laser ablation all proved highly effective. In a further study, the efficacy of CO2-laser ablation was also demonstrated in terms of its ability to produce atomically clean surfaces with a resultant increase in surface energy and observed initial single lap shear (SLS) joint strength and durability, as measured using statically stressed SLS joints (Critchlow et al. 1997) with metallic adherends. In a related study by Bedwell et al. (1998), CO2-laser ablation was studied to establish the effectiveness of simply removing adventitious organic contamination from industrially sourced hot dipped galvanized (HDG) mild steel. This laser-based process was demonstrated to be more effective than double-degreasing combined with ultrasonic immersion in acetone for the removal of organic contamination from a previously oil-contaminated surface. In addition, the TEA CO2-laser was shown to modify both the near surface chemistry and the topography; the precise modifications being dependent upon the degree of treatment used. Importantly, in bond durability trials, the changes introduced to the laser-treated HDG surface were shown to provide equivalent or better adhesion levels compared with degreasedonly adherends; see Table 3. Furthermore, in durability trials, using statically stressed SLS joints immersed in hot water, a commercially available phosphatebased, chemical conversion coating was shown to provide bond durability significantly inferior to the TEA CO2-laser treatment. It should be noted though that a chromate-based process gave markedly superior bond durability in these trials; see Table 4. Lu and Aoyage (1994) also confirmed the effectiveness of laser cleaning on various substrates using a KrF excimer source. Another process which is gaining popularity is the use of CO2-cryoblasting as discussed by Brewis et al. (1999). The combined thermomechanical action of solid CO2 particles impacting upon surfaces combined with the effective solvent action

140

G. Critchlow

Table 3 Initial joint strengths (kN) of SLS joints as a function of surface treatment Treatment Double degrease Gardobond 250 Gardobond 4504 Gardobond 250 + 4504 TEA CO2-laser treatment – 10 pulses TEA CO2-laser treatment – 20 pulses TEA CO2-laser treatment – 30 pulses

Joint strength (kN), 1 standard deviation 4.4  0.08 4.5  0.22 4.5  0.23 3.6  0.11 3.9  0.31 4.0  0.56 4.1  0.22

Table 4 Mean times-to-failure, t f, of SSLS joints immersed in water at 60
C at applied loads of 0.2 and 0.5 kN as a function of surface treatment Treatment Double degrease Gardobond 250 Gardobond 4504 Gardobond 250 + 4504 TEA CO 2-laser treatment – 10 pulses TEA CO 2-laser treatment – 20 pulses TEA CO 2-laser treatment – 30 pulses

Applied load (kN) 0.2 0.5 0.2 0.5 0.2 0.5 0.2 0.5 0.2 0.5 0.2 0.5 0.2 0.5

Mean t f value (h) 1 standard deviation 163  32 85  12 90  16 27  7 235  36 172  6 198  53 27  7 177  38 37  11 217  30 79  30 162  30 58  18

means that highly contaminated surfaces can be readily cleaned very quickly and over large areas. This process was shown by Brewis to effectively leave an atomically clean surface on previously contaminated aluminum and to introduce some surface texture. SLS joint testing was carried out using aluminum 5251 alloy adherends and, in terms of creating initial adhesion, this treatment proved better than simple ultrasonic degreasing and as effective as grit-blasting. A further development by Deffeyes et al. (1997) is the use of cryoblasting followed by ultravioletlight irradiation which proved to be highly effective at removing oils and waxes from aluminum surfaces. This combination process was shown to produce a zero-degree water contact angle on the treated surface, indicating complete removal of the preexisting contaminants. In summary, although externally sourced surface contamination can be accommodated by some adhesive-substrate combinations, it is generally regarded as good practice to remove such material prior to adhesive application. There are many treatments which will achieve this outcome; the preferred one will depend upon a

7

General Introduction to Surface Treatments

141

number of factors, including the chemical nature of the contaminant and adherends. Some degrease options can be considered as reasonably effective stand-alone treatments. In general terms though, mechanical, chemical, electrochemical, and some alternative treatments tend to give improved levels of adhesion and durability compared with simply degreasing.

7.3.2

Polymer Treatment: Removal of Contamination and Surface Functionalization

As previously mentioned, the surface of polymers can comprise contaminant from both internal and external sources. The aim of many polymer pretreatments is often to remove such material and to introduce or to expose desirable chemical functionality. This topic has been extensively covered by Brewis (1982) and Brewis and Briggs (1985); note that many of the treatments considered in Brewis’ early work are now favored by industry and so are relevant to this book. For some polymers, a solvent wipe will suffice to remove, for example, plasticizers, which have migrated to the air interface. However, not all solvents are compatible with all polymers, so care must be taken to select the most appropriate solvent. Polymers which are difficult to bond such as polyolefins and elastomers may require the removal of such weak boundary layers and also the introduction of relevant functional groups; these are often oxygen-containing. The most favored methods to achieve this result include: flame treatment, corona discharge treatment, and plasma treatment. These are all dry processes which can oxidize a range of polymers by using excited gases, ions, neutrals, electrons, or free radicals. An illustration of the usefulness of plasma treatment to treat fluoropolymers, namely poly (vinyl fluoride) (PVF) and polytetrafluoroethylene (PTFE), has been given by Brewis et al. (1995). In this example, argon, oxygen, air, and nitrogen plasmas were used. In the case of PTFE, there was no significant change in surface chemistry observed by XPS; however, SLS joint strengths increased from 420 to 1,860 N following argon plasma treatment for 30 s. This result was explained as due to the removal of a preexisting weak boundary layer. In contrast, with PVF, XPS showed the various plasma types to increase oxygen levels in the surface from 0.8 up to 12.4 atom %, in the case of the oxygen plasma. The introduction of this level of oxygen functionality resulted in an approximate tenfold increase in initial adhesion. Similar large increases in adhesion are observed in many studies of polymer–polymer adhesion, these are explained in terms of the removal of weak boundary layers, an increase in surface functionality or free energy, or enhanced roughening of the polymeric substrate using these methods. A recent study by Stewart et al. (2005) has demonstrated the effectiveness of air plasma treatment of polypropylene as applied to complex shapes; in this case, a car bumper, which, with an optimized treatment condition, demonstrated good adhesion post treatment. As an alternative to the above-mentioned treatments, Charbonnier et al. (2001) discussed the use of excimer laser and ammonia gas treatment for the

142

G. Critchlow

functionalization of polycarbonate and polyimide. In their study, a large increase was observed using X-ray photoelectron spectroscopy (XPS) in the levels of nitrogen, from 6.4 to 17.6 atom%, with no great change in the levels of oxygen as a result of a 3 min treatment. The improved adhesion of a subsequently applied palladium complex was attributed to this increased level of nitrogen, given the affinity of palladium to form “N-Pd” bonds. Note that a consideration of plasma polymerized coatings, which is a related technology, is outside of the scope of this contribution; attention is drawn, however, to the relevant book by Strobel et al. (1994).

7.4

Mechanical Treatments and the Role of the Interphase

Mechanical treatments generally remove friable surface layers and generate macrorough surface texture on metallic surfaces. They are generally combined, before and after, with a simple degrease process. The penetration of polymers into the asperities on, or within, surfaces has been well discussed in the literature; an early study was carried out by Packham et al. (1974a) who investigated the keying of polyethylene into the pores present on anodized aluminum. The effect of such polymer penetration and the importance of mechanical interlocking with such highly textured surfaces which will facilitate energy dissipation within bondlines was further discussed by Packham et al. (1974b) and, amongst others, Rider and Arnott (2001). One important conclusion from such studies is that the stresses applied to joints containing highly rough adherends are potentially no longer concentrated at the surfaces, but are then distributed more widely within the bondline, enabling more plastic energy dissipation to occur and increased observed levels of adhesion. Photoelastic (Packham 2010) and Moiré interferometry (Asundi 1987; Sargent and Ashbee 1985) studies have directly observed the nonuniform stress distributions within bondlines and in the region of surface asperities. One advantage of creating such asperities is that stresses are then potentially dispersed into and away from the “interphasial” region. It has been acknowledged for some years now that there is a zone between the substrate and adhesive within the bondline that is different to that of the bulk adhesive. In an early study, Allen et al. (1972) noted thin glassy layers remaining on titanium alloy lap joints after testing, free from the filler and pigmentation associated with the bulk of the adhesive. Similarly, Sharpe (1972) in 1972 commented upon the fact that, at that time, the components within a bonded system were treated as discrete entities, when, in fact there are many more layers present in a bonded joint. Since this pioneering work, many other workers have begun to study the interphasial chemistry. For example, Fondeur and Koenig (1993) conducted Fourier transform infrared (FTIR) microscopy studies on aluminum substrates coated with a dicyandimide cured epoxide adhesive. They found that there was a variation of the dicyandiamide concentration with thickness through the adhesive layer. Using both the nitrile and carbonyl peaks, they found that for untreated aluminum substrates, there was a comparatively high level of dicyandimide near

7

General Introduction to Surface Treatments

143

to the substrate surface that coincided with a high level of carbonyl material. The opposite was observed for chromic acid anodized adherends, where the dicyandimide content was at a minimum close to the substrate. Similarly, Nigro and Ishider (1989) conducted FTIR reflectance absorption spectroscopy on an epoxide resin with a BF3-monoethylamine catalyst system applied to polished 1008 Drawing Quality steel substrates. They studied the degree of conversion of the oixrane (epoxide) ring as a function of depth and found that epoxy conversion rates were higher closest to the steel substrate. The epoxy prepolymer was also applied onto the steel surface without hardener and the same phenomenon was still observed. They suggested that the steel might, in some way, interact with the epoxide adhesive causing homopolymerization of the epoxide resin at the surface of the substrate. Likewise, Roche et al. (2002) studied the resultant interphase of a DGEBA epoxide when combined with either an aliphatic, aromatic, or cyclic diamine curing agent and then applied onto aluminum or titanium surfaces. They found differences through the thickness of the adhesive layer: the Tg. was lower at the interphase for both substrates and the stochiometric ratio was higher nearer to the substrate, with an amine to epoxy ratio of 1.2 as opposed to 1 within the bulk region. Placing the DGEBA and IPDA individually in direct contact with the aluminum and titanium substrates, they found that dissolution of the metal oxides occurred in the basic IPDA liquid. Taking samples of the reacted IPDA liquor and combining them with DGEBA, they measured the Tg, Young’s modulus, and the epoxide amine ratio. They compared this with the properties of the interphasial regions and bulk polymer regions. What they found was that the modified IPDA system had the same properties as the thin layer samples with reduced Tg, and differing Young’s moduli. Safavi-Ardebili et al. (1997) conducted experiments upon acid etched A1100 aluminum substrates bonded with an epoxide adhesive, Hysol EA-9346. Sectioning the joints at varying angles, they examined the bondline with EDS, they measured the interphasial region, and found that this layer was irregular in thickness ranging from 2  10–6 to 6  10–6 m. Nano indentation revealed that there was a gradient of the Young’s modulus, which increased by approximately 13% in the interphasial region when compared to the values obtained within the bulk of the bondline. Likewise, Kollek (1985) used FTIR diffuse reflectance analysis to study the absorption chemistry of epoxide and phenolic resins on aluminum substrates. He observed that both the curing agent and the epoxy resin monomer were adsorbed on the aluminum oxide surface. They found that the dicyandiamide monomer was adsorbed by the oxide layers of the substrate and was attributed to the acid proton of the aluminum oxide reacting with and reducing the observed nitrile peaks. For the epoxide monomer, less adsorption was observed, and it was suggested that this occurred by the opening of the epoxide ring. More recent studies have further added to the understanding of polymer-metal interactions and confirm the highly variable chemical nature of the polymer-metal interphase (Gaillard et al. 1994; Bistac et al. 1997; Roche et al. 2006); particular attention is drawn to the work of Possart in this area (Possart 2005). Of critical importance, the interphase comprises a region whose chemical composition might be

144

G. Critchlow

very different from the bulk polymer. The mechanical properties of the interphase might, as a consequence, be different as observed by many of the workers in this area mentioned above. The extent of the polymer-metal interphase is also the subject of some discussion with some researchers focusing attention on the polymer up to a few nanometres from the metal surface, whereas others maintain that the region of modified chemistry and properties extends up to many micrometres. What is clear is that the chemistry and resultant mechanical properties of the interphasial polymer are very different from the known bulk polymer properties; hence, the ability of this material to sustain loads and resist the initiation of damage and its reaction to moisture, all factors which might influence failure in the interphasial region, are more difficult to predict than the bulk polymer. Other complications arise when, for example, water interacts with the interphasial polymer and accelerated corrosion can be initiated. In summary, the creation of rough surfaces provides the possibility of redistribution of stresses away from the interphasial polymer and into the bulk adhesive. This, in addition to the fact that rough surface have much higher specific surface areas for adhesive interaction and texture for mechanical attachment, means that the creation of such surfaces is highly desirable. The simplest form of mechanical treatment is to hand abrade using dedicated abrasive pads or papers, followed by a solvent, or similar, wipe to help remove residues. This process may be effective but can also introduce contamination and may be nonuniform in application (Bishopp et al. 1988). On an atomic scale, the surface can retain contaminating residues from the pads or papers which cannot easily be removed. In terms of resultant texture, on the macroscale, the hand-abraded surface tends to comprise broad flat areas of unaltered surface in between the grooves created by the treatment. The ratio of the surface area exhibiting these features is clearly dependant on the degree of treatment applied. A more controlled method of surface treatment is by grit-blasting or slurryblasting followed by degreasing. Frequently, coupling agents are also applied post grit-blasting. This topic will be covered in greater detail in ▶ Chap. 11, “Organosilanes: Adhesion Promoters and Primers.” During grit-basting, the abrasive media, which usually comprises alumina, chilled iron, or silica, is propelled to the surface under pressurized air, while in slurry-blasting alumina is usually used, which is transported via high pressure water jets. Clearly, treatment parameters such as pressure, times, and blast media can be optimized to produce the desired surface texture. In one study of the grit-blasting treatment, Critchlow and Brewis (1995a) quantified the degree of enhanced roughness created using various grades of alumina from 40/60 to 320, using profilometry, and changes to the surface chemistry, using AES. With 5251 grade aluminum alloy, grit-blasting did improve adhesion, but the importance of post cleaning to remove embedded grit and to create a subsequently wettable surface was emphasized. In general, the enhanced surface features created by grit-blasting provide reasonable levels of initial adhesion, Critchlow and Brewis (1996) in a review article identified a number of studies where this was the case. However, it should be noted that the very thin oxide-remaining post grit-blasting leaves metals susceptible to corrosion, and durability in hot–wet environments can

7

General Introduction to Surface Treatments

145

be poor. Bishopp illustrated this point by immersion of aluminum 2024-T3 clad alloy joints in 85% RH at 70
C and measuring the fall-off in peel strengths with time (Bishopp et al. 1988). Under these conditions, joints incorporating grit-blasted adherends retained only 20 N/25.4 mm compared with anodizing which retained up to 180 N/25.4 min after 10 days exposure. Bishopp suggested that in some cases, surfaces can be “too rough,” allowing air to be trapped at the adhesive-metal oxide interface, i.e., there is incomplete wetting with poor durability resulting. The possibility of using power beams for surface texturing has was originally discussed by Tavakoli (1994). In this sense, the term “power beams” includes the use of laser and electron beams for the texturing of surfaces. Significant work has been carried out in this area with the establishment of the Surfi-sculpt ® and Comeld ® processes. Some details of the structures generated by electron beam bombardment have been discussed by Tu et al. (2010). Such processes look highly promising for the generation of highly complex structures up to millimeters in scale on metallic, polymeric, and composite surfaces. In summary, good initial levels of adhesion can be observed using metallic adherends post-mechanical treatment due to the ability of such surfaces to partially negate any adverse influence of the interphase and to provide desirable texture. However, the thin oxides present after such treatments tends to cause mechanically treated joints to perform badly, particularly in durability tests carried out under hot-wet conditions. The application of primers and coupling agents is usually carried out to obviate this poor performance. Also, the use of chemical treatments can be carried out if enhanced durability is required, as discussed in the following section. The use of such mechanical methods on polymers is, to some extent, limited. Mechanical treatments such as grit-blasting can introduce high stress and damage into the substrate which metals can accommodate, but some polymers cannot. However, this is not always the case and grit-blasting is routinely used to treat carbon fiber–reinforced polymer (CFRP) composites. Attention is drawn to work of Belcher at NASA who reported that grit-blasting of CFRP performed similarly to the use of peel ply or laser ablation. The importance of using optimized mechanical methods though has been highlighted by Stone (1981) in his study on the influence of abrasion on CFRP adherends.

7.5

Chemical Treatments

7.5.1

Chemical Treatments for Metallic Adherends

These are invariably aqueous-based processes which may be applied from spray of full immersion. Importantly, surfaces are generally degreased prior to chemical treatment. In practical terms, this is to reduce drag-out contamination passing from stage-to-stage and so obviating the possibility of contamination present in the final solution. Chemical treatments can be considered as etchants or, alternatively, used to produce chemical conversion coatings.

146

G. Critchlow

Etch solutions invariably utilize strong single or mixed acids, although, in some rare instances, strong alkaline solutions can be used. The purpose of such etchants is to remove friable organic or inorganic layers, including loosely bound oxides, hydrated oxides, or other corrosion products, and to leave an enhanced surface topography. In principle, with metallic adherends, after rinse operations, the surfaces should also be of a high surface energy being atomically clean. Depending upon the treatment conditions used, i.e., solution type, time, and temperature along with substrate characteristics such as the alloy type and thermal history, the resultant topography can vary from macro- to nano-rough. The nanoscale features being regarded as particularly useful to enable the creation of a “micro-composite interphase.” The benefits of such topography have been discussed above and an illustration of the commonly used Forest Products Laboratory (FPL)–etched aluminum surface is given in ▶ Chap. 8, “Surface Treatments of Selected Materials.” Many of the common acids used for the prebond treatment of metallic adherends have been reviewed by Minford (1993) in the case of aluminum alloys and Critchlow and Brewis (1995b) for titanium alloys. The aforementioned works by Minford and Critchlow indicate that, as with mechanical treatments, acid etching can produce excellent initial levels of adhesion but, again, a relatively thin oxide is present after such treatments so that durability can be compromised. Chemical conversion coating is the simplest method of producing thicker, passivating films on metal surfaces. Chemical conversion coatings develop a passivating film by means of a complex growth and deposition mechanism. The degree of roughness introduced by such conversion coatings on surfaces can be highly variable. For example, on medium carbon steel, this variability was illustrated in the work of Critchlow et al. (2000). In this study, a range of chromate-free conversion coatings were compared with chromate-containing processes. The best-performing process, Bonderite 901-Pyrene 8–90, provided a thick, mixed oxide extending to approximately 150 nm, which was highly nodular on the nanometre scale and which contained polymeric functionality in its outer layers. It was concluded that these physicochemical properties were responsible for the excellent adhesion performance: with the nodular surface capable of interphase formation, some epoxide component possibly present to react with the adhesive and good barrier corrosion protection afforded by the oxide. With this treatment combined with a single part epoxide mean initial SLS strengths of 57.9, MPa were measured. Significantly, after 8 weeks of immersion in water at 60
C, these joints retained approximately 80% of their initial joint strength. Due to the complex nature of the conversion coating process, it should be noted that although the aforementioned Bonderite 901-Pyrene 8–90 process worked well for medium carbon steel, it will not necessarily be applicable to other adherends. There are, however, many available conversion coatings some of which are detailed below. Phosphate coatings are widely used in the automotive industry due to their ability to treat both aluminum and steel. These coatings offer good levels of adhesion and, although the corrosion resistance is generally regarded as marginally lower than that of chromate-containing coatings, joints produced offer good environmental stability. A study of chromate-phosphate-based conversion coatings has been carried out by Maddison and Critchlow (1996). This study showed that the preexisting oxide

7

General Introduction to Surface Treatments

147

on aluminum 5251 alloy was removed within the first few seconds of treatment followed by a controlled film formation up to a thickness of approximately 1,000 nm with extended treatment times up to 60 s. The resultant film composition was highly complex comprising Al, O, P, and Cr. Importantly, with reasonable treatment times, the chromate–phosphate treatment demonstrated excellent durability with a single part epoxide in impact testing carried out on SLS joints which had been immersed in with water at 75
C prior to testing. In this experiment, the chromate-phosphatetreated joints absorbed over five times more energy than alkaline-cleaned controls. Underhill and Rider (2005) propose an interesting alternative: a simple hot-water conversion coating. They have shown this method to have understandably low processing costs as well as suggesting successful adhesive performance. The same group has also demonstrated the effectiveness of boiling water treatment–applied post grit-blasting and subsequently followed by epoxide-functional silane application (gamma-GPS). Molybdenum-based coatings have been considered as alternatives due to molybdenum having an electronic structure similar to that of chromium. Molybdates are known to be corrosion inhibitors. When incorporated within aluminum oxide films, the molybdate (MoO42) is regarded as providing excellent corrosion resistance to pitting. The high cost of molybdenum conversion coatings is a possible explanation for their lack of popularity. Of particular note are the titanium-zirconium-based solutions which are the basis of many commercial conversion coating systems due to their ease of application. These coatings, for example, are widely used to help attain good lacquer adhesion in the canning industry. Titanium-zirconium coatings produce thin films in very short application times, in the order of a few seconds, and good adhesion and corrosion performance is reported for a range of aluminum alloys. On some aluminum alloys, namely 6060, preferential deposition has been found to cause localized variations in film thickness which, when coupled with the fact that these processes produce colorless films, makes quality control assessments very difficult, however. For example, Schram et al. (1995) showed such conversion coatings to extend to about 10 nm in thickness independent of treatment time. Using AES, the same workers identified a duplex structure to the coating with a simple oxide present adjacent to the aluminum 1050 alloy substrate, but a more complex aluminum oxide-zirconium fluoride plus polymer present in the outer nanometres of the film. Critchlow and Brewis (1997) also studied such films on aluminum and confirmed the highly complex, duplex nature of the resultant films. These workers demonstrated that static-stressed SLS joints immersed in water at 60
C had much greater times to failure compared with degreased-only or grit-blasted controls. Subsequent studies have shown the bestperforming conversion coatings, on aluminum alloys, at least match the durability performance of the best acid etches, such as the FPL. However, the use of hot-water immersed static SLS joints has demonstrated that titanium-zirconium conversion coatings to be less durable than the phosphoric acid anodizing (PAA) treatment. Cerium-based conversion coating processes are thought to precipitate Ce (III) hydroxide at regions of high pH. This precipitation gives the coating corrosion inhibiting properties similar to that of titanium-zirconium-based coatings. The

148

G. Critchlow

inhibition is believed to decrease the rate of cathodic oxygen reduction. A multistage process combining chemical and electrochemical treatments based on cerium and molybdenum has been studied and is shown to enhance the corrosion protection in the high-copper-content aluminum alloys, for example, the aerospace 2024 and 7075 alloys. The combined process offers good levels of adhesion on the 2xxx series alloys but mixed results are obtained on other alloys. One disadvantage of this process is the extended time taken to produce the coating, which increases production costs. In summary, the chemical treatments when applied to metals do remove preexisting friable oxides or hydrated oxides in the early stages of the treatment followed by controlled film formation. The films can offer the advantages of increased surface free energy or, in the case of conversion coatings, the possibility of polymeric functionality introduced. The conversion coating processes, in particular, can provide corrosion protection via a number of mechanisms; the result of which is the observed increase in durability of such treated adhesive joints exposed to hot-wet conditions. A range of surface morphologies can be generated by both etch and conversion coating processes, and techniques such as electron and atomic force microscopy (AFM) can be used to monitor these treatments to provide optimized surface texture which will facilitate interpenetration of the polymer into the asperities of the treated surface. Whilst, chemically treated surfaces do offer a range of physicochemical properties which enhance adhesion, the electrochemical processes generally provide the best bonding surfaces for aluminum and titanium; these are discussed in the following section. Note that anodizing of steels is not possible.

7.5.2

Chemical Treatments for Polymeric Adherends

In addition to the dry processes detailed previously, wet chemical processes can also be used to modify polymeric surfaces to introduce desirable physicochemical properties. For example, chromic acid has been used as an oxidizing agent to this end for some time (Brewis and Mathieson 2001). Given the concerns over the usage of hexavalent chromium in this field, there are ongoing efforts to find suitable replacements. In one study by Brewis et al. (1995), a proprietary etchant, Tetra-Etch, was used to treat two fluorinated polymers, PTFE and PVF. The results of this treatment are given in Table 5. XPS showed Tetra-Etch caused the PTFE surface to lose almost all of its surface fluorine with a short treatment time of 10 s. This was accompanied by a large increase in the surface oxygen levels from zero to 11.6 atom%. This increase in useful oxygen-containing functional groups enabled a tenfold increase in measured adhesion in the SLS configuration. The PVF, however, required longer treatment times with the best results observed after 60 min immersion in Tetra-Etch. Again, a large increase in adhesion was observed from 360 to 3,020 N, but in this case, there was relatively little surface modification observed by XPS. This result was attributed to the removal of a potential weak boundary layer by the Tetra-Etch treatment. Similarly, Dahm et al. (2005) studied a range of wet chemical treatments, including various oxidizing and halogenating agents on both additive-free and fully formulated

7

General Introduction to Surface Treatments

149

Table 5 To show surface compositions and SLS joint strength of PTFE and PVF polymers following Tetra-Etch treatment for various times (Brewis et al. 1995) Polymer/treatment PTFE none PTFE 10 s PTFE 1 min PVF none PVF 10 s PVF 1 min PVF 60 min

XPS (atom%) C 38.4 37.6 82.2 70.4 72.4 75.4 87.3

F 61.6 0.8 0.9 28.8 26.7 23.0 11.4

O – 11.6 16.9 0.8 0.9 1.6 1.3

Failure load (N) 420 4,280 4,260 360 800 2,080 3,020

elastomers. Of particular interest were styrene butadiene (SBS) and ethylenepropylene-dimer terpolymers (EPDM). In this study, extensive use was again made of XPS to relate changes in surface chemistry to SLS strength determined using a composite SLS joint. As with polyolefins, elastomers proved highly receptive to oxidizing agents, showing a large increase in surface oxygen content with, for example, potassium permanganate, potassium dichromate, Fenton’s reagent, and other such treatments. In general, however, the oxidizing agents provided only a slight increase in observed adhesion; this being attributed to the production of a layer of low cohesive strength at the surface of the elastomers. In particular, static secondary ion mass spectrometry (SSIMS) was used to study the SBS surface following potassium permanganate treatment, and it could be inferred that this treatment did cause some chain scission of the elastomer surface. In contrast, a number of chlorinating agents, including sodium hypochlorite solution, Chloramine-T, and trichloroisocyanuric acid (TCICA) were studied, and these also created highly modified SBS and EPDM surfaces. With these treatments, however, a large increase in adhesion was observed, by a factor of approximately x4 to x20, depending upon the specific treatment on SBS. The mechanisms by which this was achieved have been discussed by Dahm et al. (2005). In summary, wet chemical treatments can be highly effective on a range of polymers by the removal of weak layers and the introduction of desirable chemistry. As with dry processing of polymers, there is always the potential for over-treatment, causing chain scission and a net reduction in adhesion levels. For this reason, such treatments should always be optimized for a specific application.

7.6

Electrochemical Treatments

7.6.1

Introduction to Anodizing

Regarding the higher treatments, and the anodizing processes in particular, these can be applied to a number of metals but are mostly carried out on aluminum and titanium and their alloys. These treatments are in the main complex, time consuming, and costly to carry out. There are also legislative drivers which make the use of some

150

G. Critchlow

processes, especially those which utilize hexavalent chromium, highly undesirable. Other factors such as energy and chemical disposal costs also deserve consideration. However, for structural metal bonding, these treatments are highly recommended as they impart all of the fundamental physicochemical properties required to provide the best initial levels of adhesion and durability of any treatments, namely, wettability, micro- and nano-roughness, mechanical stability, and passivation. Note that to take full advantage of the porosity present on anodic oxides, it is often recommended that primers are used in combination. There exists extensive literature on this topic and attention is drawn to the highly relevant book by Minford (1993) and to ▶ Chap. 8, “Surface Treatments of Selected Materials” of this handbook. It is widely reported, and further discussed in ▶ Chap. 8, “Surface Treatments of Selected Materials,” that the most successful treatments for aluminum and titanium include direct current (DC) anodic oxidation in chromic, phosphoric, or sulfuric acid electrolytes (CAA, PAA or SAA). It is, of course, recognized that other electrolytes are used, some of which will be discussed later in this section. The Bengough–Stuart 40/50 V CAA process is currently the preferred pretreatment of European aerospace manufacturers, and on clad aluminum alloy produces a porous surface oxide typically 2–3 μm thick with a pore diameter of approximately 25–35 nm. This oxide offers possibly the best levels of adhesion and environmental resistance. PAA, as a prebond process, is preferred by Boeing Commercial Airplane Company. Boeing claim good in-service durability performance with adherends using this pretreatment. The oxide produced is described as being 400–800 nm thick incorporating 100 nm of small outwardly protruding fibrils. The phosphate content of the oxide improves stability toward water and therefore improves the hydration resistance as discussed in the following section. PAA is considered to be very tolerant to variations in processing parameters compared to CAA. PAA is claimed by Boeing to provide the most durable joints but, in some studies, it is shown to be comparable to or even worse than CAA. Sulfuric acid anodizing (SAA) has also been widely compared to CAA, PAA, and FPL etches in the review by Critchlow and Brewis (1996). As indicated by Critchlow and Brewis, SAA prepared joints usually demonstrate moderate initial joint strength and relatively poor long-term joint durability. In general, SAA-treated adherends are significantly out-performed by CAA and PAA but perform similarly to FPL etched adherends. The oxide thickness associated with SAA is of the order of 10 μm and is significantly thicker than oxides produced during other anodizing processes. For this reason, it is ideally suited to corrosion protection applications. SAA followed by a phosphoric acid dip has been shown to improve joint performance by the opening up of the otherwise highly dense porous oxide. Alternating current (AC) anodizing produces a similar but thinner surface oxide structure to the aforementioned direct current (DC) treatments, but with the advantage of the adherend being electrolytically deoxidized on the negative half of the applied voltage cycle. This reduces the importance of the preceding degrease and alkaline cleaning usually associated with conventional DC anodization. This, however, comes at a cost. The solution–adherend resistance is much less on the cathodic half of the cycle than it is on the anodic half. Consequently, the current density is

7

General Introduction to Surface Treatments

151

much larger and can potentially cause an undesirable effect known as “burning” or “powdering” on the surface. A compromise must be struck between treatment time, temperature, and applied voltage in order to minimize this effect. The thinner resultant surface oxide and limited applications are consequences of this. The two main electrolytes, indicated in the literature as being used for AC anodizing, are phosphoric acid and sulfuric acid. A number of workers have investigated both electrolytes and found promising results indicating that AC anodizing was suitable as a prebond process. Further details of this treatment, including a consideration of the mechanisms of anodic film formation, follow.

7.6.2

Structure and Composition of Anodic Films

It has been accepted for some time that an anodic oxide formed on the surface of an aluminum alloy typically consists of an inner, thin, compact oxide layer, called the barrier layer, and a porous outer oxide displaying a honeycomb-like hexagonal close-packed structure, as illustrated in Fig. 1. The thickness of the barrier layer is approximately 0.1–2.0% of the overall film thickness and is dependent upon the electrolyte concentration and the applied anodic voltage. Several authors propose the dependency to be close to 0.01 nm/V. It is suggested that the idealized honeycomb arrangement is due to repulsive interactions between the pores during anodic film growth, and that a mechanical stress resulting from the volume expansion of the forming oxide causes these repulsive forces. In doing so, the formation of a hexagonal arrangement during nucleation occurs. Production rates of between 1 and 5 μm/h are typical. Factors such as the alloy type, the thermal history of the metal, deoxidizing process used, and electrolyte agitation can be a very important factor in the formation of ordered pore arrays. When an anodic voltage is applied to aluminum in an electrolyte, a current will flow through the electrolyte to the anode and aluminum oxide or hydrayed aluminum oxide will be produced on the surface. If this oxide is soluble or partially soluble in the electrolyte, then dissolution will occur creating the porous oxide layer. The porosity of the oxide will therefore be a function of the solubility of the oxide in the electrolyte, and the film will only grow if this rate of dissolution is less than the rate of oxide production. An important realization is that the oxide grows from the bottom of the film rather than being deposited on the outer surface as in electrodeposition. For this reason, it is pivotal in the understanding of porous oxides that the outer surface, being formed first, has been in contact with the electrolyte for the longest time and may have been notably modified. The growth of the oxide film is reported to be made possible by the solid-state migration of the Al3+ and O2 ions which move outward and inward, respectively, through the film. This movement is driven by the electric field created during anodization. The development of the new material at the metal/film interface is due to oxidation of aluminum atoms to form Al3+ combining with the O2 ions present in the electrolyte. This migration results in new material being formed in two places, the metal/film and the film/electrolyte interfaces. The second phase elements

152

G. Critchlow Pore diameter 25nm

Cell wall thickness 10nm

Oxide thickness 4 µm

Barrier layer

Base Metal

Fig. 1 To show the idealized cell structure of a porous aluminum oxide grown post chromic acid anodization

present on the alloy surface are either removed or oxidized along with the aluminum. A further mechanism, field-assisted dissolution, may occur due to the of the electric field concentration across the barrier layer which is thermally enhanced due to localized Joule heating and the aforementioned ionic mobility. This effect causes the aluminum to oxygen bonds in the oxide lattice to weaken, resulting in a dissolution at the oxide/electrolyte interface. The chemical and field-assisted dissolution take place with widely differing rates, with the field-assisted being reportedly up to 3,000 times greater. Note that the electrolytically driven growth competes with these dissolution mechanisms to produce the final anodic film under steady-state conditions. The resultant surface features associated with the anodizing are further detailed in ▶ Chap. 8, “Surface Treatments of Selected Materials.” It is, however, sufficient to state that the microscopically rough features produced can increase the specific surface area of an anodized adherend by a factor of several hundred. It is widely accepted that wetting by an adhesive, or primer, into a porous anodized structure such as the PAA oxide or the CAA oxide is possible and is necessary for the formation of a strong bond via the formation of a “micro-composite” region due to this penetration of adhesive into the oxide. The microstructure generated by the CAA process on both bare and clad aluminum alloy is illustrated in Fig. 2. Note the non-idealized structure which is typical of that on the higher aluminum alloys.

7

General Introduction to Surface Treatments

153

200nm

a

c

200nm

b

200nm

200nm

d

Fig. 2 CAA 40/50 V processed, 2024-T3 bare; plan view (a) and cross-section (b), 2024-T3 clad; (c) plan view and (d) cross-section

7.6.3

Physical Factors Affecting Anodic Oxides

Physical properties such as the temperature of the electrolyte and the anodizing voltage have been shown to have a marked effect on the anodic oxide film formation. For example, an increase in local temperature due to convection heat transfer enhances the field-assisted dissolution at the pore bases and results in an increase in current density. The anodizing voltage has been shown to affect most of the overall film properties such as the barrier layer thickness, cell wall thickness, pore diameter, and the overall thickness of the oxide layer. The electrolyte temperature also has an effect on these properties due to its influence over the dissolution power of the solution. Increasing the temperature produces thinner porous and barrier layers, larger pore diameters, and thinner cell walls. Reducing the temperature has the opposite effects. The alloy used also affects the oxide composition. The copper rich 2xxx series alloys adversely affect the film morphology and integrity. The oxidation of copper and the inclusion of Cu 2+ in the anodized alumina is associated with O2 evolution which causes localized film cracking and can lead to nonuniform current flow through the film.

7.6.4

Alternatives to Phosphoric and Chromic Acid Anodizing

As detailed above, anodizing of aluminum and titanium alloys provides a surface that is corrosion resistant and which may be receptive to applied organic layers. Such

154

G. Critchlow

functionality is achieved by engineering a hydrated oxide film with controlled nanometre-scale porosity within an adherent coating measuring typically 2–3.5 μm in thickness. The 20–30 nm diameter (Ф) pores created during chromic acid anodizing (CAA), for example, enable interphase formation with the organic layer giving excellent adhesion, while the oxide present provides barrier corrosion protection. There is, however, a conflict between the requirement for good corrosion resistance, which dictates a compact pore structure with thick cell walls and small diameters, and adhesion, which requires good interlocking and a large pore diameter. The former case is satisfied by the sulfuric acid anodizing (SAA)–based processes with Ф  5 nm, and the latter by phosphoric acid–based processes, where Ф 50 nm. The CAA process, in some respects, provides a compromise; it is, however, very widely used but for a number of reasons drop-in replacements are being sought (Critchlow et al. 2006). One option is the use of a more benign electrolyte, such as boric-sulfuric acid. In one study on boric-sulfuric acid anodizing (BSAA), Critchlow et al. (2006) modified a standard process by varying the deoxidizing stage prior to anodizing, and introduced a post-treatment stage in order to optimize structural adhesion between a range of aluminum alloys and a single part epoxide. It was shown that BSAA could provide all of the physicochemical parameters required for excellent adhesion being capable of replacing CAA for adhesively bonded components. The conclusions from the study by Critchlow et al. (2006) were that increasing the temperature of the electrolyte from room temperature to 35
C increased the pore diameter of the BSAA oxide film, which allowed more complete surface wetting by the adhesive/primer and maximized the pore penetration depth of the primer. This resulted in joints which replicated CAA for environmental durability and adhesion performance. Importantly, however, duplex structures offer the advantages of a large pore diameter in the outer layers of the anodic oxide and very small pores in the inner layer close to the metal to give the optimum adhesion and corrosion performance. There are a number of methodologies which might be used to achieve a resultant duplex structure. Yendall (unpublished work) has investigated the use of an electrolytic deoxidizing process (EPAD) followed by a conventional SAA. Figure 3 shows

1µm

EHT = 5.00 kV WD = 5 mm

Signal A = InLens Photo No. = 2042

300nm

EHT = 5.00 kV WD = 5 mm

Signal A = InLens Photo No. = 2043

Fig. 3 (a) Low-magnification FEGSEM image to show the through thickness duplex structure of the EPAD plus SAA oxide on clad 2024-T3 aluminum, and (b) the surface region at high magnification

7

General Introduction to Surface Treatments

155

the structure formed using an optimized EPAD plus SAA process. A number of process variables, for example, immersion times and temperatures plus acid electrolyte types and concentrations were considered to achieve this optimized structure. 2024-T3 clad aluminum alloy bonded with the BR127/FM73 primer/adhesive system (Cytec) exposed to 60
C and tested using the standard wedge test configuration showed comparable results, with optimized processes, to CAA-treated controls. In contrast, the SAA only treated aluminum alloy performed very unfavorably in this test due to the lack of primer penetration; see Fig. 4. Other electrolytes have also attracted much recent attention such as the tartaric-sulfuricbased process. Although DC anodizing does provide many of the desirable features required for optimized bonding, alternating current (AC) or combined AC then DC combination processes do offer some further advantages and will be discussed in the following sections.

7.6.5

Alternating (AC) Current Anodizing

Apart from the obvious use of alternating current, the main difference between AC and DC anodizing is the increased porosity and ductility of AC films. The formation of the oxide is essentially the same as for direct current anodizing with the exception of the extra time during the cathodic half-cycle when extra fieldassisted dissolution can take place, this being the cause of the increased porosity. Fortunately, this dissolution reaction is not as efficient as the anodic production reaction and dissipates much of the energy input by means of hydrogen evolution and heat. This results in an overall film thickening during AC anodizing. The proposed reactions for AC anodizing in a sulfuric acid electrolyte are given below [1] and [2]:

100 hours 24 hours 5 hours Io

70

60

Crack Ext (mm)

50

40

30

20

10

0 CAA 40/50V

(NaOH)+ SAA 40g/l 26°C

(EPAD)+ SAA 40g/l 26°C

(EPAD)+ SAA 40g/l 35°C

(EPAD)+ SAA 40g/l 35°C + (PAD)

(EPAD)+ SAA 180g/l (EPAD)+ SAA 180g/l 20°C 35°C

Fig. 4 To show the results of wedge testing with clad 2024-T3 aluminum exposed to deionized water at 60
C

156

G. Critchlow

• Anodic reactions [1]: – Film formation: 2Al + 6OH ! Al2O3 + 3H2O + 3e – Film dissolution: Al ! Al3+ + 3e – Oxygen evolution: 4OH ! 2H2O + O2 + 4e – Sulfate oxidation: 2SO42 ! S2O82 + 2e • Cathodic reactions [2]: – Hydrogen evolution: 2H+ + 2e ! H2 – Reduction of sulfate: SO42 + 8H+ + 6e ! S + 4H2O or SO42 + 10H+ + 8e ! H2S + 4H2O – Alumina reduction: Al2O3 + 6H+ + 6e ! 2Al + 3H2O – Aluminum deposition: Al3+ + 3e ! Al There is a tendency during AC anodizing for compounds from the electrolyte to be deposited within the anodic film, this can be an advantage if, for example, phosphates ions are incorporated, for reasons discussed in the following section. The oxide morphology of an AC anodized surface does not differ widely from the oxide produced using a DC counterpart when using either sulfuric or phosphoric acid electrolytes. The oxide retains the columnar porous structure with the only differences being the reduced thickness and increased pore diameter due to the extra dissolution time allowed on the cathodic part of the voltage cycle. The overall thickness of the oxide is dependent on the electrolyte composition, but more importantly on the electrolyte temperature as this has, by far, the greatest effect in increasing the dissolution rate of the oxide. Specifically, with electrolytes comprising of sulfuric acid, there are two main processing difficulties; the discoloration of the oxide due to the deposition of sulfur compounds and the difficulty in growing a thick oxide layer without burning the surface. AC anodizing in hot sulfuric and hot phosphoric acids gives clear advantages compared with conventional DC treatments. During the cathodic half-cycle, hydrogen evolution produces a cleaning effect of the surface of the substrates, possibly reducing or even eliminating the need for multistage degreasing and deoxidizing steps. The resultant anodic films are produced in treatment times typically in the order of 10 to 60 s. Hot AC SAA is typically conducted in an electrolyte of 15% sulfuric acid at a temperature of around 80
C, whereas hot AC PAA has parameters of 10% acid concentration and 50
C. The current densities for both these processes are held at 103 A/m2 rms. Work by Johnsen et al. (2004) has shown AC anodizing to be comparable in mechanical performance to DC anodizing in equivalent electrolytes.

7.6.6

Alternating Current–Direct Current Anodizing (ACDC Anodizing)

The duplex structure created by first AC anodizing aluminum followed by a DC anodize was optimized by Cartwright (2005). Figure 5 shows the evolution of

7

a

c

General Introduction to Surface Treatments

200nm

EHT = 5.00 kV WD = 5 mm

Detector = InLens Photo No. = 251

Date :1 Apr 2004

–5s

200nm

EHT = 5.00 kV WD = 5 mm

Detector = InLens Photo No. = 264

– 120 s

157

b

Date :1 Apr 2004

d

200nm

EHT = 5.00 kV WD = 4 mm

Detector = InLens Photo No. = 258

Date :1 Apr 2004

Detector = InLens Photo No. = 267

Date :1 Apr 2004

– 30 s

200nm

EHT = 5.00 kV WD = 5 mm

– 240 s

Fig. 5 (a–d) FEGSEM micrographs showing the development the AC anodized 2024-T3 clad aluminum alloy surface with application periods ranging from 5 to 240 s in the phosphoric/sulfuric acid electrolyte

the honeycomb-like porous oxide created by the AC phase. Figure 6 shows the fully formed optimized ACDC structure in cross-section at various magnifications. A range of tests have been conducted using 2xxx (clad and bare), 5xxx, 6xxx, and 7xxx (clad and bare) aluminum alloys. These tests include measurement of both adhesion and corrosion performance. In all cases, with all alloys, the ACDC process performed favorably compared with either SAA or CAA controls. Examples of wedge test data are presented by Cartwright (2005); see Table 6 which compares wedge test data for various ACDC parameters against the 40/450 V Bengough–Stuart CAA process. Note, however, that peel, butt, lap-shear (static exposed and fatigue) tests have also been carried out. All tests indicating comparable levels of adhesion between ACDC and CAA treated alloys (Shenoy 2009). Similarly, linear polarization and other corrosion studies have shown that the ACDC process gives comparable corrosion resistance to the CAA. Importantly, the ACDC process is carried out in a weak acid electrolyte with only a 5% total acid content and containing no hexavalent chromium. Furthermore, the total processing time for ACDC is approximately one tenth of that required to anodize to the standard CAA specification. These advantages make the ACDC anodizing process of industrial interest (Cartwright 2005).

158

G. Critchlow

Fig. 6 FEGSEM images to show the structure of optimized ACDC oxides on clad 2024-T3 clad aluminum. Top image shows the through thickness, middle image the outer oxide, bottom image the inner oxide

1µm

EHT = 5.00 kV WD = 3 mm

Signal A = InLens Photo No. = 2898

200nm

EHT = 5.00 kV WD = 3 mm

Signal A = InLens Photo No. = 2901

200nm

EHT = 5.00 kV WD = 3 mm

Signal A = InLens Photo No. = 2899

7

General Introduction to Surface Treatments

159

Table 6 Wedge crack length data to show the influence of varying the AC anodizing parameters and electrolyte temperature for the ACDC process; note a 10 m, 20 V DC process has been deposited beneath each AC film (Cartwright 2005) AC anodizing parameters 10 V 120 s 35
C 10 V 120 s 50
C 10 V 240 s 35
C 10 V 240 s 50
C 15 V 120 s 35
C 15 V 240 s 35
C 40/50 V CAA

7.7

Exposure time (h) 1 Mean crack 33.25 length (mm) Σ 0.82 Mean crack 31.33 length (mm) Σ 0.37 Mean crack 31.33 length (mm) Σ 0.56 Mean crack 33.50 length (mm) Σ 0.19 Mean crack 31.33 length (mm) Σ 0.82 Mean crack 31.92 length (mm) Σ 0.29 Mean crack 31.50 length (mm) Σ 0.95

4 33.25

8 33.25

24 35.00

48 35.38

72 35.50

96 35.50

0.82 31.50

0.82 31.58

0.74 32.83

0.75 33.33

0.54 33.33

0.54 33.33

0.41 31.42

0.45 31.42

1.07 33.00

1.18 33.33

1.18 33.42

1.18 33.42

0.69 33.67

0.69 33.67

0.79 34.50

0.95 34.83

0.99 35.25

0.99 35.42

0.25 31.67

0.25 32.25

0.84 33.08

0.45 33.25

0.69 33.25

0.91 33.25

1.25 32.17

2.07 32.58

1.65 33.58

1.57 34.08

1.57 34.25

1.57 34.92

0.56 33.10

0.75 33.35

1.11 33.90

0.90 34.45

1.11 34.45

1.41 34.45

1.09

0.84

1.37

1.90

1.90

1.90

Conclusions

A large range of processes exist which might be used to modify the surface of adherends prior to the application of an adhesive. All such treatments modify the surface physical and chemical properties to one degree or another. Important surface properties which might be modified include: wettability, roughness, mechanical, and electrochemical stability. The last point is particularly significant if joints are to experience environmental ageing under corrosive conditions. The requirements of a particular treatment depends upon the demands put upon the joint; a simple degrease operation may suffice, and, alternatively, a more complex process may be required with well-defined surface topography and chemistry to meet the demands of structural bonding. Importantly, legislative drivers intended to reduce operator health and safety risks or environmental impact are currently in place which promote the need for more acceptable surface treatments such that the range of options available today may be reduced in the near future. Extensive efforts are, of course, ongoing to help develop

160

G. Critchlow

drop-in replacements particularly for the high-performance chromate-containing treatments. A number of drop-in replacements do, however, exist today which provide all of the functionality of the chromate-containing treatments and are likely to see further industrial usage.

References Allen K, Allen HS et al (1972) Faraday special discussion. Chemical Society, London, p 38 Anderson G (1993) J Adhes 41:129 Ansanfar A, Critchlow G et al (2009) Rubber Chem Technol 82(1):113 Asundi A (1987) Int J Adhes Adhes 7(1):39 Averill A, Ingram J et al (1998) Trans Inst Met Finish 76(3):81 Bedwell K, Critchlow G et al (1998) Trans Inst Met Finish 76(5):203 Bishopp J, Kim E et al (1988) J Adhes 26:237 Bistac S, Vallat M et al (1997) Appl Spectrosc 51(12):1823 Brewis D (1982) Surface analysis and pretreatment of plastics and metals. Applied Science Publishers, London Brewis D, Briggs D (eds) (1985) Industrial adhesion problems. Orbital Press, Oxford Brewis D, Mathieson I (2001) Adhesive bonding to polyolefins. RAPRA Rev Report 12(11) Brewis D et al (1995) Int J Adhes Adhes 15:87 Brewis D et al (1999) Int J Adhes Adhes 19:253 Cartwright T (2005) PhD thesis, Loughborough University Charbonnier M, Romand M et al (2001) J Adhes 75:381 Critchlow G, Brewis D (1995a) Int J Adhes Adhes 15(3):161 Critchlow G, Brewis D (1995b) Int J Adhes Adhes 15(3):173 Critchlow G, Brewis D (1996) Int J Adhes Adhes 16(4):255 Critchlow G, Brewis D (1997) J Adhes 61:213 Critchlow G, Cottam C et al (1997) Int J Adhes Adhes 17:143 Critchlow G, Webb P et al (2000) Int J Adhes Adhes 20:113 Critchlow G, Yendall K et al (2006) Int J Adhes Adhes 26:419 Dahm R, Brewis D et al (2005) J Adhes 81:59 Debski M, Shanahan M et al (1986) Int J Adhes Adhes 6(3):145 Deffeyes J, Lilenfield H et al (1997) SAMPE J 33(1):58 Fondeur F, Koenig J (1993) J Adhes 43:236 Gaillard F, Romand M et al (1994) J Adhes 46:227 Hong S, Boeno F (1995) J Adhes 49:133 Johnsen B, Lapique F et al (2004) Int J Adhes Adhes 24:153 Kollek H (1985) Int J Adhes Adhes 5:75 Litchfield R, Critchlow G et al (2006) Int J Adhes Adhes 25(5):295 Lu Y-F, Aoyage Y (1994) Jpn J Appl Phys 33(2B):430 Maddison A, Critchlow G (1996) J Adhes 55:273 McNamara D, Venables J et al (1979) In: Proceedings of the national SAMPE conference, p 740 Minford J (1981) Aluminium 57(10):657 Minford J (1993) Handbook of aluminum bonding. CRC Press, Boca Raton Nigro J, Ishider H (1989) J Appl Polym Sci 38:2191 Olsson-Jacques C, Wilson A et al (1996) Surf Interface Anal 24:569 Packham D (2010) J Adhes 86:1231 Packham D, Bright K et al (1974a) J Appl Polym Sci 18:3237 Packham D, Bright K et al (1974b) J Appl Polym Sci 18:3249 Possart W (ed) (2005) Adhesion: current research and applications. Wiley, Weinheim

7

General Introduction to Surface Treatments

161

Rance D (1985) Thermodynamic approach to adhesion problems. In: Brewis D (ed) Industrial adhesion problems. Orbital Press, Oxford Rider A, Arnott D (2001) J Adhes 75:203 Roche A, Bouchet J et al (2002) Int J Adhes Adhes 22:431 Roche A, Aufrey A et al (2006) J Adhes 82(9):867 Safavi-Ardebili V, Sinclair A et al (1997) J Adhes 62:93 Sargent J, Ashbee M (1985) J Adhes 18(3):217 Schram T, Goeminne G et al (1995) Trans Inst Met Finish 73(3):91 Shanahan M (2005) In: Packham D (ed) Handbook of adhesion, 2nd edn. New York, Wiley Sharpe L (1972) J Adhes 4:51 Shenoy V (2009) PhD thesis, Loughborough University Smith T, Crane R (1980) In: Proceedings of the national SAMPE symposium, p 25 Smith B, Brewis D et al (2000) Trans Inst Met Finish 78(2):56 Stewart R, Goodship V et al (2005) Int J Adhes Adhes 25:93 Stone M (1981) Int J Adhes Adhes 5(1):271 Strobel M et al (eds) (1994) Plasma surface modification: relevance to adhesion. VSP, Utrecht Tavakoli S (1994) Assem Autom 14(4):36 Tu W, Pihua W et al (2010) In: Proceedings of the world congress on engineering, London Underhill P, Rider A (2005) Surf Coat Technol 192:199

8

Surface Treatments of Selected Materials Guy D. Davis

Contents 8.1 Introduction: Requirements of Surface Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.1 Surface Pretreatment Steps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.2 Degradation Processes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.1.3 Processibility Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2 Specific Metal Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.1 Aluminum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.2 Titanium . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.2.3 Steel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3 Other Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.1 Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.2 Plastics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.3 Ceramics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.3.4 Wood . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

164 165 167 173 174 174 181 184 188 188 189 190 191 192 193

Abstract

Proper treatment of an adherend surface is one of the most important factors in assuring high initial strength and extended durability of high-performance adhesive joints. There are several requirements for a good surface preparation: (1) The surface must be cleaned of any contamination or loosely bound material that would interfere with the adhesive bond. (2) The adhesive or primer must wet the adherend surface. (3) The surface preparation must enable and promote the formation of chemical and/or physical bonds across the adherend/primer–adhesive interface. (4) The interface/interphase must be stable under the service conditions for the lifetime of the bonded structure. (5) The surface formed by the treatment must be reproducible. G. D. Davis (*) ElectraWatch, Inc., Linthicum Heights, MD, USA e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_8

163

164

G. D. Davis

In this chapter, high-performance surface treatments for several metals and other materials are discussed. Surface treatments of aluminum and other metals are used to illustrate how proper surface preparations meet these requirements.

8.1

Introduction: Requirements of Surface Treatments

Proper treatment of an adherend surface is one of the most important factors in assuring high initial strength and extended durability of high-performance adhesive joints (Landrock 1985; Kinloch 1987; Clearfield et al. 1990; Minford 1993; Pocius 1997; Chaudhury and Pocius 2002; Davis 2003). Surface general principles, treatments, and assessments are treated in ▶ Chaps. 7, “General Introduction to Surface Treatments,” ▶ 9, “Surface Characterization and Its Role in Adhesion Science and Technology,” and ▶ 10, “Use of Surface Analysis Methods to Probe the Interfacial Chemistry of Adhesion” respectively. In the present chapter, surface treatments of selected materials are given. In the case of a structural bond, a strong adhesive is of little value unless the stresses can be transferred via the adhesive from one adherend to the other. The interface or interphase (a three-dimensional interfacial region with properties different from those of either material) between the adhesive and the adherend is critical to this stress transfer. The goal of a surface treatment is to form a strong and stable interface or interphase that is stronger and more durable than the adhesive being used so that joint failure is cohesive within the adhesive, both initially and throughout the joint’s service lifetime. There are several requirements for a good surface preparation: (1) The surface must be cleaned of any contamination or loosely bound material that would interfere with the adhesive bond. (2) The adhesive or primer must wet the adherend surface. (3) The surface preparation must enable and promote the formation of bonds across the adherend/primer–adhesive interface. These bonds may be chemical (covalent, acid–base, van der Waals, hydrogen, etc.) or physical (mechanical interlocking) in nature. (4) The interface/interphase must be stable under the service conditions for the lifetime of the bonded structure. A stable interface/interphase is critical for structures expected to be in service for a long time, e.g., aircraft, but is less critical for structures whose lifetime is expected to be short, e.g., some packaging, or if the bond will be protected from the environment over its lifetime. (5) The surface formed by the treatment must be reproducible. Reproducibility requires that the prepared surface be independent of material variables, such as surface contamination, mill scale, and, ideally, even alloy or heat treatment. Reproducibility is facilitated if the treatment has wide processing windows so that processing variables, such as pH, solution contamination, solution chemistry, temperature, and time, do not have to be narrowly controlled. The robustness provided by the wide processing windows increases reliability and decreases costs associated with stringent process controls and parts scrapped due to out-of-specification processing. Different industries assign different roles to surface preparation. The aerospace industry, with relatively low throughput, but long expected lifetime (several decades), carefully prepares the surface prior to bonding during original manufacture

8

Surface Treatments of Selected Materials

165

and repair to assure the best practical durability. The automotive industry, with high throughput, prefers almost no surface preparation and uses contamination-tolerant adhesives and overdesign of joints. The construction industry follows a similar approach. The packaging industry, with very high throughput but relatively low requirements for stress and durability, also minimizes surface preparation by using relatively pure aluminum that reportedly can be more easily wetted and bonded by adhesives. This chapter will focus on surface treatments used for high-performance bonds, primarily in the aerospace industry.

8.1.1

Surface Pretreatment Steps

Cleaning the surface of undesirable organic and inorganic material is the first step in preparing an adherend for bonding. The organic contamination originates from rolling oils at the mill, oils and greases applied for corrosion protection, machining oils and release agents used during fabrication, and contamination from the atmosphere. Unless the adhesive is specifically designed to absorb this material, such contamination will prevent wetting and formation of strong bonds across the interface and the joint will perform poorly. The amount of acceptable contamination varies and can be very small. Silicones and other mold-release agents are designed to be very effective at preventing adhesives and primers from wetting the surface. Even small amounts can cause bond failure. Aluminum surfaces with one third to one molecular layer of silicone will not bond well. Unfortunately, silicones are very mobile and can readily migrate during solvent cleaning, transfer from tooling or handling, or diffuse during curing operations at elevated temperatures. Silicones and silicone-containing materials, such as certain pressure-sensitive adhesive tapes and mold-release agents, need to be carefully marked and kept away from bonding operations. The sensitivity of a system to contamination depends on many factors. Rubbers are more tolerant of contamination than epoxies. Rougher surfaces can be more tolerant than smoother surfaces. In general, waterborne or high-solid adhesives and primers are more sensitive to organic surface contamination than solvent-borne materials. Thus, contamination may become a greater issue as environmentally acceptable adhesives and coatings are introduced. This is illustrated in Fig. 1, which shows the solvent-borne system to be very tolerant of grease, while the aqueous system is more sensitive. The type of contamination is also important as illustrated in Fig. 2, which shows that some contamination will cause the rubber–steel peel strength to decrease by 50% at levels less than 1 μg/cm2 while others require more than 100 μg/cm2 or even enhance the bond strength. Because one must always assume the contamination to be detrimental, the first step in an adherend preparation process is usually a degreasing step. Traditionally this was done with chlorinated solvents, but regulations have required substitution of more environmentally acceptable cleaners. In addition to organic contamination, most as-received metal alloys have a thick oxide/hydroxide film formed during heat treatment, fabrication, and other processing.

166

G. D. Davis 40,000

Peel strength (N/m)

30,000

Solvent-based adhesive/primer

20,000

10,000 Aqueous-based adhesive/primer 0 1

1,000

10 100 Contamination (µg/cm2)

Fig. 1 Peel strength of nitrile butadiene rubber (NBR) as a function of grease contamination level on grit-blasted steel using a solvent-borne primer/adhesive and an aqueous primer/adhesive Contaminated rubber–rubber peel strengths 10,000

Peel strength (N/m)

Machine oil #1 8,000 Machine oil #2 Lubrication grease

6,000

Curing agent 4,000 Plasticizer

2,000

Corrosion preventative grease

0 0

10 100 Contamination level (μg/cm2)

1,000

Fig. 2 The effect of surface contamination on the peel strength of rubber–steel bonds (Adapted from Davis 1993)

The coating’s composition, morphology, thickness, and mechanical properties will vary from fabricator to fabricator and from lot to lot. In the case of aluminum alloys containing magnesium, such as the 5000, 6000, and 7000 series alloys, the mill scale can be all or mostly MgO because of the high mobility of Mg in the alloy and its preferential oxidation when exposed to the atmosphere at high temperatures. In copper-containing alloys, such as the 2000 and 7000 series, there often is a buildup

8

Surface Treatments of Selected Materials

167

of copper at the oxide/metal interface as aluminum is preferentially oxidized. This copper can serve to promote corrosion via galvanic coupling with the aluminum and reduce bond durability. The effect of magnesium is less clear and differences in surface morphology and other variables may have played a role in conflicting results. To assure reproducible bonding surfaces, the second step in treatment process is generally a deoxidizing step so that oxidizing or coating process begins with bare metal (or a thin native oxide). The final step in providing an acceptable bonding surface is to grow or deposit an oxide or other surface film designed to promote good, stable bonds with the primer or adhesive. As mentioned earlier, the bonding can be either chemical or physical. Durable covalent chemical bonding across the interface can be achieved with phenolic adhesives, but not the more common epoxy adhesive without the use of coupling agents or sol–gel films. In the other cases, suitable surface morphology to promote physical bonding (mechanical interlocking) is critical. This is discussed below.

8.1.2

Degradation Processes

For most structural joints using epoxy adhesives and metallic adherends, secondary or dispersive bonds are the only ones formed across the interface. Because metal oxide surfaces are polar, they attract water molecules that can disrupt these dispersive bonds across the interface. This disruption can be seen thermodynamically by the work of adhesion in an inert medium, WA, which can be represented as (Kinloch 1987) W A ¼ γa þ γs  γas

(1)

where γa and γs, are the interfacial free energies of the adhesive and substrate, respectively, and γas is the surface free energy. In the presence of a liquid such as water, the work of adhesion WA1 becomes W A1 ¼ γa1 þ γs1  γas

(2)

where γa1 and γs1 are now the interfacial free energies of the adhesive/liquid and substrate/liquid interfaces, respectively. In an inert environment, the work of adhesion for a bonded system will be positive, indicating a stable interface, whereas in the presence of water, the work of adhesion may become negative, indicating an unstable interface that may dissociate. Table 1 shows, in fact, that moisture will displace epoxy adhesives from iron (steel) and aluminum substrates and promote disbonding (Kinloch 1987). In contrast, although moisture weakens epoxy/carbon fibers bonds, these remain thermodynamically stable. Experience with both metal and composite joints confirms these predictions. The data presented in Table 1 illustrate the potential disastrous results when relying solely on dispersive bonds across the interface between an epoxy adhesive

168

G. D. Davis

Table 1 Work of adhesion for various interfaces (Kinloch 1987) Interface Epoxy/ferric oxide Epoxy/alumina Epoxy/carbon fiber

Work of adhesion (mJ/m2) In inert medium In water 291 255 232 137 88–90 22–44

Interfacial debonding after immersion Yes Yes No

and metals or ceramics. Even though good initial bond strengths may be obtained with cohesive failure (within the adhesive), bond durability in the presence of moisture is likely to be poor with interfacial failure. To illustrate this danger, demonstration specimens can be produced that exhibit good initial strength, but fall apart under their own weight when a drop of water is placed at the crack tip. Because moisture will break the dispersive interfacial bonds, mechanical interlocking on a micro- or nanoscopic scale is needed between the adhesive/primer and adherend for good durability. In these cases, even if moisture disrupts interfacial chemical bonds, a crack cannot follow the convoluted interface between the polymer and oxide and the joint remains intact unless this interface or the polymer itself is destroyed. In addition, the micro- or nanocomposite interphase formed with the polymer and oxide allows a gradient of mechanical properties between the adhesive and the metal so that there is less buildup of interfacial stresses. The scale of the microscopic surface roughness is important to assure good mechanical interlocking and good durability. Although all roughness serves to increase the effective surface area of the adherend and, therefore, to increase the number of primary and secondary bonds with the adhesive/primer, surfaces with features on the order of tens of nanometers exhibit superior performance to those with features on the order of microns. Several factors contribute to this difference in performance. The larger-scale features are fewer in number for a given area and generally are smoother (even on a relative scale) so that interlocking is less effective. Depending on the particular treatment used, there may also be loosely bound material that prevents bonding to the integral adherend surface. In addition, the larger-scale roughness frequently allows trapped air and surface contaminants to remain in the microscale valleys. These unbonded regions limit joint performance by reducing both chemical and physical bonds and serving as stress concentrators. In contrast, smaller-scale micro- or nanoroughness tends to be more convoluted in morphology and generates strong capillary forces as the primer wets the surface, drawing the polymer into all the “nooks and crannies” of the oxide and displacing trapped air and some contaminants to form a micro- or nanocomposite interphase. Indeed, cross-sectional micrographs show complete filling of the nanopores (Venables 1984; Clearfield et al. 1990). This dependence on the degree and scale of roughness is illustrated in Fig. 3, which shows wedge test results of titanium bonds with several surface preparations (Brown 1982). Because the titanium surface is stable under these conditions, differences in the joint performance can be attributed solely to differences in the polymerto-oxide bonds and correlate very well with adherend roughness. The poorest

8

Surface Treatments of Selected Materials

169 1µm

50 45

MPF

Crack length (mm)

40 35

PF

30 25

DA

20

DP LP

15 CAA

10

TU 1µm

5 0 0

200

400 Time (h)

600

0.2 µm

Fig. 3 Wedge test results for Ti adherends with several different surface treatments having differing degrees and scales of roughness. Also shown are micrographs of selected specimens showing improved bond performance with increasing micro- or nanoroughness. Specimens were exposed to 100% relative humidity at 60  C. Treatment key: phosphate fluoride (PF), modified phosphate fluoride (MPF), Dapcotreat (DA), dry hone Pasa Jell (DP), liquid hone Pasa Jell (LP), Turco (TU), and chromic acid anodization (CAA) (Wedge test data are from Brown (1982))

performing group of pretreatments (Class I: phosphate fluoride [PF] and modified phosphate fluoride [MPF]) produced relatively smooth surfaces. The intermediate group (Class II: Dapcotreat [DA], dry hone Pasa Jell [DP], liquid hone Pasa Jell [LP], and Turco [TU]) exhibited macrorough surfaces with little microroughness. They had significant improvements in durability over the smooth adherends, but not as good as the Class III pretreatment (chromic acid anodization [CAA]), which provided a much evolved nanoroughness. Because of the importance of nanoscale roughness, high-quality micrographs with magnifications in the order of 50,000X are critical in evaluating surface morphology. Once high-quality physical bonds are formed across the interface, the interphase will not fail unless the polymer disengages from oxide or the interphase degrades in some manner to break the interlocking or allow crack growth through the oxide or at the oxide/metal interphase. The mechanism of degradation will depend on the adherend material and the exposure conditions. For aluminum adherends, moisture causes hydration of the surface, i.e., the Al2O3 that is formed during the surface treatment is transformed into the oxyhydroxide AlOOH (boehmite) or trihydroxide Al(OH)3 (bayerite) (Davis et al. 1982; Venables 1984). The transformation to the hydroxide leads in an expansion of the interphase (the volume occupied by the hydroxide is larger than that originally occupied by the Al2O3). This expansion and the corresponding change in surface morphology and structure induce high stresses

170

G. D. Davis

at the bondline. These stresses coupled with the poor mechanical strength of the hydroxide, promote crack propagation near the hydroxide/metal interface. The rate of hydration of the aluminum oxide depends on a number of factors, including surface chemistry (treatment), presence of hydration/corrosion inhibitors in the primer or applied to the surface, temperature, and the amount of moisture present at the surface or interface. One surface treatment that provides an oxide coating that is inherently hydration resistant is phosphoric acid anodization (PAA). Its stability is due to a monolayer of phosphate incorporated into the outer Al2O3 surface during anodization; only when this phosphate layer goes into solution does the underlying Al2O3 hydrate to AlOOH. The hydration process is illustrated in the surface behavior diagram of Fig. 4 (Davis et al. 1982; Davis 1986). It shows hydration to occur in three stages: (1) a reversible adsorption of water, (2) slow dissolution of the phosphate layer followed by rapid hydration of the freshly exposed Al2O3 to AlOOH, and (3) further hydration of AlOOH to Al(OH)3. Although the evolution of surface chemistry depicts the hydration of bare surfaces, the same process occurs for buried interfaces within an adhesive bond. This was first demonstrated by using electrochemical impedance spectroscopy (EIS) on an adhesive-covered Forest Products Laboratory (FPL, a sodium dichromate + sulfuric acid etch) aluminum adherend immersed in hot water for several months (Davis et al. 1995b). EIS, which is commonly used to study paint degradation and substrate corrosion, showed absorption of moisture by the epoxy adhesive and subsequent hydration of the underlying aluminum oxide after 100 days (Fig. 5). At the end of the experiment, aluminum hydroxide had erupted through the adhesive. Later, crosssection micrographs of bonded tapered double-cantilever beam specimens cyclically loaded during immersion in water revealed the very early stages of hydration (prior to substantial changes in the oxide morphology and volume) in advance of the crack tip (Kinloch et al. 2000). Similar early stages of hydration have also been seen using high-frequency dielectric spectroscopy (Affrossman et al. 2000). Subsequent investigations proved that identical hydration reactions occur on bare aluminum surfaces and bonded surfaces, but at very different rates of hydration Fig. 4 Surface behavior diagram showing hydration of the phosphoric acid anodized (PAA) aluminum surface. Hydration occurs in three stages: I. Reversible adsorption of moisture; II. hydration of the Al2O3 to AlOOH; and III. further hydration to Al(OH)3 (Data are from Davis et al. (1982) and Davis (1986))

AlPO4

I II III H2O

Al(OH)3

AlOOH

Al2O3

8

Surface Treatments of Selected Materials

171

1.E + 07

Low frequency impedance (Ω)

Moisture absorption by adhesive

1.E + 06

Breach of adhesive by hydration products

1.E + 05

Incubation period before substrate hydration Hydration of Al2O3 substrate

1.E + 04 0

50

100 Immersion time (days)

150

200

Fig. 5 Low-frequency electrochemical impedance of an epoxy-coated Forest Products Laboratory (FPL)-etched aluminum adherend as a function of immersion time in 50  C water (Adapted from Davis et al. (1995b))

(Davis et al. 2000). An Arrhenius plot of incubation times, ti, prior to hydration of bare and buried FPL surfaces, for the equation 1 / k ¼ AeEa =RT , ti

(3)

where R is the universal gas constant and T is the absolute temperature in Kelvin, showed that the hydration process exhibits the same energy of activation, Ea, (82 kJ/mole) regardless of the bare or covered nature of the surface (Fig. 6). On the other hand, the pre-exponential factor, A, and the rate of hydration vary dramatically, depending on the concentration of moisture at the oxide interface/surface available to react. The epoxy-covered surfaces have incubation times (and rate constants, k) three to four orders of magnitude longer than bare, immersed specimens, reflecting the limited amount of moisture absorbed by the epoxy and free to react with the oxide. The slow rate of hydration for buried surfaces is good from a service point of view, but makes the study and evaluation of the durability of surface treatments difficult unless wedge tests (ASTM D3762) or similar tests are used to accelerate the degradation. The stress at the crack tip of a wedge test specimen, together with the presence of moisture at the tip, serves to make this test specimen more severe than soaked lap-shear specimens or similar types for which moisture must diffuse from an edge and is only present at the interface in relatively low concentrations. As a result, the wedge test is a better laboratory evaluation of relative durability. Furthermore, from a production perspective, Boeing has correlated the results of wedge tests from

172

G. D. Davis 1.E + 00

1.E – 01 Incubation time (min–1)

Bare substrate

1.E – 02

1.E – 03

1.E – 04

Epoxy covered substrate

1.E – 05

1.E – 06 2.8E– 03

2.9E – 03

3.0E– 03

3.1E – 03

3.2E –03

3.3E –03

Temperature (K–1)

Fig. 6 Arrhenius plot of incubation times prior to hydration of Forest Products Laboratory (FPL)etched aluminum under various conditions (Adapted from Davis et al. (2000) with additional data)

actual aircraft components with their in-service durability (Marceau and Thrall 1985). Specimens fabricated from aircraft components that had exhibited service disbonds showed significant crack growth during the first hour of exposure whereas those fabricated from good components showed no crack growth during this time period. In contrast, lap-shear specimens and porta shear specimens all demonstrated high bond strengths regardless of the service conditions. Wedge tests can be used as a quick quality control test in the production facility (crack length after 1 h must be below a given value) or the laboratory to study long-term stability of a surface/ interface. In principle, lap shear or similar tests can also be used to obtain the same information. However, the time for moisture to diffuse into these specimens and for the specimens to come into equilibrium with the test environment is so long that most investigators stop the test before useful durability information is obtained. Unlike aluminum, titanium adherends are stable under conditions of moderately elevated temperatures and humidity. In the wedge test results of Fig. 3, the adherend surfaces underwent no change in morphology. Failure of the CAA specimens remained within the adhesive, with the physical bonds provided by the microscopically rough oxide remaining intact. For moderate conditions, the key requirement for a titanium treatment is a convoluted nanorough surface to promote physical bonding. At elevated temperatures where titanium alloys would be the adherend of choice, a different failure mechanism becomes important. Because the solubility of oxygen in titanium increases with temperature, the oxygen in a CAA or other oxide dissolves into the metal, leaving voids or microcracks at the metal–oxide interface and embrittles the surface region of the metal. Consequently, stresses are concentrated at small areas at

8

Surface Treatments of Selected Materials

173

the interface and the joint fails at low stress levels. Such phenomena have been observed for adherends exposed to 600  C for as little as 1 h or 300  C for 710 h prior to bonding (Clearfield et al. 1989) and for bonds using a high-temperature adhesive cured at elevated temperatures. To prevent this failure mode, thick oxides, such as those grown by CAA, should be avoided in high-temperature applications.

8.1.3

Processibility Issues

For a surface treatment process to be practical, not only does it have to produce a clean, stable surface suitable for bonding to the primer/adhesive, but it must be able to do this reliably in a production environment. An important aspect of processibility is the acceptable windows for the different process variables. That is, how much can a parameter vary from its nominal value before the performance of the bonded structure is compromised and degraded? The narrower the acceptable window, the tighter the processing controls must be. Tight controls, in turn, are associated with high cost for implementation of the controls, the inspection of the part or processing variable, and the waste from out-of-specification parts that must be reprocessed or discarded. In contrast, a robust process will have wide processing windows so that the prepared surface will still be acceptable despite the inevitable processing discrepancies that occur. The critical processing parameters (those which must be controlled most closely) will depend on the specific process, the material, and the performance requirements. Typical parameters include solution composition/concentration, pH, temperature, and contamination of cleaning, deoxidizing, etching/anodization/deposition, and rinsing solutions, time of each step and time between each step, anodization voltage and current, ambient temperature and humidity, and airborne contamination. An example of process controls is the acidification of rinse water baths following FPL or other chromic acid etches (CAE). Due to the nature of the processing, some etch solution drag-out is inevitable as parts are transferred from the etch bath to the rinse bath. In such cases, the pH of the rinse bath in the vicinity of the aluminum surfaces can be quite low. If low pH rinse water is allowed to dry on the oxide, the surface may become sufficiently acidic to react with the amine-curing agents that typically are in epoxy primers, leading to an undercured epoxy at the oxide interface (McNamara et al. 1983). Similar surfaces bonding with a partially cured (b-staged) film adhesive do not suffer bonding problems, presumably because the curing agent in these films is not sufficiently mobile to diffuse to the interface and be neutralized. Because the acidic residue cannot usually be detected on the surface during production, a two-step rinse after etching/anodizing is recommended. Because of the critical nature of adhesive bonds in the aerospace industry, among others, it is essential to inspect the process to assure quality. This inspection can involve monitoring the critical process parameters, testing of witness specimens prepared at the same time as the structure, or evaluating the surface to be bonded. The best surface to inspect is the actual surface to be bonded, but techniques available for inspection are limited. Witness specimens allow a broad range of

174

G. D. Davis

techniques, including those requiring small sizes, vacuum, or highly trained personnel. They also allow bonding and destructive testing. The most common inspection of actual bonding surfaces is the water-break test. Water, being a polar molecule, will wet a high-energy surface, such as a clean metal oxide, but will not wet a low-energy surface, such as most organic materials and will “bead up.” As a part is removed from a rinse bath or spray, if the water flows uniformly over the entire surface, the surface is clean. However, if the water beads up or does not wet an area, that area likely has an organic contaminant that must be cleaned. A separate set of processibility issues involves health, safety, and environmental concerns. Ozone-depleting compounds (ODCs), such as trichloroethane and trichloroethylene, have been banned. Hexachromium is a carcinogenic toxin and is heavily regulated. Disposal is becoming increasingly difficult and expensive. Other materials are either strong acids or strong bases and require appropriate safety precautions. As a result, considerable research has aimed at developing more environmentally acceptable and less hazardous processes. These will be discussed below along with the more conventional treatments.

8.2

Specific Metal Treatments

In this section, the most common or promising new treatments for aluminum, titanium, and steel are discussed. For more details, the reader is referred to books such as Wegman (1989), Cagle (1973), and Minford (1993) and review articles such as Crithlow et al. (2006) for comprehensive descriptions of the most relevant processes.

8.2.1

Aluminum

Aluminum surfaces are most commonly prepared for adhesive bonding in aerospace applications either by etching or anodization in acid solutions. (For less stringent strength and durability requirements, mechanical abrasion can be adequate.) Common preparations result in nanorough adherend morphologies, which studies have shown yield the best overall bond durability. The two widely used treatments are phosphoric acid anodization (PAA) and chromic acid anodization (CAA). Other treatments include Forest Products Laboratory (FPL) etch, or similar chromic acid etches (CAE), P2 etch, and the boric acid/sulfuric acid anodization (BSAA). These treatments are summarized in Table 2. In addition, sol–gel or grit blasting/silane coupling agent treatments are showing promise as environmentally acceptable processes that can be used for repair.

Anodization Phosphoric acid anodization was developed by Boeing to improve the performance of bonded primary structures. Bonds formed with PAA-treated adherends exhibit

8

Surface Treatments of Selected Materials

175

Table 2 Selected surface treatments for aluminum adherends Process PAA CAA BSAA FPL P2

Solution H3PO4 H2CrO4 H2SO4 + B2O3 + H2O Na2CrO7  2H2O + H2SO4 + H2O FeSO4 + H2SO4 + H2O

Process parameters 20–25  C, 25 min, 10 V 32–38  C, 8 V/min to 40 V, hold 55 min 25–30  C, ramp to 15 V, hold 20–25 min 68  C, 15–30 min 60–70  C, 8–15 min

superior durability during exposure to humid environments compared to those formed with FPL-treated adherends, especially when epoxy adhesives are used. In addition, PAA bonds are less sensitive than FPL bonds to processing variables, such as rinse-water chemistry and time before rinsing. As a result, the PAA procedure has become the treatment of choice in the USA for critical applications. Chromic acid anodization is widely used to improve the corrosion resistance of aluminum surfaces. The use of the good protective coating on the aluminum protects the metal interface and increases the bond durability of the joint. CAA is widely used for aerospace applications in Europe, although it is not as popular as PAA in the USA. Boric acid/sulfuric acid anodization also avoids the chromates of the older processes. It provides an oxide similar to the CAA oxide – thick relative to PAA (0.5 μm) with thick cell walls and narrow pores. Wedge test results indicate durability similar to CAA and PAA. Although there are some differences in detail among the anodization processes used commonly for Al adherend preparation, they share many common features: cleaning, deoxidizing, and anodization. The anodization step for PAA is performed at a constant voltage, whereas for CAA, voltage is applied to the adherend in a gradual or stepwise fashion until the maximum voltage is reached, and is then held for roughly 30 min. Both anodization processes are more costly and time-consuming than the FPL process but are justified by the increase in bond durability as shown by the wedge test results of Fig. 7. The PAA oxide morphology has an open nanorough morphology consisting of narrow-walled pores with whiskers sticking up above the pores, as the isometric drawing (adapted from Venables et al. (1979)) and the high-resolution stereo scanning electron micrograph (SEM) in Fig. 8 show. The total oxide thickness is 400 nm. This type of nanoroughness can provide more mechanical interlocking than most other treatments for improved bond strength and durability, but this increase occurs in the interlocking only if the polymeric polymer or adhesive completely wets and penetrates the pores of the oxide. The nanoscale pores create capillary forces that help this penetration. Cross-sectional SEM views of PAA coated with typical epoxy primers show such penetration (Venables 1984). In such cases, strictly interfacial bond failure is very unlikely. Chemically, the PAA film is amorphous Al2O3 (Venables 1984) with the equivalent of a monolayer of phosphate incorporated onto the surface (Davis et al. 1982). Alloying constituents of the adherend are not generally found in the as-anodized

176 90 FPL

80 Crack length (mm)

Fig. 7 Wedge test results for Forest Products Laboratory (FPL)-etched and phosphoric acid anodized (PAA) adherends (above) and PAA, chromic acid anodized (CAA), and boric-sulfuric acid anodized (BSAA) adherends (below). The upper wedge test involved FM-123 adhesive; the lower wedge test involved FM-94 adhesive so the crack lengths cannot be compared from one test to the other. Exposures were 50  C at 95% relative humidity (RH)

G. D. Davis

70 60 50 40

PAA

30 20 10 0 0

20

40

60

80 100 Time (h)

120

140

160

55 Crack length (mm)

PAA

50

CAA BSAA

45 40 35 0

200

400 Time (h)

600

800

oxide. Depending on the storage conditions, some water can adsorb on the surface, but is readily removed by heating or storing in a dehydrating environment. The morphology of the CAA oxide duplicates the morphology of the surface prior to anodization as oxidation occurs at the interface without dissolution of the existing oxide. The total oxide thickness is 1–2 μm, which is much greater than that provided by PAA. The barrier layer at the bottom of the columns is also relatively thick (40 nm), due to the high anodization voltages used. The CAA oxide is Al2O3; the upper portion is amorphous but the lower part may be crystalline. In contrast to the phosphate incorporated in the PAA oxide, little Cr is incorporated into the CAA oxide. Because the morphology of the outer surface of the CAA oxide is strongly dependent on the process steps used just prior to the CAA treatment, the type of pretreatment can be chosen to enhance the bonding. Samples that are FPL-etched and then anodized have a typical FPL shallow cell-and-whisker morphology. Surfaces that are PAA treated before the CAA anodization have a deeper cell structure, typical of PAA. Finally, surfaces that have a smooth tartaric acid-anodized (TAA) oxide before CAA treatment retain a smooth surface which is typical of the TAA treatment. The morphology of CAA oxides can also be altered by varying the processing conditions, including using higher temperature anodizing solutions and postanodizing phosphoric acid etches (Arrowsmith and Clifford 1985). One advantage

8

Surface Treatments of Selected Materials

Fig. 8 Phosphoric acid anodized (PAA) aluminum morphology. Above: Isometric drawing adapted from Venables et al. (1979). Below: High-resolution stereo scanning electron micrograph (SEM). Note: viewing this stereo micrograph and others in this chapter with a stereo viewer provides a 3-D image of the surface

177 ~10 nm

~400 nm

“Whiskers” ~100 nm

Porous oxide

~400 nm

Barrier oxide

PAA oxide

0.4 μm

to the CAA process is that the oxide is less friable, i.e., less susceptible to mechanical damage, than the PAA oxide. Bonds made with PAA adherends exhibit excellent inherent durability with failure normally occurring within the adhesive (Venables 1984; Chaudhury and Pocius 2002) due to both the hydration resistance of PAA surfaces and the evolved nanoroughness and subsequent interlocking that occurs between the oxide and the adhesive (Venables 1984). Hydration of PAA surfaces occurs via a three-step process, as discussed above. It is the second step (hydration to boehmite) that causes crack propagation. The initial stability of the PAA surface results from the very thin layer of phosphate ions that is incorporated into the oxide during anodization. FPL oxides with an adsorbed layer of phosphonate-containing hydration inhibitor derive their hydration resistance from a similar mechanism (Davis et al. 1985). In all cases, the bond-degradation processes can be retarded by using moisture-resistant adhesives or primers. CAA oxides protect the metal surface from hydration because of their inherent thickness; the important factor for bond durability is the stability of the outer oxide

178

G. D. Davis

structure when water diffuses through the bondline to the polymer/oxide interface. Because hydration rates are a function, in part, of the thickness of the barrier layer (which is directly proportional to the anodizing voltage), typical CAA processes yield oxides that are more resistant to hydration than FPL surfaces. Direct comparisons of the normal surface treatments based on durability tests show that CAA adherends and PAA adherends perform equally well (Fig. 7).

Etches The FPL and other chromic-sulfuric acid etching procedures are the oldest surface pretreatments for aluminum adherends other than simple degreasing or mechanical abrasion. In addition to being used as a complete adherend pretreatment, FPL was used as the first step in other pretreatments, such as PAA and CAA although it has largely been replaced as part of the PAA process for environmental reasons. The presence of hexavalent chromium in the FPL/CAE and CAA solutions led to alternative treatments, such as P2 and BSAA and, later, sol–gel processing. The P2 etch, avoids the use of toxic chromates, but still provides the complex oxide surface morphology that is crucial to a “mechanically interlocked” interface and strong bonding. Ferric sulfate is used as an oxidizer in place of sodium dichromate (Rodgers 1981). The P2 solution produces an oxide morphology similar to what is seen on chromic-sulfuric acid etch surfaces over a broad range of time–temperature–solution concentration conditions. Mechanical testing indicates that P2-prepared surfaces are equivalent to FPL-prepared specimens. The FPL oxide morphology is shown in the isometric drawing (adapted from Venables et al. (1979)) and high-resolution SEM stereo micrograph in Fig. 9. The oxide consists of a network of shallow pores and protrusions or whiskers on top of a thin barrier layer. This nanoroughness provides some mechanical interlocking between the adhesive and the oxide surface that is critical for the durability of epoxy-bonded structures, but is not sufficient to avoid interfacial failure if the oxide is stable enough to hydration (Davis et al. 1982). Chemically, the FPL film is amorphous with varying quantities of adsorbed water that can be removed by heating, vacuum exposure, or adsorption with certain organic hydration inhibitors. Other Treatments Sol–Gel Processing Sol–gel or “solution–gelation” is based on hydrolysis and condensation reactions to form inorganic polymer networks. The coating is schematically shown in Fig. 10. The Boegel process involves a dilute aqueous alkoxide solution containing zirconium isopropoxide (tetra-n-propoxyzirconium, TPOZ) and a silane coupling agent (Blohowiak et al. 1996; Park et al. 2000). The silane constituent can be chosen to be compatible with the primer and to form strong, durable chemical bonds. A glycidoxyl group, such as that of glycidoxytrimethoxysilane (GTMS), is typically used for epoxies. The hydrolyzed aluminum oxide surface promotes the condensation reactions. The part to be treated can be sprayed, drenched, or immersed. No rinsing is required.

8

Surface Treatments of Selected Materials

Fig. 9 Forest Products Laboratory (FPL)-etched aluminum morphology. Top: Isometric drawing adapted from Venables et al. (1979). Bottom: High-resolution stereo scanning electron micrograph (SEM)

179 ~5 nm

“Whiskers” ~40 nm

Shallow porous oxide ~40 nm

FPL oxide ~5 nm

Barrier oxide

0.4 μm

A thin film (typically 50–200 nm) serves as a hybrid inorganic/organic interphase between the metal (oxide) surface and the organic primer or adhesive. The surface is silane-rich and the interface near the metal substrate is zirconia-rich with Zr-O-metal substrate chemical bonds formed with the substrate. The graded coating allows covalent bonding throughout the deposited interphase film. In this way, the sol–gel treatment differs from the traditional oxide treatments discussed above in that chemical bonding and not physical bonding governs joint strength and durability. Sol–gel films deposited on a grit-blasted aluminum surface give performance close to PAA bonds with little crack growth (Fig. 11). Failure is generally cohesive. Silane/Grit-Blast Treatment The grit-blast/silane (GBS) treatment was developed in Australia for use with bonded repair patches. One example is the repair of outer wing fatigue cracks, in which over 1,000 bonded doublers have been applied to primary structure with no crack growth or patch failures for 20 years. The US Air Force has extended the technology and optimized a grit-blast/silane surface treatment for patching other aircraft. In contrast with the oxide treatments discussed above where mechanical interlocking is the key to obtaining good bond strength and durability, the GBS treatment utilizes the functionality of the silane to form chemical bonds with both

180

G. D. Davis

Adhesive NH2

NH2 o

NH2 o

O

O

O

HO

Si

O

Zr

O

Si

O OH

O

Si

O O

O

Zr

Zr

Zr O

O

O

O

Si

Zr O

O

O

O

O

Zr O

Zr Zr

Si

Zr O

O

O

O Zr

O

Zr O

OH

Si

O

Si

O

O

O Si

O

O

Si

O O

o

O O

O

O

O

NH2 o

Si Si

O

O Zr

O

O Si

Zr

O

NH2

O

O Si

O

Graded sol–gel interphase

o

O

O

Substrate

Fig. 10 Schematic of sol–gel surface (Adapted from Blohowiak et al. (1996)) 3.4 Grit blast/sol–gel

Crack length (cm)

3.3

3.2

PAA

3.1

3

2.9 1

10

100

1,000

Time (h)

Fig. 11 Wedge test results for Phosphoric acid anodized (PAA) and grit-blasted sol–gel treated aluminum adherends. Note the expanded crack-length scale (Data are from Blohowiak et al. (1998))

the surface and the adhesive/primer to couple the two materials (Plueddemann 1982; Bascom 1990). Silane coupling agents have the form R-SiX3, where R is an organic function group and X is a hydrolysable group. The GBS treatment uses γ-glycidoxypropyltrimethoxy silane (γ-GPS), in which R is an epoxy group connected to the Si by a propylene chain and X is a methoxy. In an aqueous solution, the methoxies hydrolyze to form the trisilanol (R-Si-(OH3)). For this application, the

8

Surface Treatments of Selected Materials

181

epoxy group reacts with epoxy adhesive while the silanols react with the metal oxide surface. These bonds are stable against moisture attack and can provide excellent bond durability. One issue with silane-coupling treatments is that the performance can be highly dependent on processing conditions, especially hydrolysis and drying conditions. Nonetheless, process controls suitable for field application of bonded repair patches are possible. With such controls, Optimized grit-blast/silane treatments can provide wedge test durability as good as PAA with failure entirely cohesive within the adhesive.

8.2.2

Titanium

Titanium alloys are particularly attractive for aerospace structures due to their high strength-to-weight ratio. In addition, they retain their mechanical properties at high temperatures, so that they can be used as components in structures where operating temperatures up to 370  C (700  F) are expected for short times or up to 177  C (320  F) for extended times. The desire to use adhesively bonded titanium structures at elevated temperatures has been a driving force in the development of hightemperature adhesives (those curing at temperatures up to 400  C). In this section, surface preparations for titanium and the durability of the subsequent adhesive bonds are described. Selected treatments are shown in Table 3. Like Al, durable surface preparations for titanium can be achieved by forming oxides in anodizing and/or etching solutions. Typically, anodization results in the best bond durability for Ti alloys, primarily due to the nanorough surface morphology that results from the treatment. The most commonly used anodization process is CAA (but in a different form than for aluminum adherends) (Venables 1984; Clearfield et al. 1990). The Pasa Jell 107 and Turco 5578 etches are also commonly used. In early studies, the durability of Ti-6Al-4V adherends was determined as a function of surface preparation for ten adherend preparations (Brown 1982) and was discussed above (Fig. 3). The general result was that surface preparations that produce oxides with no roughness (macro- or micro-) yield the poorest bond durability; those that produce a large degree of macroroughness, with little or no micro- or nanoroughness, fall into an intermediate class of moderate-to-good durability; and those preparations that produce oxides with significant nanoroughness

Table 3 Selected surface treatments for titanium adherends Process CAA AP Turco 5578 Pasa-Jell 107

Solution CrO3 + NH4HF2 + H2O NaOH + H2O2 Proprietary, includes NaOH Proprietary

Process parameters 20–25  C, 10 V for 20 min 65  C for 20 min 80–95  C for 15 min 2rp should ideally be selected. The dependence of GIIc (and also GIIIc) upon bondline thickness was observed by Chai (1986, 1988) who noted that values of GIIc and GIIIc were also dependent upon bondline thickness for brittle adhesives. Adhesive layer thickness should therefore be considered carefully when designing adhesive-jointed structures and test coupons.

20.4

Peel Testing

20.4.1 Introduction to Peel Testing The adhesive-joint fracture tests discussed above have been concerned with joints formed between substrates that deform in a linear-elastic manner during the loading of the joint such that the concepts of linear-elastic fracture mechanics apply. Such tests are used routinely to measure fracture energy, Gc. However, there is also a

544

B. R. K. Blackman

requirement to be able to measure the performance of adhesive joints formed between substrates that may deform extensively during an adhesion test. Examples include the adhesion of flexible layers such as food packaging laminates, electronic components, and the structural bonding of relatively thin sheet metal in the automotive industry. In his forward to the special issue on peel in the International Journal of Adhesion and Adhesives (Moore 2008), Moore distinguished between two general types of peel test, namely, those that are dependent upon geometry and those that are geometry independent. Geometry-dependent peel tests are numerous and these typically measure the peel force per unit width of joint, commonly called the peel strength. Geometry-independent peel tests, on the other hand, measure a geometry-independent parameter such as fracture energy, Gc with units J/m2. Using fracture mechanics and the concept of energy balance, significant progress has been made in recent years toward the goal of developing and refining geometryindependent peel tests (Kinloch et al. 1994; Kinloch and Williams 2002). At least in principle, a peel test can now be used to determine the adhesive fracture energy of a joint even when extensive plastic deformation of the peel arm(s) has occurred during the test.

20.4.2 Geometry-Dependent Peel Tests As discussed by Moore (Moore 2008), several peel test configurations have been developed over the years and many developments and modifications to these have been variously proposed in the literature. It is not the aim of the present chapter to review these techniques, but rather the aim here is to identify some of the main types of peel test that fall into this category and discuss their advantages and their limitations. In the so-called fixed-arm peel tests, one substrate is fixed to a rigid base (the rigid substrate) and the other (the flexible substrate) is peeled by a test machine at a defined applied peel angle, which can be varied between tests. Figure 9a shows a schematic fixed-arm peel test. Standardized peel tests describe both 90
and 180


a

b

Peel force

Peel force

Fig. 9 (a) The fixed-arm peel test arrangement and (b) the T-peel test arrangement

20

Fracture Tests

545

fixed-arm peel tests (BSI 1993a, b), but the peel angle can be varied typically from 45
to 180
. Peel tests are typically performed at a constant rate of peeling and the standards specify such details as the width, thickness, and width–thickness ratio of the substrates. The apparatus used to control the peel angle is also described in the various standards, for example, by attaching the peel arm to a cylinder mounted on bearings that is then rotated to create the peel. A floating roller and a climbing drum apparatus are specified in two tests that have been standardized by ASTM (2004a, 2006). The average peel force per unit width (the peel strength) is reported. For the peeling of two flexible substrates, the T-peel test is often adopted. In the T-peel test, the two flexible substrates are bonded over part of their length and the unbonded ends are peeled at 180
and the peel strength is determined as above. A T-peel test has been standardized (BSI 2005). It is common for the substrates in a T-peel test to take up an asymmetric profile when the failure path runs along one adhesive– substrate interface. The T-peel can then become a “Y-peel” as the local peel angle differs from the 180
peel angle applied. In such cases, the effective peel angle between the bonded portion of the joint and the load-line should be measured. For the peeling of two rigid substrates a fixture such as the floating roller peel test can be used. Such a fixture incorporates two rollers, and the peel arm is forced through the rollers at constant speed such that it slides on the roller and curves through a defined angle during peel. Again, the average peel force per unit width is measured and reported as the peel strength. Clearly, while comparative measurements can be made using the above types of peel test, the results obtained will depend upon such factors as the mechanical properties of the peel arms, their thickness, and also upon such factors as the applied peel angle, and the effective “local” peel angle at the peel front. Also, in a study to evaluate these different types of peel test, Kawashita et al. (2005) demonstrated that in tests such as the floating roller peel test, frequently the peel arm does not conform to the roller geometry. This was identified using a modified and partially cutaway apparatus, which allowed the conformation to be evaluated. Such peel tests can be valuable to industry for the purposes of selecting a suitable adhesive for a given substrate material, for developing a surface pretreatment, or for investigating the effects of altering one test parameter (e.g., temperature, humidity or rate of peel). However, the results obtained in one peel test are rarely transferable to other geometries and have limited predictive capability. For this reason, much recent attention has been directed toward the development of geometry-independent peel tests, as is discussed in the next section.

20.4.3 Geometry-Independent Peel Tests The peel strength determined in the above “geometry-dependent” peel tests is a measure of the total work done on the joint during the peeling process. Due to the plastic deformation of the peel arm(s), this total energy measure can be significantly greater than the energy required to drive a crack through the joint under linear-elastic conditions. Bearing this in mind, it is easy to see how peel test results can be

546

B. R. K. Blackman

manipulated to achieve impressive numerical results by simply changing the substrate, for example, by changing from a high yield strength aluminum alloy to pure aluminum for the peel arm, significant increases in measured peel strength can be obtained for an otherwise identical joint. Recent research activity into peel testing (Kinloch and Williams 2002) has thus focused on the development of procedures (both test techniques and analysis schemes) to isolate the energy required to fracture the peel sample, termed the adhesive fracture energy, GA, from the total input energy to the system, termed G. The underlying energy balance is: GA ¼ G  Gp ,

(28)

where Gp represents the plastic energy dissipated in the peel arm. In, for example, the 90
fixed-arm peel test for flexible to rigid peel arms, energy is dissipated by local plastic bending of the peel arm at the peel front. Then, as the section of the peel arm forced through a tight curvature at peel front subsequently straightens out, there is further, reversed plastic bending in the arm. In addition, there may be sufficient tension in the peel arm to exceed the yield strength of the substrate and hence there is additional tensile plastic energy dissipation in the joint. Thus, in the above energy balance, G is a measure of how difficult it is to peel one substrate from another in the joint whereas GA is a measure of how well the substrates are stuck. As described in Kawashita et al. (2006), the ESIS TC4 committee has developed a test protocol covering both the fixed-arm and the T-peel tests, together with an analysis to obtain the geometry-independent adhesive fracture energy, GA, termed the adhesive fracture toughness in the ESIS TC4 protocol, but equivalent to the adhesive fracture energy, Gc. Because the tests require that the plastic energy dissipated in the peel arms be determined, it is also necessary to perform an additional tensile test on the peel arms. The tests developed accommodate situations where the adhesive layer thickness is negligible (e.g., when two polymer films are molded together as is common in packaging laminates) and also when the adhesive layer thickness is not negligible, that is, ha > 0; thus the deformation of the layers needs to be taken into consideration in the analysis. For this latter case, the tensile modulus of the adhesive material is required. This may be available from adhesive manufacturers’ data sheets, otherwise a tensile test on the bulk adhesive will be needed. In the fixed-arm peel test as shown schematically in Fig. 9, the substrates (which can be the same or different materials) are bonded along their length but excluding one section, typically 30 mm in length, which is left unbonded to allow for gripping in a test machine. The size of the test specimen and indeed the test jig to be used are not defined precisely in the ESIS protocol. The protocol permits applied peel angles in the range of 0–180
to be selected although in practice the fixed-arm peel tests are most commonly performed with an applied peel angle of 90
. The protocol additionally requires peel test to be performed at applied angles of 45
and 135
. During the development of the procedures and analysis, there was great interest in applying different peel angles to the joints, as although the total energy applied to the specimen, G, will be a strong function of applied peel angle, the adhesive fracture

20

Fracture Tests

547

toughness, GA, should be independent of peel angle, as has indeed been shown to be the case in the round-robin tests coordinated by ESIS-TC4. Perhaps the most difficult experimental aspect of the fixed-arm peel test is ensuring that the fixed arm is (and remains) rigidly attached to the base. Also, the peel force is usually applied via the crosshead of a testing machine and the base should be fixed to a linear bearing trolley such that the specimen can slide with minimum friction during the peeling of the arm. If the fixed arm is attached to the base using an adhesive, it is important to ensure that this does not fail during the test. The fixed arms are commonly bolted to the free-sliding base. When bolts are used, it is necessary to accommodate the bolt holes in the fixed arm only at the peeling end of the joint, and through both the peel arm and fixed arm at the other end. When this is done, the second bolt is outside of the length of joint to be peeled in the test. It is also possible to use a wider fixed arm than peel arm, such that the extra width of the fixed-peel arm can be used to bolt down this arm to the sliding base. The fixed-arm peel test is conducted at a rate of 10 mm/min when the applied angle is 90
and when other angles are used, then at speeds to give the same equivalent crack speed. The peel force is measured as a function of peel-arm displacement and an average force is determined over a minimum of 30 mm of peel fracture. If peel proceeds via a stick-slip mechanism, then average peel forces at initiation (cohesive fracture) and arrest (interfacial fracture) are determined separately. Figure 10 shows the steady-state peel force obtained in a fixed-arm peel test.

140

120

100 Peel force (N)

Steady-state peel force 80

60

40

20

0

0

20

40 60 Displacement (mm)

80

100

Fig. 10 Peel force versus crosshead displacement obtained during fixed-arm peeling at 10 mm/min at an angle of 90
. The adhesive used was Bondmaster ESP110 and the peel arm was 0.1 mm thick aluminum

548

B. R. K. Blackman 100

E2

80

Stress (MPa)

E1 60

40

Yield co-ordinates, (εy, σy)

20

0

0

0.01

0.02

0.03 Strain

0.04

0.05

0.06

Fig. 11 Stress–strain data obtained for an aluminum peel arm, showing the initial modulus E1, the initial plastic modulus E2 used in the bilinear model, and the yield coordinates

In this test, an aluminum peel arm of thickness 0.5 mm has been bonded to a rigid aluminum alloy base using Bondmaster ESP110 epoxy adhesive. This was then peeled at an angle of 90
and at a rate of 10 mm/min. As mentioned above, in addition to the fixed-arm peel test, it is necessary to perform a tensile test on the peel arm. This is performed at the same test speed as used in the peel test. The stress–strain curve is determined. Five parameters are obtained from the tensile test, namely, E1, E2, α(E2/E1), ey, and σ y. These are shown in Fig. 11. The elastic modulus (E1) is defined as the slope of the stress–strain curve at small deformations and before yielding has occurred. In order to define the plastic modulus (E2), it is suggested that the tensile specimen be allowed to continue extending for as long as possible. E2 is drawn to the slope of the initial plastic region. This straight line should be drawn approximately between the limits of ey–10ey. The parameter α is then given by the ratio of the two slopes, E2/E1. If the peel arm work hardens, a work hardening coefficient is determined from the power law: σ ¼ AeN :

(29)

If the “work hardening” portion of the stress–strain curve exhibits a negative slope, then values of α and E2 are set to zero (and not negative) (Moore and Williams 2007). The plastic bending in the fixed-arm peel test can be modeled using largedisplacement beam theory with modifications for plastic bending (Kinloch and Williams 2002). Much of the development of the correction procedures has focused on how the

20

Fracture Tests

549

Fig. 12 T-peel schematic specimen showing asymmetric loading due to differing peel-arm stiffness

Peel force Peel arm 1

q f Peel arm 2

stress–strain behavior of the peel arm is treated. Indeed, solutions have been formulated for three cases: the various approaches investigated and adopted have been (1) the use of a bilinear model to fit the data, (2) the application of a power law, and (3) the digitization of the complete stress–strain curve. The digitization procedure provides the most accurate results but frequently the simple bilinear model of the stress–strain behavior is sufficient. The analysis of the test results is most conveniently performed using software developed at Imperial College called ICPeel. This software is freely available from the Imperial College Web site http://www.imperial.ac.uk/meadhesion. Preparing T-peel test specimens is simpler than preparing fixed-arm peel specimens as there is no need to attach one substrate to a fixed base and no need for a linear bearing trolley to accommodate the specimen sliding. A typical T-peel test specimen is shown schematically in Fig. 12b. Two flexible substrates are bonded with a 30 mm disbond, as above. The substrates may or may not be of the same material. According to the protocol (Moore and Williams 2007), loading is performed at 10 mm/min. Figure 12 shows the peel angles that may be obtained in the T-peel test. If the loading remains symmetric in the test, then both the peel angles will be 90
. However, if the substrates are of different stiffness, or the crack runs to an interface such that the adhesive is all on one side, then the peel angles will be φ and θ (rather than 90
) as shown in Fig. 12. To account for the plastic work done on the substrates, a tensile test on the arms is again necessary. If the substrates are dissimilar, then tensile tests on each peel-arm material is required. The analysis of the resulting tensile test is as described above for the fixed-arm peel test. The analysis of the T-peel test data is similar to that for the fixed-arm peel test, except that the peel arms are treated individually. Only one of the peel angles needs to be considered since φ = π  θ. For the two peel arms individually: G1 ¼

P P ð1 þ cos ϕÞ and G2 ¼ ð1  cos ϕÞ, b b

(30)

550

B. R. K. Blackman

where subscripts 1 and 2 refer to the individual peel arms. Determining the adhesive fracture toughness, GA, for the two peel arms individually then: ðGA Þ1 ¼ G1  ðGP Þ1 and ðGA Þ2 ¼ G2  ðGP Þ2 :

(31)

For the complete T-peel joint combining the individual contributions above leads to: GA ¼ ðGA Þ1 þ ðGA Þ2 :

(32)

The main limitation of the peel tests described above is that the adhesive fracture toughness, GA, is determined from the difference between two numbers, that is, G  Gp. When modern, structural adhesives are employed to bond the substrates, these two numbers may be large, and as a consequence, the repeatability of the GA values determined may be poor. Recent research has investigated the application of various validity criteria for peel tests. One such criterion has been to limit the size of the plastic correction term, Gp, to a fixed percentage of G, for example, a limit of (Gp/G) < 80% has been utilized. Another criterion proposed has been to limit the maximum strain in the peel arm.

20.5

Relationship Between Gc (Bulk, Fracture, and Peel)

As discussed in Sect. 1, fracture tests on bulk adhesive specimens are frequently undertaken as part of the materials development process. Loading is necessarily in mode I in bulk tests as even if other loading modes are applied, the crack path will tend to deviate to run locally in mode I. Values of KIc or GIc are determined, and these relate to the toughness of the bulk adhesive. The tests are very sensitive to the initial notching conditions and the values are those associated with crack initiation (unless steady-state propagation is clearly observed in the CT specimens). Fracture tests on adhesive joints are more complex, as the loading mode, the adhesive, the substrate, and the interface or interphase region all have an effect on the values of Gc determined. In addition, both initiation and steady-state propagation values of Gc are usually determined, allowing the resistance curve, that is, the plot of Gc, versus crack length to be drawn. Comparisons of Gc values between bulk and mode I adhesive-joint fracture tests can be drawn, but considerable care is required when making this comparison. For example, it is often observed that for the DCB or TDCB tests, provided the adhesive layer is sufficiently thick to avoid the effects discussed in Sect. 3.4 and provided the crack grows cohesively through the adhesive layer with no rising R-curve observed, the value of GIc from the joint is equivalent to the value of Gc obtained in the bulk test. Clearly, if the bondline thickness is insufficient to allow for the full development of the plastic zone or if a weak interface causes the crack failure path to deviate from running cohesively in the adhesive, then lower values of Gc may be measured in the joint than in the bulk. If the adhesive joint exhibits rising R-curve behavior, which may occur if additional energy absorbing

20

Fracture Tests

551

mechanisms develop as the cracks propagate, then in such cases only the initiation values of GIc from the joint may be comparable to the bulk values. Also, when comparing bulk and joint adhesive properties, the state of cure of the adhesive should be shown to be equivalent as the exact processing conditions are affected by factors such as the thermal inertia of the joint and potential exothermic reactions of the bulk. Measurement of the glass transition temperature of the cured adhesive in the bulk and joint can be compared using techniques such as differential scanning calorimetry. Finally, peel tests are used routinely by industries to assess adhesion. In geometry-dependent peel tests, G is the total energy dissipation on peeling and is significantly affected by factors such as the material properties of the peel arm, its thickness, and the angle of peel. In broad terms, these tests measure how difficult it is to peel the arms. This value of G may be many times greater than the value of GIc determined using the DCB or TDCB test. The degree of adhesion is just one of several factors that influence this measure. In geometry-independent peel tests, the energy dissipated plastically in the peel arms is determined and taken away from the total energy dissipation G to yield the adhesive fracture toughness, GA. It is the value of GA that may be considered equivalent to GIc, provided that the failure paths in the peel test and in the adhesive joint are both cohesive, in the adhesive layer. It should be noted that in the fixed-arm peel test, failure paths often occur close to the adhesive–substrate interface and thus may, in principle, yield values of GA that are lower than the values of GIc determined using DCB or TDCB specimens in which cohesive cracks run through the center of the adhesive layer. This difference may be insignificant for brittle adhesives where the plastic zone size is small, but may be larger for tough adhesives. T-peel tests are sometimes preferred for this reason, as similar failure paths can be observed to the LEFM tests.

20.6

Conclusions

This chapter has described and discussed the fracture tests commonly performed on both bulk adhesive specimens and on adhesive joints. Whenever possible, fracture mechanics approaches have been followed with the goal of determining geometryindependent toughness values. Fracture tests on bulk adhesives are commonly conducted as part of the materials development process. Key to obtaining valid fracture results in bulk tests is to ensure (1) the generation of sufficiently sharp initial notches and (2) the manufacture of sufficiently good quality test specimens that pass the linearity and size criteria, as imposed by the standard, and thus allow the plane strain fracture toughness of the adhesive to be measured. The fracture toughness values obtained in bulk tests (i.e., KIc and GIc) will not be influenced by “adhesivejoint” factors such as constraint or substrate effects, or the quality of interfacial adhesion achieved. Fracture tests on adhesive joints are always more complex than tests on bulk adhesive specimens and a G rather than a K approach is invariably followed for their analysis. The adhesive is present as a thin layer, it may be constrained by the

552

B. R. K. Blackman

presence of nearby substrates and the failure paths may be influenced by the poor adhesion with the substrate. Fracture tests on adhesive joints are commonly conducted in mode I (tensile opening mode), mode II (in-plane shear mode), and mixed-mode I/II (combinations of mode I and II). The key to success in all LEFM fracture mechanic tests is to ensure that the substrates do not deform plastically during loading. Plastic deformation of the substrates would violate the assumptions of LEFM and invalidate the results. In mode I, the DCB and TDCB test specimens are almost universally employed and these tests have been standardized. Either test may be used for the determination of GIc. The choice of which to use is likely to depend upon factors such as the toughness of the adhesive, the properties of the substrate material employed, and whether or not crack length measurement is possible. In mode II, the ELS, ENF, the 4-point loaded ENF, and the tapered ENF (TENF) test specimens have all been employed to measure GIIc values in adhesive joints with various levels of success. There is currently no agreed test standard for the determination of GIIc in adhesive joints, but efforts are underway in the ESISTC4 (a European Fracture Committee) to develop a mode II standard, based on the ELS test method. The situation for mixed-mode fracture is similar, with several test specimens having been proposed but none yet standardized for adhesive joints. Values of Gc obtained from fracture tests on adhesive joints can be sensitive to variations in bondline thickness, and to assess this possibility reference should be made to the size of the plastic zone in relation to the bondline thickness. Peel testing of flexible substrates can be considered as a class of fracture test, albeit one accompanied by extensive plastic deformation of the peel arm(s). A multitude of peel tests have been proposed and typically these determine a value of the peel strength, that is, a measure of how difficult it is to peel the arms during the test. Peel tests are thus highly geometry dependent. Recent research has focused on test and analysis methods to obtain geometry-independent values, for example, adhesive fracture toughness – one measure of how well the substrates are bonded. The method requires additional information about the stress–strain characteristics of the peel arms and also of the adhesive layer, so that the energy dissipated in these can be determined and accounted for in the fracture toughness values determined. Finally, a major goal of fracture mechanics has been to derive geometryindependent values of the fracture parameter Gc. Great progress has been made in this endeavor with the determination of GIc from standardized bulk and adhesivejoint tests and GA from geometry-independent peel tests. However, great care is required with the interpretation of these fracture parameters.

References ASTM (1990) ASTM D3433. Annual book of ASTM standards. Adhesives section 15, Philadelphia. 15.06 ASTM (1999) ASTM D5045. Determination of the fracture toughness of polymers ASTM (2004a) Standard test method for floating roller peel resistance of adhesives. West Conshohocken, PA: ASTM International

20

Fracture Tests

553

ASTM (2004b) Standard test method for mixed mode I-mode II interlaminar fracture toughness of unidirectional fiber reinforced polymer matrix composites. ASTM Stand D6671M:1–14 ASTM (2006) Standard test method for climbing drum peel for adhesives. West Conshohocken, PA: ASTM International Baldan A (2004) Review: adhesively-bonded joints and repairs in metallic alloys, polymers and composite materials – adhesives, adhesion theory and surface pretreatment. J Mater Sci 39:1–49 Berry JP (1960) Some kinetic considerations of the Griffith criterion of fracture-I; equations of motion at constant deformation. J Mech Phys Solids 8:207–216 Bishopp JA (2005) Surface pretreatment for structural bonding. In: Cognard P (ed) Handbook of adhesives and sealants, vol 1. Elsevier, pp 163–214 Blackman BRK (2008) Delamination in adhesively bonded joints, Chapter 22. In: Sridharan S (ed) Delamination behaviour of composites. Woodhead Publishing, Cambridge Blackman BRK, Dear JP et al (1991) The calculation of adhesive fracture energies from DCB test specimens. J Mater Sci Lett 10:253–256 Blackman BRK, Hadavinia H et al (2003a) The calculation of adhesive fracture energies in mode I: revisiting the tapered double cantilever beam (TDCB) test. Eng Fract Mech 70(2):233–248 Blackman BRK, Kinloch AJ et al (2003b) Measuring the mode I adhesive fracture energy, GIC, of structural adhesive joints: the results of an International round-robin. Int J Adhes Adhes 23:293–305 Blackman BRK, Kinloch AJ et al (2005) The determination of the mode II adhesive fracture energy, GIIC, of structural adhesive joints: an effective crack length approach. Eng Fract Mech 72:877–897 Brunner AJ, Murphy N et al (2009) Development of a standardised procedure for the characterisation of interlaminar delamination propagation in advanced composites under fatigue mode I loading conditions. Eng Fract Mech 76(18):2678–2689 BSI (1993a) Peel test for a flexible bonded-to-rigid test specimen assembly. 90 degree peel, Adhesives. London: BSI BSI (1993b) Peel test for a flexible bonded-to-rigid test specimen assembly. 180 degree peel, Adhesives. London: BSI BSI (2001) Determination of the mode I adhesive fracture energy, GIC, of structural adhesives using the double cantilever beam (DCB) and tapered double cantilever beam (TDCB) specimens. BS 7991 BSI (2005) T-Peel test for flexible-to-flexible bonded assemblies, Adhesives. London: BSI BSI (2009) Structural adhesives. Guidelines for the surface preparation of metals and plastics prior to adhesive bonding. London: BSI, 35 pp Chai H (1986) Bond thickness effect in adhesive joints and its significance for mode I interlaminar fracture of composites. In: Seventh conference on composite materials: testing and design ASTM STP 893. American Society for Testing and Materials, Philadelphia Chai H (1988) Shear fracture. Int J Fract 37:137–159 Davies P, Blackman BRK et al (2001) Mode II delamination. In: Moore DR, Pavan A, Williams JG (eds) Fracture mechanics testing methods for polymers adhesives and composites, ESIS publication 28. Elsevier, Amsterdam/London/New York, pp 307–334 Dillard DA (2010) Improving adhesive joint design using fracture mechanics. In: Dillard DA (ed) Advances in structural adhesive bonding. Woodhead Publishing, Oxford/Cambridge/New Delhi, pp 350–388 Dillard DA, Singh HK et al (2009) Observations of decreased fracture toughness for mixed mode fracture testing of adhesively bonded joints. J Adhes Sci Technol 23(10–11):1515–1530 Edde FC, Verreman Y (1995) Nominally constant strain energy release rate specimen for the study of mode II fracture and fatigue in adhesively bonded joints. Int J Adhes Adhes 15:29–32 Fernlund G, Spelt JK (1994) Mixed-mode fracture characterization of adhesive joints. Compos Sci Technol 50:441–449 Hashemi S, Kinloch AJ et al (1990) The analysis of interlaminar fracture in uniaxial fibre-polymer composites. Proc R Soc Lond A427:173–199 ISO (2000) Standard test method for the determination of fracture toughness (Gc and Kc) – a linear elastic fracture mechanics approach. ISO Standard ISO 13586

554

B. R. K. Blackman

ISO (2001) Standard test method for mode I interlaminar fracture toughness, GIC, of unidirectional fibre-reinforced polymer matrix composites. ISO 15024 ISO (2009) Adhesives – determination of the mode I adhesive fracture energy GIC of structural adhesive joints using double cantilever beam and tapered double beam specimens. ISO Standard 25217 Kawashita LF, Moore DR et al (2005) Comparison of peel tests for metal-polymer laminates for aerospace applications. J Adhes 81:561–586 Kawashita LF, Moore DR et al (2006) Protocols for the measurement of adhesive fracture toughness by peel tests. J Adhes 82:973–995 Kinloch AJ (1987) Adhesion and adhesives: science and technology. Chapman & Hall, London/New York Kinloch AJ, Williams JG (2002) The mechanics of peel tests. In: Pocius AV, Dillard DA (eds) The mechanics of adhesion, vol 1. Elsevier, Amsterdam, pp 273–302 Kinloch AJ, Kodokian GKA et al (1992) The adhesion of thermoplastic fibre composites. Philos Trans R Soc Lond A 338:83–112 Kinloch AJ, Lau CC et al (1994) The peeling of flexible laminates. Int J Fract 66:45–70 Liechti KM, Freda T (1989) On the use of laminated beams for the deformation of pure and mixed mode fracture properties of structural adhesives. J Adhes 29:145–169 Liu Z, Gibson RF et al (2002) The use of a modified mixed mode bending test for the characterisation of mixed mode fracture behaviour of adhesively bonded metal joints. J Adhes 78:223–241 Martin RH, Davidson BD (1999) Mode II fracture toughness evaluation using four point bend, end notched flexure test. Plast Rubber Compos 28(8):401–406 Molitor P, Barron V et al (2001) Surface treatment of titanium for adhesive bonding to polymer composites: a review. Int J Adhes Adhes 21:129–136 Moore DR (2008) An introduction to the special issue on peel testing. Int J Adhes Adhes 28:153–157 Moore DR, Williams JG (2007) A protocol for determination of the adhesive fracture toughness of flexible laminates by peel testing: fixed arm and T-peel methods. ESIS TC4 protocol O’Brien TK (1998) Composite interlaminar shear fracture toughness, GIIC: shear measurement or sheer myth? In: Bucinell RB (ed) Composite materials: fatigue and fracture, STP 1330, 7th vol. ASTM, West Conshohocken, pp 3–18 Papini M, Furnlund G et al (1994) Effect of crack growth mechanisms on the prediction of fracture load of adhesive joints. Compos Sci Technol 52:561–570 Qiao P, Wang J et al (2003a) Analysis of tapered ENF specimen and characterization of bonded interface fracture under mode II loading. Int J Solids Struct 40:1865–1884 Qiao P, Wang J et al (2003b) Tapered beam on elastic foundation model for compliance rate of change of TDCB specimen. Eng Fract Mech 70:339–353 Reeder JR (1993) A bilinear failure criterion for mixed-mode delamination. In: Camponeschi ET (ed) Composite materials: testing and design, ASTM STP, vol 11. ASTM, Philadelphia, pp 303–322 Ripling EJ, Mostovoy S et al (1964) Measuring fracture toughness of adhesive joints. Mater Res Stand (ASTM Bull) 4(3, March):129–134 Swadener JG, Liechtiet KM et al (1999) Mixed mode fracture of automotive bonded joints. SAMPE-ACCE-DOE advanced composites conference, Detroit Williams JG (1988) On the calculation of energy release rates for cracked laminates. Int J Fract 36:101–119 Williams JG (2001) Determination of KIC and GIC in polymers. In: Moore DR, Pavan A, Williams JG (eds) Fracture mechanics testing methods for polymers, adhesives and composites. Elsevier Science, Amsterdam

Impact Tests

21

Luca Goglio

Contents 21.1 21.2 21.3 21.4 21.5

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Block Impact Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Impact Wedge Peel Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hopkinson Bar Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Other Test Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.5.1 Fracture Energy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.5.2 Drop-Weight-Based Tests and Alternative Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.6 Conductive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.7 Response to Environmental Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

556 558 561 563 574 574 577 585 588 589 590

Abstract

Assessing the mechanical properties of adhesives and the joints strength requires specific tests. Indeed, since polymers are, in general, sensitive to the strain rate they undergo, different phenomena can be induced when a load is applied abruptly. Under such a condition, usually the adhesive tends to react to the deformation with higher stress and lesser ductility, which causes higher resisting loads but lower absorbed energy. This is why the common standard impact tests for adhesives, involving bonded blocks or strips, aim at measuring the energy required to break the bond. However, since this result is not sufficient to characterize an adhesive, other types of tests have been introduced to obtain a deeper insight. Different types of specimens (samples of bulk adhesive, lap joint, butt joint, double cantilever beam) and test rigs (pendulum, falling weight, hydraulic actuator) have been L. Goglio (*) Department of Mechanical and Aerospace Engineering, Politecnico di Torino, Torino, Italy e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_21

555

556

L. Goglio

adopted in many works to measure the properties of interest. The fracture energy, typical of fracture mechanics, requires ad hoc tests and processing to be measured in dynamic conditions. The Split-Hopkinson pressure bar, a special apparatus conceived to test materials at high strain rate, has been also applied with success to adhesives and joints. Also test rigs based on the falling weight schemes have been widely used. A special case is that of conductive adhesives, which are getting into use to replace soldering in electronic packaging: their capability to withstand the impact that a device can undergo in use must be assessed. This chapter, mainly relying on the results available from the technical literature, presents a survey of various test methods, focusing on the related problems and achievements.

21.1

Introduction

In many of the applications in which adhesives are used, it is likely that they are exposed to loads applied abruptly. A first case is that of a bonded assembly included in a mechanism which, in its service, undergoes discontinuous loading. Examples are mechanisms which undergo intermitting contact, or machineries which exert a crushing action on a material (like mills), or devices conceived to absorb impacts regularly. A second case is that of structures which under normal service are not loaded impulsively but in special conditions (accidents of various types) can be subjected to impacts. This is especially true for automotive applications, in which the crashworthiness is of crucial importance, but it also applies to devices which accidentally can undergo impacts (for instance, due to misuse, fall, etc.). According to the case, the expectations are different: if the impacts are a normal service condition, they must be sustained repeatedly without damage; otherwise, if the impact is an exceptional case, the goal is to minimize its consequences, accepting that a certain level of damage (sometimes high) is unavoidable. It can be remarked that in the former case, the determination of the adhesive performance can be expressed in terms of force or stress, while in the latter case, the typical quantity involved is the absorbed energy. Two typical issues arise about the behavior of an adhesive subjected to impact. The first one is whether an abrupt loading can cause brittle behavior in a material that under static or quasi-static conditions would not be brittle; this aspect is critical regarding the capability of absorbing energy. The second issue is assessing the influence of the loading rate on the adhesive response, i.e., the sensitivity of the adhesive properties to the strain rate. It is clear that such issues, all concerning material response, need experimental results to be addressed. For this reason, ad hoc tests have been proposed and developed, also as international standards. The commonest is the block impact test, defined as an American Society for Testing and Materials (ASTM) standard – and subsequently also as European Standard (EN) and International Organization for Standardization (ISO) standard – in which an impact is applied, with a loading scheme similar to that of the Izod pendulum, to a block bonded to a base. Another

21

Impact Tests

557

popular test, especially in the automotive industry, is the impact wedge peel test, originally developed by Ford and presently also defined as an ISO standard; in this case the bond between two sheet metal strips is broken by a wedge that goes through the joint. These two tests are described in the next two sections of this chapter. In addition to these standard tests, a considerable amount of research on the impact of adhesives has been carried out and presented by many scientists over the last four decades. A special apparatus, developed for testing materials under high strain rate and applied also to adhesives, is the Hopkinson bar, described in a subsequent section. A selection of other special testing methods and the related modeling are described in a further section. In many cases, apart from the rupture energy (which can be obtained directly with pendulum-like devices), the measurement carried out in impact testing is the force time history. The typical feature of the latter is its oscillating behavior – posing serious problems when analyzing the curve – which is justified by two arguments. The first explanation is based on the discontinuous nature of crack propagation: when the critical condition is attained, the crack grows, and this relieves the local stresses and makes the system more compliant; thus the load decreases; then, as the impactor advances further, the cycle restarts. The second explanation considers that the specimen receiving the impact starts to vibrate, independently of the damage that can occur. Moreover, the vibrational properties are not necessarily a feature of the specimen alone, since they often depend on the system formed by the specimen and the test rig (which comes into play with its deformability). Unfortunately, in many cases it is very difficult to distinguish one contribution from the other, especially when the recorded time history is the only information available. The operation of smoothing the recorded curve (e.g., by filtering the high-frequency content) implies the risk of losing some useful information; on the other hand, the raw data can be misleading, in the sense that purely dynamic effect can be taken as material properties (e.g., force peaks at breakage). Thus, the evaluation of impact measurements often requires careful judgment, to avoid erroneous interpretation and false results. It is clearly impossible to recall in a short space all the experiences about impacts tests on adhesives reported in the literature and describe in detail the findings. The objective of the present review is to give general information about what is (and can be) done in this field regarding the experimental arrangements adopted, the types of results obtained, and the behaviors observed. Another broad and detailed state-ofthe-art review on adhesives and joints under impact has been very recently presented by Machado et al. (2017). From these reviews, an interested reader could get an address to select the types of experiments and results that fit his/her need. Of course, examining the original literature is always recommended. Among the several contributions appeared in recent times, the special issue of the Int. J. of Adhesion and Adhesives on “Impact phenomena of adhesively bonded joints,” Sato and Goglio (2015), can be mentioned. It can be also remarked that often the cited articles are not exclusively dedicated to experimental works, since they report also the numerical and/or analytical modeling of the specimen behavior. This is due both to the use of the experimental results

558

L. Goglio

to validate the models and to the additional insight given by modeling to understand the stress and strain state during the tests.

21.2

Block Impact Tests

The basic principle on which this test, given by the ASTM D950 standard and by ISO 9653:1998 and EN ISO 9653:2000 as well, relies is to apply an impact loading condition, mainly in shear, by means of a test rig similar to that used for Izod resilience measurement. Figure 1 shows the experimental arrangement. An upper block (usually of steel) is bonded onto another larger block which, in turn, is fixed to the base of the test rig. The first block is struck by a hammer, in a direction parallel (at least at the beginning of the impact) to the bond surface. The energy required to fracture the sample is obtained from that lost by the pendulum, considering a correction for the kinetic energy of the projected parts. The reason for using this type of specimen, quite different from a real joint, is double. First, the block offers a large lateral face, which is easy to hit by means of the hammer. Secondly, the scheme aims at creating a condition of pure and constant shear in the adhesive layer. A thorough analysis of this case has been carried out by Adams and Harris (1996) who, examining the features of this test, have analyzed in detail the contact conditions between hammer and side of the block. In an ideal situation, the impacting surfaces of the hammer and of the block should be perfectly parallel, giving a uniform load on the block. In practice, a certain degree of misalignment is unavoidable. The authors, by means of finite element analyses, have compared (Fig. 2) the ideal case (I) with the limit cases in which – due to misalignment – the contact occurs at the upper edge of the block (case II) or near to the bond (case III). Their main findings are that, even in case of perfect alignment (Fig. 3, case I), the stresses in the

Fig. 1 ASTM block impact test (From Bezemer et al. 1998, copyright Elsevier)

21

Impact Tests

559

Fig. 2 Possible modes of impact of the pendulum against the block: (I) perfectly parallel; (II) far from the bond; (III) close to the bond (From Adams and Harris 1996, copyright Elsevier)

adhesive are far from being constant: the shear stress peaks near to the bond end at the impacted side; also the peel stress is significant and peaks near to the bond end at the impacted side; then it decreases monotonically assuming negative sign at the opposite bond end. The non-constancy of the shear stress is generated by the “differential straining” mechanism due to the deformability of the block (see ▶ Chap. 24, “Analytical Approach”) that emphasizes the stress value at the end of the bond. The peel stress can be primarily regarded as a bending stress caused by the rotation of the impacted block, which – also in this case due to the deformability – is not linearly distributed. In case of misalignment, the distribution of the shear and peel stress is strongly influenced, as shown in Fig. 3 (cases II and III), due to the changes of moment value and the compressive local deformation of the block (this effect can be regarded as a compression that “irradiates” from the contact point). If the contact occurs far from the bond (case II), the bending moment is higher, and the peak of peel stress increases, while the local deformation of the block has little influence and the shear stress is not much affected. Conversely, if the contact occurs close to the bond (case III), the combination of low bending moment and compressive local deformation of the block causes a compressive peel stress in the adhesive close to the impacted bond end. At the same time, the local deformation of the block increases the differential straining at the same bond end, which causes a higher peak of shear stress. Obviously, these features refer to the stress distribution at the beginning of the impact; then the situation evolves depending on the deformations (and damage) that take place. Thus, in global terms, the loading conditions and the strain energy are influenced, depending on the specific case; this aspect is important, since the elastic energy in the steel blocks can be even higher than the rupture energy of the adhesive. From all these reasons, it comes out clearly that the block impact test is not suitable to give absolute figures about the energy absorption of an adhesive, due to the dependence of the results on the test rig and testing conditions. The only usefulness of the test is that of comparing the behavior of different types of adhesives, but also in this perspective the results must be considered carefully, keeping in mind the importance of the possible misalignments.

560

L. Goglio

a

y F

6.0

t xy /F –3 10 mm–2

(25 mm wide)

x

4.0 III I 2.0 II

0

b

5

10

5

10

x mm

15

20

15

20

10.0 8.0 6.0 II

s yy /F –3 10 mm–2

4.0 I

2.0 0

–2.0

III

x mm

–4.0 –6.0

Fig. 3 Stress distribution on the adhesive midplane of the impact block ASTM 250 for different loading cases: (a) shear stress; (b) peel stress (From Adams and Harris 1996, copyright Elsevier)

An alternative specimen for impact testing of adhesives, aiming at a stress distribution mainly in shear and suitable for a speed higher than that reachable with a pendulum, has been proposed by Bezemer et al. (1998). The specimen in this case is formed by a ring bonded to a coaxial rod; the radial gap defines the desired thickness of adhesive. In the test, the rod is hit axially – and pushed away

21

Impact Tests

561

from the ring – by a projectile released from a drop tower or launched by an air gun, depending on the speed which is wanted; obviously, also static loading on a standard test machine is possible, especially for the sake of comparison. The contact force between projectile and rod is measured and recorded during the test; hence the displacement can be calculated by integration of the time history, and, in turn, also the energy can be obtained. The authors recognize that, though the motion of the rod with respect to the ring generates a tangential relative displacement, the strains and the stresses are not of pure shear; thus also a considerable normal stress is applied to the adhesive, in spite of the expectation. On the other hand, compared to the block impact, this solution still offers the advantages of a lesser sensitivity to misalignment and broader range of possible loading speed. However, this type of test is not widespread and has not reached the level of a standard.

21.3

Impact Wedge Peel Tests

Another objection against the block impact test is that the adherends involved (blocks) are thick and this does not reflect real applications, in which bonding is usually adopted to join thin parts. From this viewpoint, it appears more realistic to test a bond which joins two metal strips, highly deformable both elastically and plastically. This is just what is done with the “Impact wedge peel test,” foreseen by the standards ISO 11343:2003 and EN ISO 11343:2005. The specimen (Fig. 4) is made by two metal strips, formed and bonded together to obtain the shape of a Y. The strips are 90 mm long and 20 mm wide; their thickness can range between 0.6 and 1.7 mm. The bonding length is 30 mm, without any pre-cracking or crack initiator; the not bonded zones form the arms of the Y, which are clamped when the specimen is mounted on the test rig. Even if, to obtain the Y shape, the strips can be formed after bonding, it is recommendable to preform them, to avoid the risk of damaging the adhesive in the forming operation. The specimen is loaded by a wedge which is one side of a rectangular shackle and goes through the bond, thus separating the strips in peeling mode. The pull is applied to the shackle by means of a pendulum Wedge

Specimen grip Substrates Adhesive Wedge retaining shackle

Fig. 4 ISO 11343 wedge impact peel test specimen (From Blackman et al. 2000, copyright Kluwer)

562

L. Goglio

or an actuator, with typical test speeds of 2 or 3 m/s depending on the material of the strips (lower value for steel, higher value for aluminum). During the test, the impact force is measured and recorded for subsequent processing. A broad study of this test has been carried out by Blackman et al. (2000), who – on the basis of a series of tests carried out on several types of adhesives, at different speeds and temperatures, and using a high-speed hydraulic machine – have analyzed the effect of the specimen geometry and type of crack growth. The onset and the propagation of the rupture have been observed by means of high-speed photography. Furthermore, the authors have reviewed the methods proposed by the standard to extract the results and have applied a finite element model to reproduce the fracture of the specimen. Two types of crack growth can be observed in this test, stable and unstable. In case of stable growth it has been observed that the crack tip runs ahead the wedge at approximately constant offset; thus the propagation speed is given by the wedge speed. The force time history exhibits initial peaks, due to dynamic effects originated by the sudden contact of the impactor with the specimen, and then a plateau appears and lasts for most of the test. Furthermore, the elevated plastic deformation of the metal strips that takes place involves a considerable amount of strain energy. In case of unstable growth, typical at low temperature (
40  C in these experiments), or with a brittle adhesive, the crack runs rapidly through the whole bond, at a speed much higher than that of the wedge. The force time history after the initial peaks does not exhibit any plateau and oscillates close to zero. After the test, the strips appear almost undeformed, i.e., little plastic deformation takes place, and the measured energy is much lower than in the previous case. According to the ISO standard, the average cleavage force can be calculated from the time history by considering the data between 25% and 90% of the time interval from the beginning to the end of the curve. The energy is given by the area under the curve between the same end points multiplied by the test speed. The reason for this assumption is to discard the data corresponding to the initial peaks and the “tail” of the curve. Blackman et al. (2000) remark that such procedure works well in case of stable crack growth, but it fails and can lead to false results in case of unstable growth, which does not exhibit any plateau. An important aspect of the impact wedge peel test is that in case of stable crack growth; part of the energy is spent for deforming the metal strips (this aspect does not affect the unstable crack growth, which, however, is not suitable for the abovementioned reasons). Thus, also with this test, the measurement does not yield a property of the adhesive in se but of the assembly formed by adhesive and strips. Another factor that potentially affects the results is the friction between the wedge and the surfaces of the specimen during the fracturing process. However, the finite element simulations performed by Blackman et al. (2000) show that this aspect has minor importance, the leading factor being the adhesive fracture energy Gc. From the knowledge of the latter parameter, which can be measured with a fracture mechanics test such as the double cantilever beam, the time history of the impact wedge test can be reconstructed with finite element modeling.

21

Impact Tests

21.4

563

Hopkinson Bar Apparatus

In addition to the limitations already mentioned, the two tests described in the previous sections present the further restriction that the strain rate attained in the adhesive cannot be much high, since it usually does not exceed 102 s
1. When testing at higher strain rates is needed, the suitable equipment is the so-called Hopkinson bar, which in the last years has become probably the most used by the researchers of the field. This name is due, historically, to the apparatus used at the beginning of the twentieth century by B. Hopkinson to study the pressure waves generated by explosions or projectiles. In the experiment, the explosion (or the projectile) attained one end of a long and thin bar suspended horizontally, creating a pressure pulse which propagated through the bar until the other end, where a short cylinder of the same diameter of the bar was provisionally attached by means of a grease film or magnetically. This cylinder, called time-piece, was projected by the pulse against a ballistic pendulum which gave the measure of the momentum contained in it; the momentum remained in the bar was measured by observing its oscillation. By varying the length of the time-piece, it was possible to assess the length of the pressure pulse. The version of the apparatus used nowadays was introduced by Kolsky (1963), who added a second bar, from which the name “Split-Hopkinson pressure bar” comes from; the specimen of material to be tested is inserted between the two bars, as shown schematically in Fig. 5a. The projectile, usually fired by means of a pneumatic gun, impacts the first bar (incident bar), generating the incident pulse which, at the bar/specimen interface, is partially reflected and partially propagates in the specimen. From the specimen, the pulse is transmitted to the second bar

a

Strain gauges

Projectile

Second (transmitter) bar

First (incident) bar Specimen

b

Incident pulse

Reflected pulse

Distance

Transmitted pulse

Time

Fig. 5 Schematics of the Split-Hopkinson pressure bar (a) and related Lagrangian diagram (b)

564

L. Goglio

Fig. 6 Split-Hopkinson pressure bar: (a) overall view of the rig including three bars for different test conditions; (b) detail of the bending fixture (Politecnico di Torino, DYNLAB)

(transmitter bar). The situation is described graphically by the so-called Lagrangian diagram presented in Fig. 5b. A concrete example of Split-Hopkinson pressure bar is shown in Fig. 6. The pulses are measured by means of strain gages placed on both incident and transmitter bar, thus their time history can be stored by means of a transient recorder, usually a digital oscilloscope or an acquisition board. From such measurements the stress (σ), strain (e), and strain rate ðe_Þ in the specimen can be obtained as A0 etra ðtÞ A ð c0 eðtÞ ¼
2 eref ðtÞdt L σ ðtÞ ¼ E0

e_ðtÞ ¼
2

c0 eref ðtÞ L

(1) (2) (3)

21

Impact Tests

565

where, respectively, E0, A0, and c0 are Young’s modulus, cross section, and wave velocity of the bar, A and L are the cross section and the length of the specimen, etra is the strain pulse transmitted in the second bar, and eref is the strain pulse reflected in the first bar. Equations 1, 2, and 3 are based on the assumption that the specimen has reached equilibrium conditions; this requires that its length be smaller than the length of the pulse. Clearly, the measurement cannot be taken directly at the interface; thus the strain gages are placed at a distance from the end of each bar; this is potentially a source of error since the pulses during their travel (from the bar end to the gage) change in shape and assume a delay that must be evaluated. As far as the adhesives are concerned, the Hopkinson bar has been applied to the characterization of materials and also butt or lap joints. The arrangement described above is suitable for testing in compression (that can be applied by simple contact); in case of tension, the system must be modified, usually by means of a threaded connection. Moreover, it is required to generate a tensile pulse in the bars, and this can be obtained in two ways. The first is to use a reverse loading system, so that the impact of the projectile applies tension to the first bar, as done by Yokoyama (2003) and also (actually to test polymers, but the case would apply also to adhesives) by Chen et al. (2002). The second and more popular solution is to generate, as usually, an initial compressive pulse that travels through both bars; the transmission of this pulse from the first to the second bar is given by a split ring, inserted between them, which prevents the specimen from being loaded in compression (Fig. 7). It is also possible to avoid the use of the split ring if the dimensions of the threaded parts of specimen and bars are accurately manufactured to obtain a precise fit of the ends of the bar (Fig. 8). When the compressive pulse reaches the end of the second bar, it is

Specimen

Split ring

Fig. 7 Use of threaded specimen and split ring to carry out tensile tests with the Hopkinson bar (From Goglio et al. 2008, copyright Elsevier)

Fig. 8 Use of threaded specimen without split ring to carry out tensile tests with the Hopkinson bar: (a) tubular specimen; (b) solid specimen (From Bragov and Lomunov 1995, copyright Elsevier)

566

L. Goglio

reflected back as a tensile pulse which stresses the specimen, while the ring gives no contribution under tension. Several studies based on the Hopkinson bar have been presented by Yokoyama and co-workers. In Yokoyama and Shimizu (1998), the shear strength of pin-andcollar specimens (ASTM D4562) bonded with cyanoacrylate has been studied with the Hopkinson bar and compared to the static case. Both steel and aluminum have been considered as adherend materials. Later, Yokoyama (2003) has tested under tension butt joints made with cyanoacrylate adhesive and two different types of adherends, namely, a bearing steel and a high-strength aluminum alloy. The impact tests have been carried out on a Hopkinson bar; other comparative tests at intermediate and low loading rates have been performed on a standard test machine. The experimental work is supported by a finite element analysis of the stress state in the joint. In both studies, the main findings are that the joint strength, expressed as peak force or stress, increases with the loading rate, while the deformation decreases. There is an optimum value of adhesive thickness (about 35 μm), for which the tensile strength is maximized; furthermore, the strength is influenced by the material of the adherends, steel, or aluminum. In the case of butt joints also the absorbed energy has been evaluated, by integrating the load-deformation curve. It has been noticed that the amount of energy is influenced by the material of the adherends; as in case of steel adherends, it decreases significantly with the loading rate. Again, Yokoyama et al. (2012a, b) have studied with the Hopkinson bar the response of bulk specimens made of two different structural adhesives (epoxy and methacrylate) tested in compression. A tapered striker has been used to apply a nearly constant strain rate. Both adhesives have shown marked strain rate dependence, and it has been found that a modified Ramberg-Osgood model predicts well the stress-strain curve in loading, but not the strain softening after yielding and the unloading phase. Recently, Yokoyama and Nakai (2015) have tested brim-and-crown butt joints, fabricated with aluminum alloy or pure titanium and joined by an epoxy. A Hopkinson apparatus modified by adopting a tubular output bar has been used. The experiments have shown that the tensile strength increases significantly with the loading rate and decreases under increasing adhesive thickness, for both adherend types. Goglio et al. (2008) have tested a bicomponent epoxy under tension and compression, to assess the effect of the strain rate on the strength of this adhesive. Also, in this work both the Hopkinson bar and a hydraulic test machine have been employed, depending on the desired strain rate (from 10
3 to 3  103 s
1). Under compression, the stress-strain curve of the adhesive exhibits an initial linear (elastic) rise, then a wide plateau, and again a rise with increasing slope. These features remain the same under any strain rate values, as it can be observed in Fig. 9: increasing the strain rate, the yield stress increases; thus the plateau is shifted upward and starts at larger strain. This observation has led to define a dynamic factor as k = σ dynamic/σ static, obtained by the ratio (averaged over the plateau) of the points on the curve at a chosen rate to the points on the static curve. The available data have been used to define prediction formulae for k as a function of the strain rate, based on

21

Impact Tests

567

400 (a) Dynamic test (3.0⫻103 s–1) (b) Dynamic test (2.0⫻103 s–1)

350

(c) Dynamic test (1.5⫻103 s–1)

Stress (MPa)

300

(d) Dynamic test (1.2⫻101 s–1) (e) Dynamic test (5.0⫻100 s–1)

250

(f) Dynamic test (5.0⫻10–1 s–1) a

(g) Static test (1.0⫻10–3 s–1)

200

d

150

b

e

c

100

g f

50 0

0

0.05

0.1

0.15

0.2

0.25 Strain

0.3

0.35

0.4

0.45

0.5

Fig. 9 Dynamic stress-strain curves under compression at different strain rates, case of cold cured adhesive (From Goglio et al. 2008, copyright Elsevier)

a log-log fit as shown in Fig. 10. The two popular models, to account for the strain rate influence, of Cowper-Symonds and Johnson-Cook have been applied to this case. None of them can reproduce the experimental data over the whole range tested (10
3-103 s
1); thus they can be conveniently replaced by a poly-linear (in log-log scale) fit in the form k ¼ Ae_n , where A and n are coefficients determined over each chosen range. From Fig. 10, it can be also appreciated that the k factor assumes, at the highest strain rate tested, values close to 3; this gives a measure of the elevated sensitivity of the adhesive to the strain rate. Testing under tension at high strain rate presents particular problems. Apart from the abovementioned modifications to the apparatus required to apply positive stress, the measurement is hindered by the different response of the material. Indeed, the material usually behaves as brittle, and the specimen cracks soon after the stressstrain curve has reached its maximum. The only value that can be obtained is in practice the 0.2% (or other chosen percentage) proof stress. However, this value is also strongly dependent on the strain rate. There are other specific problems, which affect testing with the Hopkinson bar, that deserve to be mentioned. One concerns the first part of the stress-strain diagram, corresponding to the elastic behavior and, therefore, useful to evaluate the elastic modulus of the material. This part of the diagram corresponds to a time range of the test in which the equilibrium conditions still have not been achieved in the specimen, since as a

568

L. Goglio

3 Experimental data Cowper–Symonds model

Dynamic factor

Johnson–Cook model

2

1 10–4

Polylinear fitting

10–2

100

102

104

–1

Strain-rate (s ) Fig. 10 Dynamic factor k (case of cold cured adhesive): fitting of the experimental values with Cowper-Symonds, Johnson-Cook, and poly-linear models (From Goglio et al. 2008, copyright Elsevier)

practical rule it can be assumed that equilibrium is achieved after the waves have traveled back and forth more than three times along the specimen (Chen et al. 2002). Thus, the values of the elastic modulus can be affected by uncertainty and must be regarded with care. A general survey – not specifically related to adhesives – of the problems concerning the measurements by means of the Hopkinson bar has been presented by Bragov and Lomunov (1995). Apart from presenting nonconventional applications of the Hopkinson bar (such as dynamic hardness and toughness testing), these authors discuss the problems of signal synchronization and strain rate constancy. The former problem is originated by the difficulty in recognizing the exact beginning of the three measured pulses (incident, reflected, transmitted), which in this paper is corrected by checking that the sum of the incident and reflected pulses equals the transmitted strain pulse. The latter problem is due to the fact that the strain rate, according to Eq. 3, is proportional to the reflected pulse, which decreases as the specimen hardens and its cross section increases during the test. The remedy is to modify the shape of the incident pulse, by means of a tapered projectile or an auxiliary specimen (pulse shaper) placed at the impacted end of the first bar: in this way the incident pulse assumes an increasing amplitude, while the reflected pulse tends to be constant. Zhao and Gary (1996), presenting a study related to the measurement of small strains with the Hopkinson bar, have considered two important sources of error. The first is the dispersion (related to the difference between phase and group velocities of the waves), which the authors correct with a numerical procedure. The second is the determination of the time delay between the signals, for which an iterative procedure

21

Impact Tests

569

is proposed. These authors also propose an identification technique to treat the data corresponding to the initial transient of the measurement, when the equilibrium in the specimen is not yet attained. In a general perspective, not solely related to testing of adhesives, the problem of the dispersion has been considered by Tyas and Watson (2001). They have proposed to correct the measurement by means of two factors, termed M1 and M2, one relating measured surface strain to average strain over the cross section and the other relating average axial strain to average axial stress (dynamic elastic modulus). Another source of error is the effect of the nonuniform distribution of stress and strain over the cross section, related to radial inertia. This problem, considered again in general case by Forrestal et al. (2007), becomes important when the diameter of the bar is large and, therefore, the wave is no longer mono-dimensional. The corrections which are proposed are based on the theoretical solution for the wave propagation. Sen et al. (2011) have carried out tests on single-lap joints of different overlap lengths; they have found that the (apparent) strength expressed as average stress decreases by increasing the overlap – in accordance to what is found also under static loading – and that the edges of the adhesive are subjected to the maximum strain. An important remark about this type of measurements is that, as stated previously, in Hopkinson bar testing, the stresses should be uniform throughout the length of the sample, which is more difficult to achieve in case of lap joints. Similarly, Challita et al. (2011) have carried out quasi-static, intermediate, and high strain rate tests (the latter on a Hopkinson bar) on double-lap joints, using compression-shear and tensile-shear specimens. It has been found that the shear strength increases for increasing strain rate until 103 s
1 then decreases. Correction coefficients based on the previous work Challita and Othman (2010) have been used to estimate the maximum stress, accounting for the nonuniform distribution. Testing with the Hopkinson bar has been applied also to nonconventional cases, to assess the material response in special conditions. For instance, Martinez et al. (1998) have studied the behavior of a confined adhesive, i.e., when it cannot freely expand laterally under compression. The interest for such case is related to the use of an adhesive to join, in an armored panel, the front ceramic layer to the metal backing plate. To this aim, the adhesive specimen in the Hopkinson bar tests has been surrounded by a confinement tube. The study, carried out on several types of adhesives, has evidenced an optimum value of the adhesive thickness in the armor, in order to ensure enough strength to contain the ceramic fragments and the minimum reflection of the pressure wave. A work by Adamvalli and Parameswaran (2008) has studied the influence of the temperature on the strength of single-lap joints, bonded with an epoxy adhesive, tested at different loading rates (Fig. 11). It has been found that, on the one hand, at the highest loading rate tested (4.7 MPa/μs), the strength is more than three times the corresponding static value; on the other hand, an increase of temperature from 25  C to 100  C reduces by 35% the strength. Ravi Sankar et al. (2015) have used a Hopkinson bar to test lap joints with splitcylinder geometry, made with similar or dissimilar metals (steel, aluminum) joined by an epoxy. The results, supported by finite element modeling to assess the stress

570

L. Goglio Lap joint specimen

1

Incident bar Pulse shaper Striker bar

2

Transmitted bar

Strain gage

1

2

Strain gage

Fig. 11 Setup of the Hopkinson bar for testing lap joints at high temperature (From Adamvalli and Parameswaran 2008, copyright Elsevier)

distribution, indicate that the strength is dictated by the less stiff adherend. A further interpretation by these authors is that the strength measured using the Hopkinson bar represents the volume averaged shear stress in the adhesive layer, as stated by Challita and Othman (2010). The method of the Hopkinson bar testing has also been applied to specimen types other than cylinders or lap joints. Sato and Ikegami (1999) have measured the dynamic strength of an epoxy adhesive by testing butt joints of tubes subjected to shear and normal stress. Their equipment (Fig. 12) include a Hopkinson bar suspended vertically and loaded at the upper end by hydraulic actuators which apply, before the beginning of the test, a torsional and a tensile preload. Nearby the lower end, the bar is clamped by a mechanism that can be released rapidly. The specimen is formed by an upper short tube (load input tube), screwed at its upper end to the Hopkinson bar, and a long lower tube (load output tube), instrumented with strain gages. When the clamp is released, the elastic energy accumulated in the bar creates the waves that travel from one tube to the other, thus loading the adhesive layer. The reason for an elevated length of the output tube is to avoid that the back reflection of the waves at its lower end could interfere with the useful signals. The result is a chart, of the type shown in Fig. 13, reporting the failure conditions under different combinations of peel and shear stress. Also, cases of failure under static loading are reported in the diagram, for comparison with the dynamic case; it is evident that the stress values under impact are approximately double than the static ones. The authors have also tried to interpolate the experimental points with the popular Tresca and von Mises failure criteria, but the best fit (solid lines in the figure) is given by a polynomial in the form   0 0 0 τd2 σ d
σ d þ σ d τ2d ¼ 0

(4)

where σ d and τd are the applied normal and shear stresses and σ d0 and τd0 are the strength values measured experimentally. This formula, although not supported by a theoretical background, has also the merit of predicting an increase of the bearable

21

Impact Tests

Fig. 12 Schematic of the modified Hopkinson bar used by Sato and Ikegami (1999) to test butt joints of tubes under tension and torsion

571 Tensile and torsional actuators

Hopkinson bar

Clamping system

Input tube

Adhesive

Strain gages

Output tube

shear stress in case of negative (i.e., compressive) normal stress; this can be noticed by extrapolating to the left the curves in Fig. 13. An evolution of the Hopkinson bar is the test rig presented by Kihara et al. (2003) to measure the shear strength of adhesive layers (Fig. 14). A prism of hexagonal cross section, screwed between bars 1 and 2, is bonded to a pair of plates which, in turn, are screwed to the additional bars 3. When the pulse, generated from top by the impact of the projectile 0 against bar 1, travels through the prism, it propagates also laterally to the bonds and the plates. The skew slit contained in each plate acts as a reflector which deviates most of the pulse toward bar 3. In addition to the strain gages placed on bar 1 and 2, as usually in the Hopkinson bar scheme, a rosette is applied on each lateral plate to measure the strains in a point close to the adhesive layers. A peculiar phenomenon noticed by the authors is that in case of low incident pulse (in bar 1), the adhesive withstands the shear stress and is fractured by the tensile normal stress generated by the reflection in the plate. Conversely, in case of high incident pulse, it is the shear stress (in combination with compressive stress) that causes the fracture of the adhesive. By means of this apparatus, the authors have

572

L. Goglio 60

Shear stress (MPa)

50 Dynamic tests

40 30 20

Static tests

10 0 0

10

20

30

40 50 60 Peel stress (MPa)

70

80

90

Fig. 13 Limit curves for dynamic and static stress of tubular butt joints subjected to normal and shear stresses based on the results of Sato and Ikegami (1999)

obtained a value of shear strength comparable to that obtainable by cylindrical butt joints subjected to torsion. Wang and Xu (2006) have performed tensile dynamic tests with the Hopkinson bar on convex-edge joints, shaped in such a way to eliminate the edge singularity in the interface of dissimilar materials, so obtaining a nearly uniform stress distribution and a more correct measurement of the strength. This specimen type is not exactly a bonded joint; however the findings can be useful to design a butt-bonded specimen. Gilat et al. (2007) have studied the response of two epoxy resins, one untoughened and one toughened, in the strain rate range 10
5–700 1/s, under tensile and torsion loading. Tests at low and medium strain rates have been carried out on a hydraulic (biaxial) machine, those at high strain rates with a tensile or a torsional Hopkinson bar. The latter have been adapted by clamping, respectively, a force or a torque which is then released. A transition from ductile to brittle behavior has been noticed under tension under increasing strain rate; conversely, ductile behavior has been observed under torsional shear at any strain rate. In a comprehensive work concerning the mechanical properties of the crashoptimized adhesive Betamate 1496 V, May et al. (2015) have used a modified Hopkinson apparatus in which no input bar is foreseen: the striker hits directly the specimen, and the force is measured by a film transducer attached to the end of the output bar. The deformation of the specimen is captured by means an electro-optical system based on a high-speed camera which keeps track of a pair of markers painted on the edges of the specimen. Very recently, Neumayer et al. (2016) have used a Hopkinson bar to study the impact response of butt- and lap-shear joint specimens, measuring the strain in the adhesive by means of Digital Image Correlation (DIC), achieving a better accuracy than in traditional way. It has been noticed that DIC has given a 12.28 times higher

21

Impact Tests

573 φ19

M6

400

Adhesive layer

80

34 10

R

.4

P

0

19 Impact face

Neck

14.0

Gauge1

355

M4

18.1

10.8

1

Distortional wave

115

15

30

19.4

Adhesive layer

Dilatational wave

(Details)

P

3

2

P

3

345

95

x

R

Gauge1 y Gauge2 φ12

Unit: mm

Rubber V1

Vo Surface plate

Fig. 14 Test rig to measure the impact strength of adhesive layers (From Kihara et al. 2003, copyright Elsevier)

574

L. Goglio

adhesive stiffness and a 1.83 times lower energy in case of butt joints (respectively, 6.13 and 1.29 times in case of lap-shear joints).

21.5

Other Test Methods

In addition to the test schemes described in the previous sections, some other methods have been applied to test adhesives and joints under impact loading. Most of them are based on schemes that reprise static cases, for the geometry of the specimens or the loading system. The reason for this is not only to recover known cases but – most of all – to evaluate easily the changes in the joint response induced by the loading rate.

21.5.1 Fracture Energy Regarding the measurement of the adhesive fracture energy Gc, under impact (or, in general, dynamic) conditions, the double cantilever beam test and the tapered double cantilever beam test (Fig. 15) must be adapted to account for the effect induced by the loading rate. Indeed, in statics the base for the measurement is the formula (Williams 1984): Gc ¼ Fig. 15 Double cantilever beam specimen (a) and tapered double cantilever beam specimen (b)

a

P2c dC 2b da

(5)

P Loading block hs

P

b P

P

hs

21

Impact Tests

575

where Pc is the critical load for crack propagation, b is the joint width, a is the crack length, and C is the specimen compliance, i.e., the displacement to force ratio (δ/P). In case of an impact test, the main problem which arises in evaluating the fracture energy is that (as already remarked previously) the force-displacement curve exhibits strong oscillations and, therefore, it is tricky to read from it the relevant information. To avoid this problem, expressions that do not directly make use of load values must be adopted, such as for the double cantilever beam specimen (Kinloch 1997): Gc ¼

3 F δ2c h3s Es 16 N 2 ða þ χ 1 hs Þ4

(6)

where F and N are correction factors which, respectively, account for the large deflection of the beam and the stiffening effect of the loading block at the ends of the arms, χ 1 is another correction factor accounting for beam root rotation, and hs and Es are, respectively, the height of the beams and Young’s modulus of the substrate material. The critical displacement for crack propagation δc and the crack length a can be monitored by means of an optical system, as long as the test goes on. In the case of the tapered double cantilever beam specimen, the formula accounts also for the variable beam height. A sophisticated analysis about the measurement of the fracture energy at high rate of loading has been presented by Blackman et al. (2009). These authors have used the double cantilever test and the tapered double cantilever test to evaluate the behavior of an epoxy adhesive, under loading conditions varying from quasistatic to 15 m/s. The experimental setup includes a hydraulic actuator to apply the load, a high-speed camera, and the suitable electronic devices for measuring, amplifying, processing, and storing the data (Fig. 16). Four different types of crack growth have been identified in this work, namely, (1) slow-rate stable, (2) slow-rate unstable, (3) fast-rate unstable, and (4) fast-rate stable. Here, “stable” is termed the case of steady propagation, while “unstable” is a discontinuous growth, showing sequences of arrest and propagation. It is interesting to notice that the unstable crack growth can occur at intermediate load rate in the tested range; the difference between slow rate and fast rate depends on whether the kinetic energy of the specimen arms is negligible or not with respect to the fracture energy of the adhesive. It is also remarkable that, at the highest load rate adopted in the tests, the crack growth is again stable; obviously in this case the dynamic effects (importance of the kinetic energy, uncertainty on the measured load) are important. For instance, in case of fast-rate loading and double cantilever beam specimen, the formula for the corrected fracture energy (i.e., accounting for the dynamic effects) is GIc ¼

3 Eh3 ðV=2Þ2 t2 F 33 EhðV=2Þ2
4 ða þ ΔI Þ4 N 2 140 c2L

for type 3 crack growth and

(7)

576

L. Goglio

Instron high-rate VHS Ram position High-speed camera Camera trigger Lost motion device Camera output Mode I geometry Imatek system (Data acquistion and video synchronization)

Load

Gain amplifier

Piezoelectric load cell

Digital image and load signal analysis

PC Fig. 16 Experimental setup for high rate testing of adhesives under high rate of loading (From Blackman et al. 2009, copyright Elsevier)

GIc ¼

3 Eh3 ðV=2Þ2 t2 F 111 EhðV=2Þ2
4 ða þ ΔI Þ4 N 2 280 c2L

(8)

for type 4 crack growth. In both Eqs. 7 and 8, E is Young’s modulus of the substrate material, h is the height of the beam, V is the velocity applied by the actuator, t is the time, a is the crack length, ΔI is the correction factor for beam root rotation, F and N are – again as in Eq. 6 – the correction factors for large deflection of the beam and stiffening effect of the loading blocks, and cL is the longitudinal wave speed in the substrate material. The first term in Eqs. 7 and 8, which can be obtained from Eq. 6, has the meaning of energy calculated as in static conditions, from which the second term, corresponding to the kinetic energy of the beams, is subtracted. Analogous formulae, of higher complexity due to the non-constancy of the beam section, are available for the tapered double cantilever beam specimen (see again Blackman et al. 2009).

21

Impact Tests

577

21.5.2 Drop-Weight-Based Tests and Alternative Tests Also in the field of adhesive joining, drop-weight machines are probably the test rigs that have been used first to generate impact conditions. Beevers and Ellis (1984) have carried out tests on joint specimens fabricated both from mild steel and aluminum coupons, bonded with a toughened epoxy. A drop-weight machine equipped with a strain-gage transducer has been used for impact testing. Compared to the static case, at 4 m/s impact, the steel joints have exhibited a significant increase in ultimate load, while only a small change has been observed for the aluminum joints. This is ascribed to the different strain rate response of the two materials. Jordan (1988) has described the use of a special drop-weight testing machine, defined “instrumented guillotine.” Its peculiar features are the use of a counterweight to adjust the impact energy, from a maximum of 800 J to values lower than 100 J, and a striker of linear shape. The apparatus has been used to test sandwich-type specimens, comparing the impact behavior of epoxy- and acrylic-bonded joints with riveted or spot-welded ones. Regarding the measurement of the impact strength in terms of force or stress, a first case is represented by the simple butt joint of hollow cylinders, loaded axially in tension. A thorough evaluation of it has been presented by Sawa et al. (2002, 2003). In these works, impact tests have been carried out on specimens made of two hollow steel cylinders bonded together (Sawa et al. 2002) and a hollow steel cylinder bonded to a hollow aluminum cylinder (Sawa et al. 2003); in both works the adhesive is an epoxy, and the outside and inside diameters of the cylinders are 15 and 5 mm, respectively. Figure 17 shows the experimental setup; it can be noticed Fig. 17 Schematics of the impact test on cylindrical butt joints used by Sawa et al. (2002) and Sawa et al. (2003)

Adhesive

Adherends (cylinders)

Falling weight

578

L. Goglio

that the load is applied by releasing, from a chosen height, a weight that hits the flange of the coaxial guide bar which, in turn, applies the tension to the lower cylinder. The measurements are taken by means of strain gages placed on the surface of the cylinders, close to the bond. The papers describe in detail the finite element elastoplastic modeling of the experiments, which aims at assessing the state of stress and strain in the adhesive; in particular, the peak stress occurs at the outside edge of the interface, its value decreases as the adhesive thickness increases (while the opposite trend occurs under static loading). When the specimens are formed by two steel cylinders, fracture always occurs at the lower adhesive/metal interface, which is reached first by the pulse arriving from bottom; in case of dissimilar adherends, fracture occurs at the adhesive/aluminum interface. The impact energy is higher in case of steel-aluminum specimens than in case of steel-steel, about 1.7 times if the lower cylinder is of aluminum, and about 1.5 times if the upper cylinder is of aluminum. Conversely, the opposite behavior is observed for the peak load, which is higher in case of steel-steel specimen. Another notable case is the study of lap joints under different impact conditions. A reason in favor for this choice is that the single-lap joint is the commonest type of specimen used for static testing; thus it appears straightforward to test the same geometry under dynamic conditions, to evaluate the influence of the loading rate without introducing additional (and possibly misleading) effects due to a change of geometry. Goglio and Rossetto (2008) have tested single-lap specimens using a modified instrumented Charpy pendulum (Fig. 18). The specimen is mounted on the hammer by means of a fixture that pulls its front half; the back half of the specimen is connected to a transverse tail. When, at the end of the fall, the hammer reaches its lowest position, the tail impacts a pair of stoppers fixed to the pendulum base, and

a

b Striker

F

Details

1

C

1. U shaped bar 2. Plate 3. Connectors 4. Screws C tup T transducer

T

2

3

4

Specimen

F/2

F/2

Tail

Fig. 18 Charpy pendulum adapted to test lap joints: (a) overall view; (b) fixation of the specimen on the hammer (From Goglio and Rossetto 2008, copyright Elsevier)

21

Impact Tests

579

this causes tension in the specimen. This scheme is opposite to that used in other works, such as by Harris and Adams (1985), in which the specimen is fixed at its back half to the pendulum base; the hammer during its fall hooks the front half specimen, thus applying the traction. The hammer is instrumented with a piezoelectric load cell, so that the force can be measured and recorded. The scheme reproduces, in dynamic conditions, the usual stress state of the single-lap joints, with shear and peel stresses concentrated at the overlap ends. It is worth noting that in this kind of test, although the hammer speed is moderate (some m/s), for adhesive layers of some tenths of mm the magnitude of the resulting shearing strain rate is 103 s
1. In the work by Goglio and Rossetto (2008), several combinations of bonded length (3, 8, 12.5 mm), adherend thickness (1.5, 3 mm), and adhesive thickness (0.5, 1 mm) have been tested, with the aim of producing failure under different blends of stress. In all cases the adherends were of mild steel, and the adhesive was a bicomponent epoxy; failure occurs in the adhesive, while the adherends remain undamaged. The main result is represented by the graph of Fig. 19 in which (with the same style as of Fig. 13) the maximum values of stress (calculated with a monodimensional model) corresponding to failure conditions are reported. Also the points corresponding to failure under static loading are plotted for comparison. The increase of stress values changing from static to impact conditions is evident; regarding the effect of the adhesive thickness, it must be remarked that the thinner layer, under the same speed applied by the pendulum, undergoes a higher strain rate. A drawback of the use of a pendulum is that, due to the transfer of energy from the hammer to the specimen, the speed during the impact is not exactly constant (although the speed drop is usually limited). A solution to avoid this, used by Cayssials and Lataillade (1996), is to replace the pendulum with a heavy rotating

70

sR,max (MPa)

Impact, t = 1 mm

60

Impact, t = 0.5 mm

50

Static, t = 0.5 mm

40 30 20 10 0 0

10

20

30 40 50 tR,max (MPa)

60

70

Fig. 19 Failure locus described by normal and tangential maximum stresses; bars show 1 standard deviation (From Goglio and Rossetto 2008, copyright Elsevier)

580

L. Goglio Hammer approaching the specimen

Specimen in idle position Hammer impacting the specimen

Rotating wheel

Specimen in impact position

Fig. 20 Schematic of the inertia wheel used by Cayssials and Lataillade (1996)

wheel, termed “inertia wheel,” schematized in Fig. 20. Once the wheel (driven by a motor) has reached the desired speed, the specimen is set in position and struck by a hammer fixed at the wheel periphery. Due to the high moment of inertia of the wheel, the reduction of speed is negligible. Also Marzi et al. (2015) have used a rotary device, to assess the dynamic strength of bonded T-peel specimens. Their work has especially aimed at evaluating the performance of a 3D optical systems designed to measure the deformations. A further alternative to the drop-weight scheme is the use of the “quick crusher,” mounted on a standard tensile testing machine. With this device, the jaws are initially kept open, while the crosshead accelerates; then, they suddenly clamp the specimen, applying an impact load. Schiel et al. (2015) adopted it to assess the impact behavior of bonded joints between zinc-coated or painted sheet metal, typical of the automotive industry. They have found that, in joints fabricated with epoxy adhesives, the weak point is the metal-coating interface or the coating itself. Another mode in which a lap joint can be impact loaded is out of plane bending. The reason of interest for this case is that, although bonded panels are designed and tested considering their service condition (which, usually, causes in plane loading), the capability of sustaining unexpected transverse impact loads must be assessed as well. For instance, in aerospace construction cases of interest are impacts with debris (in service) and fall of tools (during manufacturing or maintenance). To validate a spring-mass model of a lap joint under three-point bending, Pang et al. (1995) have carried out low-velocity impact tests on joints of composite laminates, hit by the hemispherical impactor of a drop tower instrumented with semiconductor strain gages to measure the force. Two different overlap lengths and two different adherend thicknesses have been used in the experiments; also the drop height has been varied to obtain different velocities in the range 3.6–4.6 m/s approximately. In this way, the authors have measured the force time history in the

21

Impact Tests

581

Fig. 21 Schematic of the transverse impact tests carried out by Higuchi et al. (2002)

Falling weight Position of the strain gages

impacts and, in particular, the peak force, to correlate it with joint geometry and impact velocity. Similar experiments have been carried out by Higuchi et al. (2002) to validate the finite element analysis of a single-lap joint. The specimens used in this work were mild steel plates bonded with an epoxy adhesive; strain gages have been applied (in longitudinal direction) on the mid thickness of the adherends to measure the local strain during the impact (see Fig. 21); the impact velocity was about 1 m/s. The goal of this work is to assess the stress distribution on the adherend-adhesive interface which is obtained from finite element results, since it is not possible to measure it directly (the model is validated by comparing its results in points at mid thickness of the adherends, where experimental measurements have been obtained with strain gages). The main findings concern the maximum principal stress σ 1, which attains its maximum near the edge of the interface of the adhesive with the upper (impacted) adherend. It has been noticed that, unlike the case of static loading, such principal stress increases for increasing adherend Young’s modulus, overlap length, and adherend thickness; this is due to the fact that, increasing these parameters, the joint reacts with higher stiffness to the displacement imposed by the impactor. Conversely, increasing the adhesive thickness, the principal stress decreases, as it happens in the case of static loading. Vaydia et al. (2006) have considered the case of bonded composite panels, again impact loaded in three-point bending. In this work, the specimens were epoxycarbon strips manufactured from carbon fabric; three distinct types of adhesives were used, namely, two different bicomponent epoxies and a nanoclay-reinforced (7%) epoxy. The latter case is considered to test the strength of an adhesive composition suitable to improve its performance under harsh condition (moisture uptake), although such high nanoclay content is expected to reduce the mechanical properties. Impact is applied by the tip of a drop-weight machine, with the setup shown in Fig. 22; the velocity is in the range 1–2 m/s. An extensive finite element analysis was carried out to understand the response of the joint to this loading condition. Model results confirm that there is a strong stress concentration at the ends of the overlap; in particular the peel component has a negative peak at the end of the upper (i.e., impacted) adherend and a positive peak at the end of the unloaded adherend. This fact can be understood by means of a simple argument about the rigid body relative rotation of the two adherends induced by the impact. Also, the shear stress peaks at the ends of the overlap, but the intensity is lower. The examination of

582

L. Goglio

Support Adherend

Tup

Adhesive

150 mm

Fig. 22 Schematic of the transverse impact testing of bonded composites (From Vaidya et al. 2006, copyright Elsevier)

the broken specimen suggests that the crack is initiated by the shear stress and then propagated by the dominant peel stress. In addition to the strength of adhesives under impact, also the properties related to dynamic deformability are of interest. Sato and Ikegami (2000) have studied the cases of lap joints and scarf joints subjected to axial impact loads. Preliminarily, they have measured the attenuation of the axial strain waves in a special specimen of epoxy adhesive, manufactured (by molding the adhesive resin) as a cylinder of length 900 mm and diameter 20 mm. From these results, the complex compliance of the material was obtained and, in turn, used to identify the parameters of a five-element Voigt model of viscoelasticity. These material data have been introduced in the finite element simulation of the impact tests on the joints. The joints used in these tests were made by aluminum strips 4 mm thick – bonded with the abovementioned epoxy – and adopting three different kinds of geometry, namely, single lap (square ended), tapered, and scarf. The equipment used for carrying out the tests is shown in Fig. 23; the impact bar driven by the cam hits the upper end of the specimen applying the impact load. The main difference between the types of specimen is that in the lap joint, the offset of the adherend midplanes creates a bending moment, from which high stresses originate; in the tapered lap joint this effect is mitigated; for the scarf joint the adherends are aligned and the stresses propagate easily through the joint. The measurements confirm these remarks; in the tapered lap joints (overlap length 100 mm, taper length 50 mm), the peak stresses are reduced about 50% with respect to the single lap (overlap length 100 mm); in the case of scarf joint, the stresses are even much smaller and smoothly distributed, since the stress concentration is relieved. A comparison of the effect of the strain rate on the mechanical response of a brittle and a ductile adhesive has been carried out by Sugaya et al. (2011), using a drop-weight machine in the tests at higher rates. These authors have found that for the brittle adhesive, the elastic modulus is insensitive to the strain rate, while under increasing strain rate, the tensile strength and the strain at failure increase (the latter remained anyhow small, as only the elastic range was extended). The effect of the strain rate on the ductile adhesive is to increase the tensile strength and to reduce the strain at failure.

21

Impact Tests

583

Cam

Impact bar

Motor

Rubber insulator Specimen Configuration of impact system 600 mm

Impact system Applied load

3,000 mm

Frame

Base

Strain gauges

Adhesive joint

700 mm

Specimen

DC strain amp.

Digital storage scope

1,500 mm Fig. 23 Experimental equipment for impact testing of lap and scarf joints (From Sato and Ikegami 2000, copyright Elsevier)

Liao and Sawa (2011) have studied the strength of steel shafts bonded in epoxy hollow cylinders and subjected to axial push-off loads. To support their finite element modeling, they have carried out experiments on a drop-weight machine, and the strains in the cylinder have been measured with strain gages. Good agreement has been found between calculated and measured results. Al-Zubaidy et al. (2012) have tested experimentally the impact behavior of adhesive bonds between steel plates and CFRP sheets, mainly used for retrofitting existing steel structures. Double strap specimens with different numbers of CFRP layers, fabricated with an epoxy adhesive, have been subjected to impact loads up to

584

L. Goglio

5 m/s applied by means of a drop-weight machine. The results have shown the existence of an effective bond length, above which a further increase in overlap does not increase the ultimate load. Compared to the static strength, the impact strength has increased at loading rate 3.5 m/s and decreased at higher rates. Liao et al. (2013) have modeled single-lap adhesive joints with dissimilar adherends (aluminum-steel) subject to tensile impact and carried out experiments on a drop-weight machine. In the finite element calculations, the CowperSymonds model has been used to account for the strain rate dependence. Finite element and experimental results have agreed in showing that (unlike the case of static loading) under impact the failure starts at the interface of the stiffer adherend. By investigating the behavior of a phenolic resin adhesive blended with rubber, using a pendulum-like testing machine, Adachi et al. (2015) have found an experimental validation of the principle of equivalence between strain rate and temperature. They advance that impact strength could be predicted by applying the principle. Kadioglu and Adams (2015) have studied the behavior of a flexible viscoelastic tape for automotive application. Single-lap joints of high-strength steel have been tested on a pendulum equipped with a load cell and a laser for measuring the displacement. Compared to static response, an increase of more than twice in adhesive shear strength has been noticed at 1.4 m/s impact speed. The impact behavior of a pressure-sensitive adhesive has been studied by Hayashida et al. (2015) using butt joint and double cantilever beam specimens and measuring the deformations by means of a high-speed camera. Transparent adherends have been used to observe cavitation and fibrillation, which are the phenomena controlling ductility for this adhesive. It has been noticed that under increasing strain rate, the strength of the butt joints increases, while the fracture energy of the double cantilever beam specimens decreases. This is ascribed to the fact that at high strain rate, the pressure-sensitive adhesive becomes brittle due to viscoelastic effects. Silva et al. (2016) have studied joints for automotive use, fabricated by combining adhesives of different types, ductile or brittle, as sketched in Fig. 24. The ductile and soft adhesive is placed in the ends of the joint overlap, to smoothen the stress concentrations, while the brittle and hard adhesive is set in the middle of the joint. Impact tests carried out with a drop-weight machine have shown that such a design exploits the best of the two kinds of adhesives, in terms of ultimate load (as shown in Fig. 25) and energy.

Fig. 24 Schematic of mixed adhesive joint

Upper adherend Flexible adhesive

Flexible adhesive Lower adherend

Rigid adhesive

21

Impact Tests

585 24,00

24 22

19,18 16,79

20

16,36

Load (kN)

18 16 11,75

14

10,00

10,15

12 10

10,67 9,23

8,20

8

4,36

5,10

6 4 2 0,69 0

8

2 ER N +X

D

P-

80

R

05

TV

05

6

10

XN

52

V1 +A

6

R

38

13

V +A

68

2

E-

2 85

80

D

38

1 AV

P-

R

00

8 P-

D

5

06

1 TV

Static

Impact

Fig. 25 Average maximum load comparison between static and impact conditions (From Silva et al. 2016, copyright Latin American Journal of Solids and Structures)

21.6

Conductive Adhesives

A field of application that has received attention in the last two decades is the use of conductive adhesives to replace soldering in electronic applications. Such change can bring advantages mainly of ecological type, eliminating the dangerous substances that can be produced by the soldering process and the need for energy to melt the solder. An important advantage given by adhesive joining is that the pitch between neighboring contacts can be shorter compared to the case of soldering, and this is beneficial as far as miniaturization is concerned. Apart from the technological differences introduced by adopting a new technology, the main concerns related to the use of adhesives in this field are unstable contact resistance and impact strength. Zwolinsky et al. (1996) have reported results obtained by a project of the National Center for Manufacturing Sciences (NCMS), carried out in the early 1990s to assess the suitability of 25 conductive adhesives available on the market at that time. In particular, regarding the assessment of the impact strength, experiments of two types were performed. In the first phase of the project, specimens made with a plastic leaded chip carrier (PLCC) bonded on a coupon were dropped repeatedly from two

586

L. Goglio

different heights, respectively, 0.457 m (18 in.) and 0.914 m (36 in.), recording the number of drops needed to detach the PLCC for each height. In the second phase, drop tests were carried out in specimens formed by a board containing a collection of components, joined by bonding. Each specimen was subjected to three drops from the height of 0.914 m (36 in.). The presence of damage due to the impacts is noticed both by visual inspection and by measurement of the electrical resistance. The result of this series of tests was that the commercial adhesives used at that time were not capable to withstand the impact condition. The authors recommend evaluating the impact performance of an adhesive with a series of 6 drops from 1.52 m (60 in.), with a specimen formed by a bonded PLCC. This recommendation is one of the bases of the NCMS specification for solder replacement adhesives. Another study of the impact performance of the conductive adhesives was carried out by Daoquiang and Wong (1999). To overcome the poor impact performance of the commercial conductive adhesives, these authors have proposed an adhesive of new formulation containing epoxy-modified polyurethane resin and bisphenol-F. The adopted impact test method is again the one proposed by NCMS and mentioned above (Fig. 26); the channel in which the specimen under test is dropped has the aim of avoiding rotation, thus forcing the part to impact the ground with a defined edge. In their work, these authors have also measured the internal damping of the adhesive, and they have noticed that the materials having high loss factor (tan δ) are also the most capable to withstand the impacts. It is evident that the abovementioned experiments obtain, as outcome, the simple result of failure or survival of the specimens. No insight is found about the stress state in the bonds; furthermore, the drop test assesses the strength of the bonded

Fig. 26 Schematics of the NCMS drop test for bonded electronic components

Specimen under test

Channel

60” ≈ 1.52 m

Floor

21

Impact Tests

Fig. 27 Falling wedge test arrangement for conductive adhesives (From Xu and Dillard 2003, copyright IEEE)

587 Guide columns Drop weight

Polycarbonate wedge

Steel pin Piano hinges DCB sample

Main view

Side view

specimen, not of the adhesive itself. More recent works aim at studying the impact behavior of the bonded electronic parts with the methods of fracture mechanics or structural dynamics. Wu et al. (2003) have tested surface mount joints in printed circuit boards, comparing the behavior of an isotropic conductive adhesive with a traditional lead-tin soldering alloy. Joint specimens have first undergone impact on a modified Hopkinson bar, specialized to create three-point bending conditions. Then, after the impact, the specimens (if not already failed) have been destructively tested under shear to assess their residual strength. It has been noticed that the residual shear strength of the joints decreases under increasing strain rate. Compared to the soldering alloy, the adhesive has exhibited lower strength, which nonetheless can be considered satisfactory for practical applications. Xu and Dillard (2003) have studied the impact strength of conductive adhesives by means of a new impact test, termed falling wedge test (Fig. 27). Its principle is similar to that of the double cantilever beam: a pair (because of the symmetry of the specimen) of wedges fall along a drop tower and with its motion splits apart the two substrates of the bonded joint, mounted vertically on the base of the test rig. To avoid the typical problems related to impact load measurement (already recalled in this section), no load measurement is carried out directly. In the test rig the deformation of the sample is captured by means of a high-speed camera, so that the displacement of the end points and the deformed shape of each beam are obtained from the images. The fracture energy is calculated as

588

L. Goglio

Gc ¼

9δ20 EI 4Ba4

(9)

where δ0 is the opening displacement, E is Young’s modulus of the substrate, I is the moment of inertia of the beam section, B is the beam width, and a is the crack length. It is easy to notice that Eq. 9 is a simplified version of Eq. 6, in which the correction factors are neglected. The specimens are coupons of printed circuit boards bonded in their central part (by means of a spacer); the adhesives are three different silver-filled epoxy adhesives, especially produced for this research. The tests have been carried out at several temperatures, between
70  C and 90  C, at a fixed impact speed of about 1.6 m/s. It has been observed that the results are influenced by the temperature when the glass transition temperature (Tg) of the adhesive is in the considered range; in particular a maximum of Gc has been noticed for a test temperature close to the Tg of the adhesive under consideration. These series of tests have confirmed the correlation between the loss factor of the material and the fracture energy. Rao et al. (2004), reprising the case considered in a work mentioned above (Daoquiang and Wong 1999), have added the study of the first natural frequencies and mode shapes of an electronic packaging formed by a PLCC bonded to a coupon of printed board. From the theoretical viewpoint, an ideal impulse would excite all natural frequencies of the system. This study has considered the first ten natural frequencies, whose range of values, obtained with a finite element model, is extended up to 2 kHz approximately. The impact behavior of the adhesive is adequate if its damping is high enough to dissipate energy over the whole range. The authors have measured the loss factor of four different adhesives, in a frequency range from 0.1 to 3,000 Hz. Comparing these measurements with the results obtained from the drop tests, it has been once again observed that the capability of withstanding the impact is related to the internal damping. A comprehensive state-of-the-art review on conductive adhesives has been presented by Irfan Mir and Kumar (2008), which considers the main related problems and, among these, the aspects regarding impact strength.

21.7

Response to Environmental Conditions

A quite peculiar case is that considered by Park and Kim (2010), who have tested single-lap joints between composite panels subjected to transverse impact of ice spheres, to reproduce the effect of hailstones against an aircraft or other structures exposed to the phenomenon. The experiments and the related finite element simulations have shown that damage occurs if the kinetic energy of the impactor exceeds a threshold. The most important failure mode is delamination, both for panels and joints; however, also adhesive failure can occur. Examination of the specimens has shown that the damage is mainly due to the stress concentration related to the joint. Thus, a joined assembly is more susceptible to this type of damage than a monolithic panel.

21

Impact Tests

589

Zhang et al. (2015) have studied the response to the strain rate after hygrothermal exposure in an experimental study on joints made of metal adherends (steel, aluminum) bonded with a toughened epoxy. By comparing the response at several crosshead speeds with the quasi-static conditions, they have noticed that: loading rate and aging can be regarded as independent influencing factors (with negligible interaction); the ultimate load is increased by loading rate and decreased by the hygrothermal exposure. The effects of these two factors can be accounted for by means of a linear empirical formula, to predict the resulting strength.

21.8

Conclusions

This chapter has described the main tests used to assess the impact strength of adhesives and joints, considering both those foreseen by international standards and those appeared in research papers. The main types of test rigs used in the experiments are pendulum, falling weight and Hopkinson bar, or evolutions of these. The major conclusions of this survey are the following: 1. Of the two standard test methods, the block impact (ASTM D950) is the easiest to carry out and does not require sophisticated equipment, but the only result it gives is the absorbed energy. Moreover, the specimen made of bonded blocks is far from the case of a typical joint, made of thin substrates; thus, this test can give (at best) only a comparison between different adhesive systems. The study by Adams and Harris (1996) has shown that even small errors of alignment can deeply modify the stress profile in the specimen; thus also the use of this test for comparison between adhesives is questionable. 2. The other test foreseen by a standard is the wedge peel impact (ISO EN 11343). Although in this case the arrangement is more representative of a real joint, the immediate answer given by the test rig is a property (energy) of the system formed by specimen and adhesive. Sophisticated data processing is required to extract further information and correlate the result of this test with the adhesive fracture energy Gc, as done by Blackman et al. (2000). These authors have also shown that in case of unstable crack growth, the ISO standard procedure to process the force time history can produce wrong results. They have also found that the effect of the friction between wedge and strips plays a minor role, compared to Gc. 3. When the mechanical properties of the adhesive at high loading rate (strain rate 103 s
1) are sought, the appropriate test rig is the Hopkinson bar. This apparatus is suitable to test specimens of adhesive under compression in straightforward manner; tests under tension can be carried out with modifications of the basic scheme. Also simple joints can be tested, by properly adapting the ends of the bars. In the last decade, this equipment has become more and more popular and adapted to tests specimens of various shapes (e.g., lap, pin-collar) under different loading conditions (e.g., tension, bending, shear).

590

L. Goglio

4. In the research works presented in the literature, many test rigs are based on the falling weight scheme; alternatively, the impact is applied by a hydraulic ram. In most cases, the specimens are single-lap joints, impacted axially (i.e., in tension) or transversally (i.e., in bending). Usually, in this kind of tests, the impact force is measured by a load cell embedded in the impactor, and the time history is recorded electronically. The typical problem of this measurement is that what is actually measured depends also on the dynamic response of the system formed by specimen and test rig; thus the observed oscillation is not only due to the start and growth of the crack in the joint. For this reason, when the determination of fracture mechanics properties (usually Gc) is sought, the use of force measurements is avoided, and the displacement is preferred instead. As a final remark, it can be added that frequently the experimental studies on the behavior of adhesive or joints under impact are presented together with finite element simulations. This interaction between testing and modeling is, in this field, especially needed, since the possibility of direct measurement of the desired quantities is limited and hindered by the complications introduced by the dynamic response of the tested system.

References Adachi T, Kataoka T, Higuchi M (2015) Predicting impact shear strength of phenolic resin adhesive blended with nitrile rubber. Int J Adhes Adhes 56:53–60 Adams RD, Harris JA (1996) A critical assessment of the block impact test for measuring the impact strength of adhesive bonds. Int J Adhes Adhes 16:61–71 Adamvalli M, Parameswaran V (2008) Dynamic strength of adhesive single lap joints at high temperature. Int J Adhes Adhes 28:321–327 Al-Zubaidy H, Al-Mahaidi R, Zhao X-L (2012) Experimental investigation of bond characteristics between CFRP fabrics and steel plate joints under impact tensile loads. Compos Struct 94:510–518 ASTM D950-03, Standard Test Method for Impact Strength of Adhesive Bonds. ASTM Int Beevers A, Ellis MD (1984) Impact behaviour of bonded mild steel lap joints. Int J Adhes Adhes 4:13–16 Bezemer AA, Guyt CB, Vlot A (1998) New impact specimen for adhesives: optimization of highspeed-loaded adhesive joints. Int J Adhes Adhes 18:255–260 Blackman BRK, Kinloch AJ, Taylor AC, Wang Y (2000) The impact wedge-peel performance of structural adhesives. J Mater Sci 35:1867–1884 Blackman BRK, Kinloch AJ, Rodriguez Sanchez FS, Teo WS, Williams JG (2009) The fracture mechanics of adhesives under high rates of loading. Eng Fract Mech 79:2868–2889 Bragov AM, Lomunov AK (1995) Methodological aspects of studying dynamic material properties using the Kolsky method. Int J Impact Eng 16:321–330 Cayssials F, Lataillade JL (1996) Effect of the secondary transition on the behaviour of epoxy adhesive joints at high rates of loading. J Adhesion 58:281–298 Challita G, Othman R (2010) Finite-element analysis of SHPB tests on double-lap adhesive joints. Int J Adhes Adhes 30:236–244 Challita G, Othman R, Casari P, Khalil K (2011) Experimental investigation of the shear dynamic behavior of double-lap adhesively bonded joints on a wide range of strain rates. Int J Adhes Adhes 31:146–153

21

Impact Tests

591

Chen W, Lu F, Cheng M (2002) Tension and compression tests of two polymers under quasi-static and dynamic loading. Polym Test 21:113–121 Daoqiang L, Wong CP (1999) High performance conductive adhesives. IEEE Trans Electron Packag Manuf 22:324–330 Forrestal MJ, Wright TW, Chen W (2007) The effect of radial inertia on brittle samples during the split Hopkinson pressure bar test. Int J Impact Eng 34:405–411 Gilat A, Goldberg RK, Roberts GD (2007) Strain rate sensitivity of epoxy resin in tensile and shear loading. J Aerosp Eng 20:75–89 Goglio L, Rossetto M (2008) Impact rupture of structural adhesive joints under different stress combinations. Int J Impact Eng 35:635–643 Goglio L, Peroni L, Peroni M, Rossetto M (2008) High strain-rate compression and tension behaviour of an epoxy bi-component adhesive. Int J Adhes Adhes 28:329–339 Harris JA, Adams RD (1985) An assessment of the impact performance of bonded joints for use in high energy absorbing structures. Proc Inst Mech Eng 199:121–131 Hayashida S, Sugaya T, Kuramoto S, Sato C, Mihara A, Onuma T (2015) Impact strength of joints bonded with high-strength pressure-sensitive adhesive. Int J Adhes Adhes 56:61–72 Higuchi I, Sawa T, Suga H (2002) Three-dimensional finite element analysis of single-lap adhesive joints subjected to impact bending moments. J Adhesion Sci Technol 16:1327–1342 ISO EN 11343 (2003) Adhesives – Determination of dynamic resistance to cleavage of highstrength adhesive bonds under impact conditions – Wedge impact method Jordan M (1988) The instrumented guillotine impact testing apparatus. Int J Adhes Adhes 8:39–46 Kadioglu F, Adams RD (2015) Flexible adhesives for automotive application under impact loading. Int J Adhes Adhes 56:73–78 Kihara K, Isono H, Yamabe H, Sugibayashi T (2003) A study and evaluation of the shear strength of adhesive layers subjected to impact loads. Int J Adhes Adhes 23:253–259 Kinloch AJ (1997) Adhesives in engineering. Proc Inst Mech Eng Part G 211:307–335 Kolsky H (1963) Stress waves in solids. Dover Publications, New York Liao L, Sawa T (2011) Finite element stress analysis and strength evaluation of epoxy-steel cylinders subjected to impact push-off loads. Int J Adhes Adhes 31:322–330 Liao L, Sawa T, Huang C (2013) Experimental and FEM studies on mechanical properties of singlelap adhesive joint with dissimilar adherends subjected to impact tensile loadings. Int J Adhes Adhes 44:91–98 Machado JJM, Marques EAS, Campilho RDSG, da Silva LFM (2017) Adhesives and adhesive joints under impact loadings: an overview. J Adhesion. https://doi.org/10.1080/00218464.2017.1282349 Martínez MA, Chocron IS, Rodriguez J, Sànchez Gàlvez V, Sastre LA (1998) Confined compression of elastic adhesives at high rates of strain. Int J Adhes Adhes 18:375–383 Marzi S, Biel A, Hesebeck O (2015) 3D optical displacement measurements on dynamically loaded adhesively bonded T-peel specimens. Int J Adhes Adhes 56:41–45 May M, Hesebeck O, Marzi S, Böhme W, Lienhard J, Kilchert S, Brede M, Hiermaier S (2015) Rate dependent behavior of crash-optimized adhesives – Experimental characterization, model development, and simulation. Eng Fract Mech 133:112–137 Mir I, Kumar D (2008) Recent advances in isotropic conductive adhesives for electronics packaging applications. Int J Adhes Adhes 28:362–371 Neumayer J, Kuhn P, Koerber H, Hinterhölzl R (2016) Experimental determination of the tensile and shear behaviour of adhesives under impact loading. J Adhes 92:503–516 Pang SS, Yang C, Zhao Y (1995) Impact response of single-lap composite joints. Compos Eng 5:1011–1027 Park H, Kim H (2010) Damage resistance of single lap adhesive composite joints by transverse ice impact. Int J Impact Eng 37:177–184 Rao Y, Lu D, Wong CP (2004) A study of impact performance of conductive adhesives. Int J Adhes Adhes 24:449–453 Ravi Sankar H, Adamvalli M, Kulkarni PP, Parameswaran V (2015) Dynamic strength of single lap joints with similar and dissimilar adherends. Int J Adhes Adhes 56:46–52 Sato C, Goglio L (2015) Guest editorial. Int J Adhes Adhes 56:1–2

592

L. Goglio

Sato C, Ikegami K (1999) Strength of adhesively-bonded butt joints of tubes subjected to combined high-rate loads. J Adhes 70:57–73 Sato C, Ikegami K (2000) Dynamic deformation of lap joints and scarf joints under impact loads. Int J Adhes Adhes 20:17–25 Sawa T, Suzuki Y, Kido S (2002) FEM stress analysis and strength of adhesive butt joints of similar hollow cylinders under static and impact tensile loadings. J Adhes Sci Technol 16:1449–1468 Sawa T, Suzuki Y, Kido S (2003) Stress analysis and strength estimation of butt adhesive joints of dissimilar hollow cylinders under impact tensile loadings. J Adhes Sci Technol 17:943–965 Schiel M, Kreling S, Unger C, Fischer F, Dilger K (2015) Behavior of adhesively bonded coated steel for automotive applications under impact loads. Int J Adhes Adhes 56:32–40 Sen O, Tekalur SA, Jilek C (2011) The determination of dynamic strength of single lap joints using the split Hopkinson pressure bar. Int J Adhes Adhes 31:541–549 Silva MRG, Marques EAS, da Silva LFM (2016) Behaviour under impact of mixed adhesive joints for the automotive industry. Lat Am J Sol Struct 13:835–853 Sugaya T, Obuchi T, Sato C (2011) Influences of loading rates on stress-strain relations of cured bulks of brittle and ductile adhesives. J Sol Mech Mater Eng 5:921–928 Tyas A, Watson AJ (2001) An investigation of frequency domain dispersion correction of pressure bar signals. Int J Impact Eng 25:87–101 Vaidya UK, Gautam ARS, Hosur M, Dutta P (2006) Experimental-numerical studies of transverse impact response of adhesively bonded lap joints in composite structures. Int J Adhes Adhes 26:184–198 Wang P, Xu LR (2006) Convex interfacial joints with least stress singularities in dissimilar materials. Mech Mater 38:1001–1011 Williams JG (1984) Fracture mechanics of polymers. Ellis Horwood, Chichester Wu CML, Li RKY, Yeung LH (2003) Impact resistance of SM joints formed with ICA. J Electron Packag 125:93–97 Xu S, Dillard D (2003) Determining the impact resistance of electrically conductive adhesives using a falling wedge test. IEEE Trans Compon Packag Technol 26:554–562 Yokoyama T (2003) Experimental determination of impact tensile properties of adhesive butt joints with the split Hopkinson bar. J Strain Anal 38:233–245 Yokoyama T, Nakai K (2015) Determination of the impact tensile strength of structural adhesive butt joints with a modified split Hopkinson pressure bar. Int J Adhes Adhes 56:13–23 Yokoyama T, Shimizu H (1998) Evaluation of impact shear strength of adhesive joints with the split Hopkinson bar. JSME Int J Ser A 41:503–509 Yokoyama T, Nakai K, Mohd Yatim NH (2012a) High strain-rate compressive properties and constitutive modeling of bulk structural adhesives. J Adhes 88:471–486 Yokoyama T, Nakai K, Mohd Yatim NH (2012b) High strain-rate compressive behavior of bulk structural adhesives: epoxy and methacrylate adhesives. JSME Int J A-Solid M 6:131–143 Zhang F, Yang X, Xia Y, Zhou Q, Wang H-P, Yu T-X (2015) Experimental study of strain rate effects on the strength of adhesively bonded joints after hygrothermal exposure. Int J Adhes Adhes 56:3–12 Zhao H, Gary G (1996) On the use of SHPB techniques to determine the behaviour of materials in the range of small strains. Int J Solids Struct 33:3363–3375 Zwolinski M, Hickman J, Rubin H, Zaks Y, McCarthy S, Hanlon T, Arrowsmith P, Chaudhuri A, Hermansen R, Lau S, Napp D (1996) Electrically conductive adhesives for surface mount solder replacement. IEEE Trans Compon Packag Manuf Technol Part C 19:241–250

Special Tests

22

David A. Dillard and Tetsuo Yamaguchi

Contents 22.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.2 Blister Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.3 Sealants and Elastomeric/Foam Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.4 Indentation Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.5 Scratch Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.6 Tack . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.7 CTE, SFT, and Residual Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

594 594 598 601 603 604 607 609 609

Abstract

This chapter gives a brief description of special mechanical tests for various types of materials and sample geometries, such as blister tests for membranes/adhesives/coatings, tensile tests and shear tests for sealants and foam adhesives, indentation and scratch tests for coatings, tack tests for pressure sensitive adhesives, and bimaterial curvature tests for characterizing residual stress, stress-free temperature, and coefficient of thermal expansion of adhesives bonded to substrates of interest. In addition, some applications of these tests are also described in this chapter.

D. A. Dillard (*) Biomedical Engineering and Mechanics Department, Virginia Polytechnic Institute and State University, Blacksburg, VA, USA e-mail: [email protected] T. Yamaguchi Machine Elements and Design Engineering Laboratory, Department of Mechanical Engineering, School of Engineering, Kyushu University, Fukuoka, Japan e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_22

593

594

22.1

D. A. Dillard and T. Yamaguchi

Introduction

In addition to standard adhesion tests or failure tests, many of which are described in other chapters of this handbook, a wide range of special tests are also performed for specific purposes in industry or in scientific research. In this chapter, several of these special mechanical tests are reviewed, including blister tests for membrane/adhesive/ coaing, tensile tests and shear tests for sealants and elastomeric foam adhesives, indentation tests and scratch tests for characterizing coating adhesion.

22.2

Blister Tests

The blister test is a method to determine the fracture (adhesion) energy between a substrate and a flexible adherend, which could consist of a plate or membrane adhered to the substrate or could simply be an adhesive layer or coating. The blister test was first introduced by Dannenberg for characterizing paint adhesion (Dannenberg 1961; Maugis 1999), although the more familiar form was introduced by Williams (1969). Blister tests are typically loaded by pressurizing the cavity between the substrate and flexible adherend and monitoring the growth of the debond as a function of applied pressure and, in some cases, time. Blister tests can be advantageous for some systems, including when pressurization may mimic actual application conditions, when relevant environments can be used as the pressurization media, and to avoid the need for mechanically gripping for loading. A wide range of blister configurations has been proposed and used, including some which are illustrated in Fig. 1. In the standard blister test, the flexible adherend is assumed to behave as a plate, with compiance based solely on plate bending. Assuming that the pressuredeflection relationship is linear, the fracture energy is easily obtained by determining the deflected shape at a given pressure using simple plate theory, resulting in:

Fig. 1 Illustrations of several blister test configurations Standard

Constrained

Pressure

Pressure

Blow-off

Shaft - loaded

Pressure

Peninsula

Island

Pressure

22

Special Tests

595

3p2 ð1  ν2 Þa4 32Eh3

G¼

(1)

where G is the applied strain energy release rate, p is the pressure, ν is the Poisson’s ratio of the plate, a is the radius of the debond, E is the modulus of the plate, and h is the thickness of the plate (Anderson et al. 1977). When the thickness of the flexible adherend is very large compared to the debond radius, the assumption of plate bending is no longer valid, and the solution approaches that of a penny-shaped crack between two bodies (Anderson et al. 1977). This results, for example, in G¼

3p2 a 2πE

(2)

for the the case of a semi-inifinite, incompressible adherend bonded to a rigid substrate. On the other hand, when the thickness of the flexible adherend is very thin, the resistance to deflection comes not from plate bending but primarily from membrane stretching, resulting in the membrane blister specimen (Gent and Lewandowski 1987). The membrane deflection solution is inherently nonlinear and more involved, and becomes dependent on the residual stress present within the membrane. For the membrane blister, the hole radius a0 (or in some cases, a larger debond initiated during fabrication) is given as the initial decohesion radius, and the membrane is inflated and effectively peeled off at the radius a (>a0). The relationship between the central deflection δ and the pressure p of a circular membrane with no residual stress is given by (Hencky 1915)
4 1=3 pa δ¼C Eh

(3)

where C is a constant dependent on the Poisson’s ratio (e.g. C = 0.662 for ν = 0.3 and C = 0.595 for ν = 0.5). The energy release rate G for this geometry is given by (Gent and Lewandowski 1987)
1=3 5CD p4 a4 5D pδ G¼ ¼ 4 4 Eh

(4)

where D = 0.518 for ν = 0.3 and 0.519 for ν = 0.5. As pressure is gradually increased, debonding is expected to occur at a critical pressure pc for a debonding energy of G c
pc ¼

4w 5CD

3=4

ðEhÞ1=4 a0

(5)

at a critical central deflection of
δc ¼

4wC3 5DEh

1=4 a0

(6)

596

D. A. Dillard and T. Yamaguchi

In all of these test configuratioins, debonding can often be induced between the adherend and substrate by injecting a gas (for example, nitrogen) or a liquid (e.g. water) under pressure through a small hole in the substrate, as shown in Fig. 2. This membrane form of the blister test has certain advantages in that mechanical testing is sometimes possible even for thin films which are, in general, difficult to evaluate with other mechanical testing approaches due to the soft or delicate nature of the film. Fracture mechanics theory can be applied to evaluate the fracture energy, and with appropriate analysis, accurate measures of debond energy can be obtained. Another advantage is the fact that the test geometry is similar to that in actual usage for coatings or paint, i.e., the evaluation of the film is possible in a geometry relevant to actual applications. While the blister test provides several advantages for the quantitative evaluation of the fracture energy, there are disadvantages. For example, in order to put holes to pressurize the film, fabrication processes of drilling or etching are often needed, and securing an apprpriate initial configuration that results in desired delamination without tearing or bursting the membrane can be involved. In some cases, an auxiliary adherend can be adhered on top of the membrane to reinforce it, though this influences the resulting deformation and energy analysis. For complex hole structures, such as the island and peninsula blister tests, discussed later, the process of undercutting the membrane can be even more complex. In addition, the blister test is limited to sufficiently flexible films like polymer coatings. For hard films, brittle fracture of the film can occur prior to the delamination under the applied pressure. Again, for such cases, a backing or auxilliary adherend can be added to reinforce the fragile coating, however, often resulting in a successful means to evaluate adhesion. In order to overcome such disadvantages, the blister test has been extended by several researchers. The inverted blister test (Fernando and Kinloch 1990; Jiang and Penn 1990; Fernando et al. 1993) has been developed to reduce the likelihood of cohesive failure or film bursting. Figure 3 schematically shows the inverted blister test. A sufficiently tough and flexible substrate of film geometry is delaminated from the coating or the adhesive adhered firmly onto the support plate, instead of delamination of the coated thin film from the substrate in the standard blister test. The island blister test (Allen and Senturia 1989), the peninsula blister test (Dillard and Bao 1991), and the constrained blister test (Chang et al. 1989) have been proposed. These tests are schematically shown in Fig. 1. In the island blister test, a small “inner island”induces larger driving force for the propagation of the delamination front than that on the periphery (the crack Fig. 2 Geometry of the membrane blister test. Fluid is injected thorugh a small hole of radius a0 under a pressure p and the membrane of thickness h is deflected by δand peeled off at radius a

h p

d

a 2 a0

22

Special Tests

597 Substrate

Fig. 3 Geometry of the inverted blister test Film

Support

initiation point for the standard blister test), which results in effective delamination at the interface between the film and the substrate island. The peninsula blister test is an extension of the the island blister test, developed to improve the stablity in that it is a constant energy release rate specimen. Both tests tend to induce high values of applied energy release rate because of the very localized debonding fronts (Lai and Dillard 1994; Liechti and Shirani 1994; Lai and Dillard 1996; Xu et al. 2006). In the constrained blister test, a transparent rigid cover plate prevents the film from bursting while permitting observation of the delamination front (Chang et al. 1989; Lai and Dillard 1990a, b). Theoretical treatment of some of these tests can also be found in Williams (1997). Although pressurization of circular or other confined cavities are popular for blister tests, mechanical loading has also been reported as with the shaft-loaded blister (Wan and Mai 1996) and is sometimes simpler to implement in the laboratory. Simple linear versions of blister tests are also popular. The central debonded region of a strip bonded along much of its length to a rigid substrate can be displaced from the substrate, inducing a peeling action. This configuration, known as the pull-off (Gent and Kaang 1986), strip blister (Liechti and Liang 1992), or V-peel (Wan 1999) test by various authors, is simple to perform and results in a particularly simple solution when the flexible adherend deformation is controlled by membrane stretching: G¼

3Fθ 8b

(7)

where F is the applied force, θ is the peeling angle in radians, and b is the width of the strip. One advantage of this geometry is that the relatively shallow debonding angles reduce plastic bending in the adherends, often resulting in practical or measured peel energies that are closer to the actual debond energy (Kinloch et al. 1994; Moidu et al. 1995, 1998b; Kinloch and Williams 2002; Kinloch et al. 2006). There are other variations of the blister test and the literature abounds with a range of solutions, so special care should be used for when applying the equations reported in the literature for various blister configurations. In particular, solutions often apply to either block-like (or semi-infinte), plate-like, or membrane-like behavior, as discussed above, so the relative thickness of the flexible substrate to the debond size is of critical importance. Transition regions have been considered by several

598

D. A. Dillard and T. Yamaguchi

authors (Anderson et al. 1977; Wan and Lim 1998). The transition from plate bending to membrane stretching has received some attention and becomes important when deflections become relatively large (say on the order of the flexible adherend thickness (Wan and Lim 1998). Finally, the presence of residual stress within membrane-like adherends can significantly affect deflections and energy release rate (Dillard and Bao 1991; Wan et al. 2003). Such residual stresses often arise when processing the materials to make specimens, and must be considered and characterized to achieve meaningful experimental results.

22.3

Sealants and Elastomeric/Foam Adhesives

Sealants share many similarities with other adhesives, but are often distinguished by their ability to fill gaps and undergo large strains while maintaining a sealing function, forming a barrier or protective layer against water, air, and other intruding media. There are a number of unique challenges for sealant formulation, joint design, and long term durability, however (Klosowski and Wolf 2015). Elastomer adhesives and foam pressure sensitive adhesive tapes are other types of adhesives, but also are used in thick bondline form. Many of the test methods and analysis procedures covered elsewhere in adhesion texts (Anderson et al. 1977; Kinloch 1987; Adams et al. 1997; Pocius 1997; Dillard 2002) and this handbook can be adapted to evaluate these softer and often thicker material systems, though complications can arise and test method adaptations are often employed. Here, the tensile or butt joint test, shear test, tensile adhesion test, and peel test are briefly reviewed with relevance to such materials. Figure 4a shows the geometry for the tensile butt joint test, also known as the poker-chip or button test. Two rigid cylinders are bonded end to end with the adhesive or sealant of interest, and the joint is tested by applying tensile forces to the cylinders, measuring the load at break and, in some cases, the load-displacement curves as well as. Alignment of these specimens can be problematic, as slight eccentricity or misalignment can significantly affect the stress state uniformity and the resulting rupture load (Anderson et al. 1988). Though the test method looks quite simple, special care is also required for the data analysis. In practice, the stress distribution at the adhesive-substrate interface is not uniform (see Fig. 4b), but depends on the elasticity (modulus and Poisson’s ratio Normalized) of the cylinders and the adhesive, and also strongly depends on the aspect ratio (thickness/diameter), especially for elastomeric adhesives. A range of solutions can be found, depending on the material and geometric parameters (Lindsey 1967; Anderson et al. 1977; Adams et al. 1978; Shephard 2002). Because of the significant difference in modulus between the substrate and adhesive layer, large triaxial tensile stresses arise within the bond. Assuming the adherends may be considered rigid, the strain can range widely, falling between the bound imposed for highly confined, uniaxial strain conditions approached as the thickness to diameter ratio approaches zero, and the uniaxial stress solution (i.e. very large adhesive thickness to diameter ratio), as respectively given by

22

Special Tests

599

z

a

b

Substrate

Adhesive

r

Stress at interface

Force

Normal stress sz

Substrate

Shear stress s rz

Force

Center

Edge

Fig. 4 (a) Geometry of a specimen of tensile test for butt joint. (b) stress distribution at the interface between adhesive and sustrate (solid lines for smaller Poisson’s ratio adhesive, dashes lines for elastomeric adhesive) a

b Substrate

Adhesive Substrate

Substrate

Adhesive Substrate

Substrate

Fig. 5 Geometry of a specimen of shear test for (a) the single-lap joint, and (b) the double-lap joint

ð1 þ νÞð1  2νÞ σ x σx  ex  ð1  νÞ E E

(8)

This shows that for highly constrained systems, Poisson’s ratio also becomes very important, as small deviations from ideal elastomer (ν ¼ 1=2 ) behavior can dramatically affect the resulting apparent stiffness of the joint (Lai et al. 1992). Specialized precision experimental methods to measure such Poisson’s ratios or bulk moduli (Rightmire 1970; Yu et al. 2001; Dillard et al. 2008; Tizard et al. 2012; Motamed et al. 2014) may be needed for some modeling purposes of highly constrained elastomers. Figure 5 illustrates two types of the shear tests, nominally used to measure the apparent shear strength(ASTM-D1002-99 1999; Packham 2005). In the single-lap joint, an adhesive is sandwiched between two plates and shearing is applied by tensle loading until failure occurs. Though this test has advantages in that the method is quite simple and typical of many joint configurations, complexities associated with load eccentricity, bending moments, and terminal regions result in non-uniform normal and shear stress states along the length of the bond (Goland and Reissner 1944; ASTM-D4896-95 2001), effects that are exaggerated by the larger bondline thickness of elastomeric material systems discussed in this section.

600

D. A. Dillard and T. Yamaguchi Pull

Fig. 6 Geometry of a specimen of tensile adhesion test for sealants, elastomers, or foam adhesives

Adherend spacer

sample

spacer

Adherend

In order to reduce the rotation and peeling resulting from the eccentricity, the double lap joint has sometimes been advocated. The double-lap joint essentially consists of two single-lap joints, and although peeling is reduced by the symmetric configuration, it is not eliminated and can continue to dominate failure (Borgmeier and DeVries 2000). Figure 6 is a schematic of the tensile adhesion test, where the adhesive is filled into the space between two adherends and two spacers. In the tensile adhesion test, after the spacers are removed from the specimen, the adhesive may be extended by external loading. When a given elongation (i.e., 25%, 100%, etc.) is reached, the specimen may be fixed for a period of time (ex. 24 hours) with spacers, and any resulting failure type (cohesive or adhesive) is recorded. Cyclic straining, including in the presence of environment, is a common means to load such specimens (ASTMC719-14 2014) for durability assessment. Finally, yet another common characterization method for many adhesives, including sealants, are a variety of peel tests. 180 peel tests are particularly common for sealant testing because of the very simple fixturing requirement, though 45 , 90 , and other angles are also used. Reinforcing wire or fabric is often embedded in the sealant to provide enough integrity to prevent peel arm rupture, and also to minimize peel arm straining, which can significantly affect the measured peel force and resulting debonding energy calculations. Stretching of the peel arm requires additional input work, which is typically accomplished through larger crosshead movement, albeit with smaller required forces (Lindley 1971). Although commonly tested at fixed displacement rates (ASTM-C794 2010), these tests can be loaded at a range of rates, temperatures, and humidities, for example, to assess long term performance through viscoelastic superposition methods. Dead loads can also be applied to obtain long-term failure rate information. Elastomeric foam tapes produced by several manufacturers are finding applications in a wide range of non-structural and even semi-structural applications, including supporting road signs, facades, ceilings, and some load-bearing supports, as well as in the assembly of some transportation vehicles and trailers. Combining a relatively thick, highly dissipative and flexible layer with adequate adhesion offers opportunities for significant energy dissipation, potentially resulting in damagetolerant solutions where design stress levels are relatively small. Some of these products can be used as alternatives to more traditional sealants, including for mounting windows in commercial buildings. Representative studies of such materials can be found in Townsend et al. (2011a, b, 2012).

22

Special Tests

22.4

601

Indentation Tests

Indentation tests can be used to determine the hardness and the elastic modulus of a material (Tabor 2000), and in some cases, the interfacial fracture resistance of thin coatings (Ritter et al. 1989, 1992). Several tests are proposed using different types of indenters such as spherical, three-sided pyramidal, or four-sided Vickers, and the tests are performed not only on a macroscopic scale but also on a microscopic one; measurement of the hardness at very small scales is called nanoindentation. Advantages of the indentation test include that only a small sample is needed, minimal specimen preparation is required, and in some cases even portable equipment can be used to measure the mechanical properties of the material. Figure 7 shows a schematic for an indentation test of a flat specimen with a spherical indenter. The depth of penetration of a rigid spherical indenter beneath the specimen is ht. at full load Pmax. When the load is removed, assuming no reverse plasticity, the unloading is elastic and at complete unloading, there is a residual impression of depth hr. If the load Pmax is reapplied, then the reloading is elastic through a distance he = ht.  hr. according to the Hertz equation: 4 P ¼ E R1=2 h3=2 e 3

(9)

where E* is the composite elastic modulus 1 1  ν2 1  ν0 2 þ  ¼ E E E0

a P

b

Elasto-plastic loading

Pmax

(10)

Elastic unloading

Elastic loading

Rr

dP/dh

M

hr h

ht hr

Ri amax

he hp

he/2

Fig. 7 Load-Displacement curve for an elastic-plastic specimen loaded with a spherical indenter (radius Ri) showing both loading and unloading response. Upon loading, there is an initial elastic response followed by elasto-plastic deformation. Upon unloading, the specimen shows elastic response and comes back to the residual displacement hr. with the residual impression of radius Rr

602

D. A. Dillard and T. Yamaguchi

where E, E0, ν and ν0 correspond to Young’s moduli and Poisson’s ratio for the substrate and indentor, repectively and 1 1 1 ¼  R Ri R r

(11)

with Ri and Rr being the indenter radius and the radius of residual impression, respectively. Using these equations with the known indenter parameters (E*, ν0, and Ri) and measured radius of residual impression Rr, the elastic modulus of the material E can be determined. The hardness H can also be quantified by H¼

Pmax πa2max

(12)

where amax is the contact radius at maximum load. Microindentation tests are performed to measure the hardness and the elastic modulus of small specimens with low applied loads. One of the most difficult problems in the microindentation test is to determine the contact radius (FischerCripps 2000). In this case, the contact radius is estimated by the following equation. amax ¼

qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffi 2Ri hp  h2p ffi 2Ri hp

(13)

where hp = ht  he/2 is the the plastic depth, as defined in Fig. 7. Oliver and Pharr (1992) proposed a method to determine the plastic depth as h p ¼ ht 

3 Pmax : 4 dP=dh

(14)

the derivative dP/dh is calculated by the initial slope of the load-displacement curve at the unloading phase (see Fig. 7). The use of indentation tests to measure coating adhesion has received wide attention, including for very thin coatings (Oliver and Pharr 2010). Delamination can result from a combination of shearing and tensile stresses developing across the interface. Indentation can induce high stresses across the interface in the vicinity of the interface, but also can induce a compressive in-plane residual stress within the coating, which can contribute to the debonding process through buckling-driven delamination (DeBoer and Gerberich 1996a, b; deBoer et al. 1997a). Overlayers or superlayers can also be added to provide additional stored energy to drive the delamination at the desired interface (Bagchi et al. 1994). A related test, the notched coating adhesion test (Chang et al. 1997; Dillard et al. 1999), involves indentation with a razor blade to sever the coating and induce local debonding, which can then be propagated by subsequent mechanical straining of the substrate. This test has also been used to assess structural adhesive durability as the diffusion times are much shorter for open-faced coatings rather than bonded adhesive joints (Chang et al. 1997;

22

Special Tests

603

Moidu et al. 1998a). Other experimental approaches are able to use residual stress alone to propagate debonding, both for tensile (Farris and Bauer 1988) and compressive coating stress states (Hutchinson et al. 2000). Comprehensive theoretical treatments of stresses and fracture of layered and coated systems can be found in Hutchinson and Suo (1992) and Freund and Suresh (2010).

22.5

Scratch Tests

The scratch test is a method to determine the adhesion strength of a coated layer, and many embodiments of this method have been proposed, including a number with commercially available instruments designed for this purpose. In the scratch test, as schematically shown in Fig. 8, an indenter or stylus is translated across a coated substrate, typically at a constant velocity under an applied normal force of increasing magnitude. Stresses are introduced at the interface between coating and substrate, and at a critical load, delamination of the coating occurs. Benjamin and Weaver (1963) performed a simple analysis for the scratch test by using the following equations. AH F ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi R2  H 2

(15)

rffiffiffiffiffiffiffi W A¼ πH

(16)

and

where A is the radius of the contact circle, R is the radius of the stylus tip, W is the applied normal load, H is the indentation hardness of the substrate, and F is the lateral force. Though they tried to evaluate the coating adhesion by the normal load W and the lateral force F at which the the coating is removed from the substrate, there existed several difficulties such as, delamination of the coating before removal, thinning of the film instead of removal, complex elasto-plastic deformation of both coating and substrate, and so on. As with many other coating adhesion tests, real complications make the analysis difficult unless simplifying assumptions are appropriate. In order to improve the scratch test, more sophisticated apparati have been developed. The normal Fig. 8 Schematic of the scratch test, in which a coated substrate is moved to the left against a fixed stylus

Normal load

motion

scratch channel

coating substrate

604

D. A. Dillard and T. Yamaguchi

load is not only controlled to determine the critical load Wc required to initiate delamination, also the optical microscopy for the monitoring the scratch process and acoustic spectroscopy for detecting acoustic emmisions have been employed. Microscratch and nanoscratch tests have been discussed widely, including in deBoer et al. (1997b). Sarin (1995) developed an alternate version of the scratch test, turning the experiment on its side and performing a micro-scratch test on the cross section of the sample, as shown in Fig. 9. After the edge of the cross section is precisely polished, the scratch test is carried out on a microscopic scale. As a result, the failure mode can be isolated and determined at the coating/substrate interface. In fact, for a titanium carbide (TiC) film coated onto a silicon nitride/titanium carbide (Si3N4 + TiC) ceramic composite, he could generate a crack along the coating/ substrate interface and thus obtained a delamination failure mode. While the scratch test is advantageous for its simplicity in the geometry and ease of the sample preparation, there are several disadvantages. The test is often limited to hard, brittle coatings or nonlinear elasto-plastic deformation complications arise, where the failure mode of the scratch test is not well understood (Bull 2001). Furthermore, the test is very sensitive to the material and the surface geometry or topography of the stylus (Oroshnik and Croll 1978). However, the scratch test is one of the most popular adhesion tests in both industry and academic research. This comes from the fact that this technique is applicable to the evaluation of a wide range of systems, and commercial equipment is available to perform a variety of functions, such as surface roughness measurement, the evaluation of mechanical properties of the coating, and so on. In particular, the scratch test appears useful for many situations involving hard and brittle coatings. For softer coating, though this test is difficult to extract a true delamination mode from, it is still useful as a complementary test of other techniques. For further information, refer to Lacombe (2006) and references therein.

22.6

Tack

Tack is surprisingly complex, with quantitative values often involving a combination of conforming, wetting, bonding, and removal phenomena. Based on the concept of quantifying the adhesion resulting after application of light and momentary pressure, Fig. 9 Scratch test using microindentor. The indentor is driven across the substrate and the adhesive on the cross section

Coating Scratch track Substrate

22

Special Tests

605

tack is intended to measure the aggressiveness of a polymer surface. Tack is widely measured on pressure sensitive adhesives (PSA), but is also of interest in probing the degree of cure or consolidation as other adhesives move from their liquid application state to their solid use state, and for other applications involving adhesion development at interfaces. An underlying concern for tack measurements is that although the intention is to measure the ability of the adhesive to quickly wet and hold a substrate, tack tests inherently involve debonding. Thus, the bonding process is confounded by the debonding process that also occurs upon separation. This involves the creation of new surfaces as cavitation and fibrillation occur, as well as significant rheological dissipation as materials stretch, draw, and break. Tack is measured by a number of different techniques, ranging from an individual’s perception of the “thumb appeal” of an adhesive to sophisticated characterization of the loads and deformations and failure mechanisms involved in bonding and debonding process. Primary tack measurement techniques include probe tack tests, loop tack, quick stick, and rolling ball techniques. These techniques are illustrated schematically in Fig. 10. The probe tack test involves bringing a probe of desired shape into contact with the adhesive with a prescribed force and duration, then withdrawing the probe at a fixed displacement rate. Some units involve a mechanically driven punch that rises to lift a weight to which a PSA tape has been affixed. The weight can be selected to provide the desired initial contact force and pressure. The probe then drops at a fixed rate, as a load cell records the force. More versatile test frames have also been used to apply controlled force application and probe removal to a wide range of specimens. Although peak values of removal force have typically been recorded and used in quantifying tack, additional information can be obtained from the load deformation trace, such as illustrated in Fig. 11. Other metrics that may be of interest include the area under the force removal curve, the amount of deformation at which the peak force occurs, or the deformation required to completely sever the bond. A range of probe configurations have been used, but flat cylindrical punches of several

Fig. 10 Illustrations of several test methods used to characterize tack of pressure sensitive adhesives: (a) flat probe tack, (b) loop tack, and (c) rolling ball tack

a

c

b

606 Fig. 11 Illustration of force vs time trace for a probe tack test of a pressure sensitive adhesive as the probe is compressed to make contact, held for a fixed dwell time, then removed

D. A. Dillard and T. Yamaguchi

Peak Force

Force

Plateau (fibrillation)

Time Dwell

millimeters in diameter are common, with rounded tips and other configurations used for certain applications. Zosel (1998) has made wide use of these probe tack techniques characterize the tack of pressure sensitive adhesives as a function of substrate roughness, contact time, contact pressure, and other factors. His work has clearly shown that the drop in load from its initial peak is often due to the onset of internal cavitation within the adhesive layer. The onset of cavitation limits the peak force, but leads to fibrillation which dissipates significant amounts of energy, increasing the energy required to sever the bond. Creton and coworkers have developed sophisticated techniques for carefully aligning the contact, photographing the resulting failure modes, and relating this behavior to material properties (Lakrout et al. 1999). They have also shown that the rate-dependent peak force is related to the frequency-dependent modulus of the adhesive, with lower modulus adhesives exhibiting the onset of cavitation at lower applied loads (Gent and Wang 1991). The rolling ball tack test involves rolling a ball down an incline and characterizing tack by the ability of the adhesive to arrest the forward motion of the rolling ball. In essence, the ball is placed a given height above a horizontal table supporting the adhesive layer; the potential energy of the ball is converted to kinetic energy as the ball rolls down an incline to the plane containing the adhesive layer, and this energy is then dissipated, primarily through viscoelastic processes, as the ball is brought to a stop. Assuming an ideal conversion of potential energy (the mass of the ball times its height above the plane) to kinetic energy, this energy is equivalent to the distance the ball travels on the adhesive layer times some average dissipative force. One caveat of the rolling ball test is that the rates of bonding and debonding can be significantly larger than encountered in most other tack tests, a fact that can significantly affect the perceived tack of rate-dependent viscoelastic materials. Other tack tests common for PSAs include the loop tack and quick stick test. The loop tack test involves bringing a loop of adhesive tape, sticky side out, into contact with a rigid adherend of specified length. After contact with what are typically very small contact pressures when the backing is flexible, the crosshead is reversed and the loop is removed from the substrate. The initial portion of the test is similar to a

22

Special Tests

607

peel test at just over 90 , albeit one that is conducted without the bonding step normally used to prepare peel test specimens. As the debonds approach one another, the stiffness of the backing becomes more significant, and the loop tack begins looking and acting more like a probe test. Here, the backing stiffness again becomes important in that it controls the size of the contact that remains at final separation; thicker and stiffer backings lead to greater contact areas at final separation, resulting in larger peak force measurements. Some users report using the peak force when recording loop tack results, while others report using the plateau that results during the initial peeling process. These are clearly very different phenomena that are controlled by different mechanisms, again suggesting the complexity of performing and interpreting tack tests in general and loop tack tests in particular (Plaut and Ritchie 2004). The PSTC procedure known as the quick stick test is related to initial phase of the loop tack test. A tape is laid on a horizontal surface, without applying external mechanical pressure, and the strip is removed by pulling on one end. Again, this test resembles a 90 peel test.

22.7

CTE, SFT, and Residual Stresses

Residual stresses are of significant importance in many adhesive applications. In some applications, including for structural adhesive joints, the residual stresses induced within a bond can be larger than stresses induced by mechanical loading the joint experiences within its service life. Because residual stresses can represent locked-in stresses that are present at all times, their duration can become a factor in joint life. Also, because they can be ever present, residual stresses can make adhesive layers and coatings more susceptible to environmental damage, including environmental stress cracking (Dillard et al. 1995). The coefficient of thermal expansion can be measured through several methods, the most direct being through linear or volumetric dilatometry. Thermal mechanical analyzers (TMA) are ideally suited for such measurements, as their components are made of very low expansion materials such as quartz. With precise temperature control, they are able to accurately measure linear dimensional changes as a function of temperature. They are often equipped to measure other properties such as softening temperature, etc. in addition to the coefficient of thermal expansion as a function of temperature. While the coefficient of thermal expansion is nominally a property of the bulk adhesive, the residual stress, often the term of interest for design, is a property of a bonded joint or structure. One of the simplest means to measure the residual stress in an adhesive layer or coating is with the use of the bimaterial curvature technique, For several centuries, people have known that when two dissimilar beams are bonded together at the interface and subjected to a temperature change, a curvature will result. This curvature is a function of the relative moduli and thicknesses of the two beams or plates, the difference or mismatch in their CTEs, and the temperature change to which they are exposed. The resulting stresses are easily determined but

608

D. A. Dillard and T. Yamaguchi

are quite nonuniform (Timoshenko 1925). A very significant simplification occurs when the thickness of one layer is significantly smaller than the other. For example, if a softer polymeric adhesive is applied as a thin coating on a stiffer substrate, the resulting curvature becomes a function of the residual stress within this coating layer. For appropriate geometries then, characterizing the curvature of bimaterial strips provides a direct measure of the residual stress experienced within the adhesive or coating layer. Dating back to the work of Stoney (1909): σ¼

Es t2s 6ð1  νs Þhρ

(17)

where σ is the equal biaxial stress in the thin film, h is the thickness of the film, Es, ts, and νs are the modulus, thickness, and Poisson’s ratio of the thick substrate, and ρ is the radius of curvature. Several improvements have been proposed to refine the equation to cover a broader range of specimen configurations (Freund et al. 1999; Yu et al. 2003). This can be applied to thermal, hygral (Li et al. 2012), or other swelling mechanisms as well. Such tests are easily performed in commercial dynamic mechanical analyzers (DMA), which boost excellent force and displacement resolution in a carefully controlled temperature chambers (Case et al. 2005). Figure 12 illustrates applications of this technique in which a bimaterial strip is cantilevered at one end and the deflection of the other end is measured. Knowing the geometry of the specimen and the moduli of substrate and adhesive layer, one can determine the difference in coefficient of thermal expansion between the substrate and adhesive as well as the stress-free temperature.

a

b First heat of quenched samples show no undershoot

c Displacement

First heat of aged samples show undershoot

Residual Stress aE

Temperature T Ts Temperature-T

Fig. 12 Illustration of bimaterial curvature technique to measure coefficient of thermal expansion and stress-free temperature of adhesive layer showing: (a) test sample, (b) representative deflection results, and (c) residual stress predictions

22

Special Tests

22.8

609

Conclusions

This chapter describes a few examples of the special tests to measure the mechanical properties of the membranes, films, adhesives, coating, sealants, and foam/elastomeric adhesives. The major conclusions are the following: 1. A wide range of blister tests have been developed and utilized to characterize adhesion, and in some cases, moduli and other properties of adhering membranes and adherends. 2. For mechanical tests of thick samples sandwiched between two substrates such as sealants, elastomeric or foam adhesives, shear tests and tensile adhesion tests have been commonly applied. 3. Indentation tests are performed to evaluate the elastic modulus and the hardness of the coating, and in some cases, interfacial fracture energies associated with coating adhesion. 4. In the scratch tests, a stylus is driven across a coated surface under a normal load. The critical load is determined by the first appearance of the failure. While the test can be applied for hard and brittle coating, quantitative evaluation for the soft and ductile coating is more complex. 5. Several tack tests, such as probe tack, loop tack, quick stick and rolling ball tack tests are used to evaluate the aggressiveness of a polymer surface. 6. The coefficient of thermal expansion (CTE) and the stress-free temperature (SFT) can be important in engineering design and are often measured to estimate residual stresses in adhesives.

References Adams RD, Coppendale J, Peppiatt NA (1978) Stress analysis of axisymmetric butt joints loaded in torsion and tension. J Strain Anal 13(1):1–10 Adams RD, Comyn J, Wake WC (1997) Structural adhesive joints in engineering. Chapman and Hall, London Allen MG, Senturia SD (1989) Application of the island blister test for thin-film adhesion measurement. J Adhes 29(1–4):219–231 Anderson GP, Bennett SJ, DeVries KL (1977) Analysis and testing of adhesive bonds. Academic Press, New York Anderson GP, Chandapeta S, DeVries KL (1988). Effect on removing eccentricity from button tensile tests. Adhesively bonded joints: testing, analysis, and design. S. Johnson, ASTM. STP-981 ASTM-C719-14 (2014) Standard test method for adhesion and cohesion of elastomeric joint sealants under cyclic movement (Hockman cycle). ASTM, West Conshohocken ASTM-C794 (2010). Standard test method for adhesion-in-peel of elastomeric joint sealants. ASTM, West Conshohocken, PA ASTM-D1002-99 (1999). Standard test method for apparent shear strength of single-lap-joint adhesively bonded metal specimens by tension loading (metal-to-metal). Annual book of ASTM standards. ASTM, West Conshohocken, 15.06: 42–45 ASTM-D4896-95 (2001) Standard guide for use of adhesive-bonded single lap-joint specimen test results, Annual book of ASTM standards, vol 15.06. ASTM, West Conshohocken, pp 419–424

610

D. A. Dillard and T. Yamaguchi

Bagchi A, Lucas GE, Suo Z, Evans AG (1994) New procedure for measuring the decohesion energy for thin ductile film on substrates. J Mater Res 9(7):1734–1741 Benjamin P, Weaver C (1963) Adhesion of metals to crystal faces. R Soc Proc Ser A 274 (1357):267–273 Borgmeier PR and KL DeVries (2000). Interpreting adhesive joint tests. 2nd international symposium on adhesive joints, Newark Bull SJ (2001) Can the scratch adhesion test ever be quantitative? Adhesion measurement of films and coatings. K. L. Mittal. Utrecht, vol 2. VSP, p 107 Case SL, O'Brien EP, Ward TC (2005) Cure profiles, crosslink density, residual stresses, and adhesion in a model epoxy. Polymer 46(24):10831–10840 Chang YS, Lai YH, Dillard DA (1989) The constrained blister – a nearly constant strain-energy release rate test for adhesives. J Adhes 27(4):197–211 Chang T, Sproat EA, Lai YH, Shephard NE, Dillard DA (1997) A test method for accelerated humidity conditioning and estimation of adhesive bond durability. J Adhes 60(1–4):153–162 Dannenberg H (1961) Measurement of adhesion by a blister method. J Appl Polym Sci 5:125–134 DeBoer MP, Gerberich WW (1996a) Microwedge indentation of the thin film fine line. 1. Mechanics. Acta Mater 44(8):3169–3175 DeBoer MP, Gerberich WW (1996b) Microwedge indentation of the thin film fine line. 2. Experiment. Acta Mater 44(8):3177–3187 deBoer MP, Kriese M, Gerberich WW (1997a) Investigation of a new fracture mechanics specimen for thin film adhesion measurement. J Mater Res 12(10):2673–2685 deBoer MP, Nelson JC, Gerberich WW (1997b) Mechanics of interfacial crack propagation in microscratching. Thin Films: Stresses Mech Properties Vi 436:103–108 Dillard DA (2002) Fundamentals of stress transfer in bonded systems. In: Dillard DA, Pocius AV (eds) Adhesion science and engineering I: the mechanics of adhesion, vol 1. Elsevier, Amsterdam, pp 1–44 Dillard DA, Bao Y (1991) The peninsula blister test – a high and constant strain-energy release rate fracture specimen for adhesives. J Adhes 33(4):253–271 Dillard D, Parvatareddy H, Clifton AP (1995). Environmental stress cracking In high performance adhesives and composites. Antec 95 – the plastics challenger: a revolution in education, conference proceedings, Vols I-Iii–Vol I: Processing; Vol Ii: Materials; Vol Iii: Special areas, pp 3971–3975 Dillard DA, Chen B, Chang TN, Lai YH (1999) Analysis of the notched coating adhesion test. J Adhes 69(1–2):99–120 Dillard DA, Mallick A, Ohanehi DC, Yu J-H, Lefebvre DR (2008) A high precision experimental method to determine Poisson's ratios of encapsulant gels. J Electron Packag Trans ASME 130 (Compendex):0310061–0310067 Farris RJ, Bauer CL (1988) A self-delamination method of measuring the surface-energy of adhesion of coatings. J Adhes 26(4):293–300 Fernando M, Kinloch AJ (1990) Use of the inverted-blister test to study the adhesion of photopolymers. Int J Adhes Adhes 10(2):69–76 Fernando M, Kinloch AJ, Vallerschamp RE, Vanderlinde WB (1993) The use of the inverted-blister test to measure the adhesion of an Electrocoated paint layer adhering to a steel substrate (Vol 12, Pg 875, 1993). J Mater Sci Lett 12(16):U1243–U1243 Fischer-Cripps AC (2000) A review of analysis methods for sub-micron indentation testing. Vacuum 58(4):569–585 Freund LB, Suresh S (2010) Thin film materials: stress, defect formation and surface evolution. Cambridge, Cambridge University Press Freund LB, Floro JA, Chason E (1999) Extensions of the Stoney formula for substrate curvature to configurations with thin substrates or large deformations. Appl Phys Lett 74(14):1987–1989 Gent AN, Kaang S (1986) Pull-off forces for adhesive tapes. J Appl Polym Sci 32:4689–4700 Gent AN, Lewandowski LH (1987) Blow-off pressures for adhering layers. J Appl Polym Sci 33(5):1567–1577 Gent AN, Wang C (1991) Fracture-mechanics and cavitation in rubber-like solids. J Mater Sci 26(12):3392–3395

22

Special Tests

611

Goland M, Reissner E (1944) The stresses in cemented joints. J Appl Mech 11:A17–A27 Hencky H (1915) Über den spannungszustand in kreisrunden platten mit verschwindender biegungssteiflgkeit. Zeitschrift fur Mathematik und Physik 63:311–317 Hutchinson JW, Suo Z (1992) Mixed-mode cracking in layered materials. Adv Appl Mech 29(29):63–191 Hutchinson JW, He MY, Evans AG (2000) The influence of imperfections on the nucleation and propagation of buckling driven delaminations. J Mech Phys Solids 48(4):709–734 Jiang KR, Penn LS (1990) Use of the Blister test to study the adhesion of brittle materials.1. Test modification and validation. J Adhes 32(4):203–216 Kinloch AJ (1987) Adhesion and adhesives: science and technology. Chapman and Hall, London Kinloch AJ, Williams JG (2002) The mechanics of peel tests. In: Dillard DA, Pocius AV (eds) The mechanics of adhesion, vol 1. Elsevier, Amsterdam, pp 273–302 Kinloch AJ, Lau CC, Williams JG (1994) The peeling of flexible laminates. Int J Fract 66(1):45–70 Kinloch AJ, Hadavinia H, Kawashita L, Moore DR, Williams JG (2006) A numerical analysis of the elastic-plastic peel test. Engineering Fracture Mechanics 73:2324–2335. (Copyright 2006, The Institution of Engineering and Technology) Klosowski J, Wolf AT (2015) Sealants in construction. Boca Raton:CRC Press Lacombe R (2006) Adhesion measurement methods: theory and practice. CRC Press, Boca Raton Lai YH, Dillard DA (1990a) An elementary plate-theory prediction for strain-energy release rate of the constrained blister test. J Adhes 31(2–4):177–189 Lai YH, Dillard DA (1990b) Numerical-analysis of the constrained blister test. J Adhes 33(1–2):63–74 Lai YH, Dillard DA (1994) A study of the fracture efficiency parameter of blister tests for films and coatings. J Adhes Sci Technol 8(6):663–678 Lai YH, Dillard DA (1996) A comparison of energy release rates in different membrane blister and peel tests. J Adhes 56(1–4):59–78 Lai YH, Dillard DA, Thornton JS (1992) The effect of compressibility on the stress distributions in thin elastomeric blocks and annular bushings. J Appl Mech-Trans Asme 59(4):902–908 Lakrout H, Sergot P, Creton C (1999) Direct observation of cavitation and fibrillation in a probe tack experiment on model acrylic pressure-sensitive-adhesives. J Adhes 69(3–4):307–359 Li YQ, Dillard DA, Lai YH, Case SW, Ellis MW, Budinski MK, Gittleman CS (2012) Experimental measurement of stress and strain in Nafion membrane during hydration cycles. J Electrochem Soc 159(2):B173–B184 Liechti KM, Liang YM (1992) The interfacial fracture characteristics of Bimaterial and sandwich blister specimens. Int J Fract 55(2):95–114 Liechti KM, Shirani A (1994) Large-scale yielding in blister specimens. Int J Fract 67(1):21–36 Lindley, P. B. (1971). Ozone attack at a rubber- metal bond. J Institution Rubber Industry 5(6): 243-248 Lindsey GH (1967) Triaxial Fracture Studies. J Appl Phys 38(12):4843–4852 Maugis D (1999) Contact, adhesion and rupture of elastic solids. Springer, Heidelberg Moidu AK, Sinclair AN, Spelt JK (1995) Analysis of the peel test – prediction of adherend plastic dissipation and extraction of fracture energy in metal-to-metal adhesive joints. J Test Eval 23(4):241–253 Moidu AK, Sinclair AN, Spelt JK (1998a) Adhesive joint durability assessed using open-faced peel specimens. J Adhes 65(1–4):239–257 Moidu AK, Sinclair AN, Spelt JK (1998b) On the determination of fracture energy using the peel test. J Test Eval 26(3):247–254 Motamed A, Bhasin A, Liechti KM (2014) Using the poker-chip test for determining the bulk modulus of asphalt binders. Mech Time-Depend Mater 18(1):197–215 Oliver WC, Pharr GM (1992) An improved technique for determining hardness and elastic modulus using load and displacement sensing indentation experiments. J Mater Res 7(6):1564–1583 Oliver WC, Pharr GM (2010) Nanoindentation in materials research: past, present, and future. MRS Bull 35(11):897–907 Oroshnik J, Croll WK (1978) Threshold adhesion failure. An approach to aluminum thin-film adhesion measurement using the stylus method. In: Adhesion measurement of thin films, thick films and bulk coatings., 2–4 Nov 1976. ASTM, Philadelphia

612

D. A. Dillard and T. Yamaguchi

Packham DE (ed) (2005) Handbook of adhesion. Wiley, Chichester Plaut RH, Ritchie JL (2004) Analytical solutions for peeling using beam-on-foundation model and cohesive zone. J Adhes 80(4):313–331 Pocius AV (1997) Adhesion and adhesives technology: an introduction. Hanser, Munich Rightmire, G. K. (1970). Experimental method for determining Poisson's ratio of elastomers. J of Lubrication Tech 92 Ser F(3): 381-388 Ritter JE, Lardner TJ, Rosenfeld L, Lin MR (1989) Measurement of adhesion of thin polymercoatings by indentation. J Appl Phys 66(8):3626–3634 Ritter JE, Sioui DR, Lardner TJ (1992) Indentation behavior of polymer-coatings on glass. Polym Eng Sci 32(18):1366–1371 Sarin VK (1995) Micro-scratch Test for Adhesion Evaluation of Thin Films. In: Mittal KL (ed) Adhesion Measurement of Films and Coatings. VSP, Utrecht, p 175 Shephard N (2002) Stresses and fracture of elastomeric bonds. In: Dillard DA, Pocius AV (eds) Adhesion science and engineering-I: the mechanics of adhesion, vol 1. Elsevier, Amsterdam Stoney GG (1909) The tension of metallic films deposited by electrolysis. Proc R Soc A82:172–175 Tabor D (2000) The hardness of metals. Oxford University Press, London Timoshenko SP (1925) Analysis of bi-metal thermostats. J Opt Soc Am 11:233–255 Tizard GA, Dillard DA, Norris AW, Shephard N (2012) Development of a high precision method to characterize Poisson's ratios of Encapsulant gels using a flat disk configuration. Exp Mech 52(9):1397–1405 Townsend BW, Ohanehi DC, Dillard DA, Austin SR, Salmon F, Gagnon DR (2011a) Characterizing acrylic foam pressure sensitive adhesive tapes for structural glazing applications-part I: DMA and ramp-to-fail results. Int J Adhes Adhes 31(7):639–649 Townsend BW, Ohanehi DC, Dillard DA, Austin SR, Salmon F, Gagnon DR (2011b) Characterizing acrylic foam pressure sensitive adhesive tapes for structural glazing applications-part II: creep rupture results. Int J Adhes Adhes 31(7):650–659 Townsend B, Ohanehi DC, Dillard DA, Austin SR, Salmon F, Gagnon DR (2012) Developing a simple damage model for the long-term durability of structural glazing adhesive subject to sustained wind loading. J Archit Eng 18(3):214–222 Wan KT (1999) Fracture mechanics of a V-peel adhesion test – transition from a bending plate to a stretching membrane. J Adhes 70(3–4):197–207 Wan KT, Lim SC (1998) The bending to stretching transition of a pressurized blister test. Int J Fract 92(4):L43–L47 Wan KT, Mai YW (1996) Fracture mechanics of a shaft-loaded blister of thin flexible membrane on rigid substrate. Int J Fract 74(2):181–197 Wan KT, Guo S, Dillard DA (2003) A theoretical and numerical study of a thin clamped circular film under an external load in the presence of a tensile residual stress. Thin Solid Films 425(1–2):150–162 Williams ML (1969) Continuum interpretation for fracture and adhesion. J Appl Polym Sci 13(1):29–40 Williams JG (1997) Energy release rates for the peeling of flexible membranes and the analysis of blister tests. Int J Fract 87(3):265–288 Xu DW, Liechti KM, de Lumley-Woodyear TH (2006) Closed form nonlinear analysis of the peninsula blister test. J Adhes 82(8):831–866 Yu JH, Dillard DA, Lefebvre DR (2001) Pressure and shear stress distributions of an elastomer constrained by a cylinder of finite length. Int J Solids Struct 38(38–39):6839–6849 Yu JH, Guo S, Dillard DA (2003) Bimaterial curvature measurements for the CTE of adhesives: optimization, modeling, and stability. J Adhes Sci Technol 17(2):149–164 Zosel A (1998) Effect of fibrillation on the tack of pressure sensitive adhesives. Int J Adhes Adhes 18(Compendex):265–271

Part V Joint Design

Constitutive Adhesive and Sealant Models

23

Erol Sancaktar

Contents 23.1 23.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classical Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.1 General Deformation Theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.2 Rubber Elasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.3 Viscoelasticity Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.2.4 Interrelationship Between the Tensile and Shear Properties . . . . . . . . . . . . . . . . . . 23.2.5 Time–Temperature Superposition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3 Advanced Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.1 Nonlinear Viscoelasticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.2 Empirical Description of Creep Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.3 Increases in Joint Strength Due to Rate-Dependent Viscoelastic Adhesive Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.4 Reductions in Stress Concentration Due to Viscoelastic Adhesive Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.5 Singularity Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.6 Bulk Adhesive as Composite Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.7 The Concept of the Interphase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.8 Damage Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.3.9 The Effects of Cure and Processing Conditions on the Mechanical Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.4 Prediction of Limit (Threshold) and Failure Stresses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

616 616 616 620 625 629 629 631 631 632 633 636 636 637 639 647 657 659 660 661

Abstract

A complete approach to modeling adhesives and sealants needs to include considerations for: deformation theories and viscoelasticity with linearity and E. Sancaktar (*) Department of Polymer Engineering, University of Akron, Akron, OH, USA e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_23

615

616

E. Sancaktar

nonlinearity considerations, rubber elasticity, singularity methods, bulk adhesive as composite material, damage models, the effects of cure and processing conditions on the mechanical behavior, and the concept of the “interphase.”

23.1

Introduction

Most adhesives are polymer-based materials and exhibit viscoelastic behavior. Some adhesives are elastomer materials and also exhibit full or partial rubberlike properties. The word “elastic” refers to the ability of a material to return to its original dimensions when unloaded, and the term “mer” refers to the polymeric molecular makeup in the word “elastomer.” In cases where brittle material behavior prevails, and especially, when inherent material flaws such as cracks, voids, and disbonds exist in such materials, the use of the methods of fracture mechanics are called for. For continuum behavior, however, the use of damage models is considered appropriate in order to be able to model the progression of distributed and non-catastrophic failures and/or irreversible changes in material’s microstructure, which are sometimes described as “elastic limit,” “yield,” “plastic flow,” “stress whitening,” and “strain hardening.” Many adhesive materials are composite materials due to the presence of secondary phases such as fillers and carriers. Consequently, accurate analysis and modeling of such composite adhesives require the use of the methods of composite materials. A complete approach to modeling adhesives and sealants, therefore, needs to include considerations for deformation theories, viscoelasticity, bulk adhesive as composite material, damage models, and the effects of cure and processing conditions on the mechanical behavior.

23.2

Classical Models

23.2.1 General Deformation Theories The deformation theory was first introduced by Hencky (1924), and can be described in the form: ϵij ¼


σ  kk δij þ ψSij : 9K

(1)

where Sij is the deviatoric stress tensor. Equation 1 reduces to the elastic stress–strain relations when ψ = 1/2 G, where G is the elastic shear modulus. If the scalar function ψ is defined as  ψ¼

 1 þΩ 2G

(2)

23

Constitutive Adhesive and Sealant Models

617

then Eq. 1 can be interpreted in the form ϵij ¼ ϵij E þ ϵij V þ ϵij P

(3)

where 

 Sij ϵij ¼ δij þ þ ΩSij 9K 2G

(4)

ϵij P ¼ ΩSij

(5)


σ  kk

and

with Ω being a scalar function of the invariants of the stress tensor, and the superscripts E, V, and P representing elastic, viscoelastic, and plastic behaviors, respectively. Consequently, the relation ϵ¼


σ  E

þ Λσ

(6)

is obtained for uniaxial tension on the basis of Eqs. 4 and 5 with Λ = 2Ω/3. Ramberg and Osgood (1943) used a special form of Eq. 6 with Λ = Kσn1 to result in: ϵ¼


σ  E

þ Kσ n

0

(7)

where K and n0 are material constants. They reported that Eq. 7 could be used successfully to describe uniaxial tension and compression behavior of various metal alloys. Equation 7 was later modified by McLellan (1969) to accommodate strain rate effects. McLellan interpreted the terms E, K, and n0 of Eq. 7 as material functions with the function E representing viscoelastic behavior and functions K and n0 representing work-hardening characteristics. The terms E, K, and n0 were all described as functions of the strain rate (dε/dt) so that rigidity, stress, and plastic flow, respectively, were all affected by variations in the strain rate. Renieri et al. (1976) used a bilinear form of rate-dependent Ramberg–Osgood equation to describe the stress–strain behavior of a thermosetting adhesive in the bulk tensile form. The bilinear behavior was obtained when log εp was plotted against log σ, where εp represents the second term on the right-hand side of Eq. 7. The model adhesive they used was an elastomer-modified epoxy adhesive with and without carrier cloth. They made several modifications on the form of the equation previously used by McLellan. First, the plastic strain εp was assumed to be a function of the overstress above the elastic limit stress (the development of over-stress approach will be presented subsequently) and second, the stress levels σ* defining the intersection point

618

E. Sancaktar 10000 9000 8000

Stress (psi)

7000 6000 5000 4000

q = 3,500 psi E = 330,000 psi

3000

5,950 psi ≤ s ∗ ≤ 7,480 psi

2000 Upper and lower stress whitening stress limits. 1000

Visually observed stress whitening stress.

0 0.00

0.02

0.04

0.06

0.08

0.10

Strain (in/in)

Fig. 1 A stress–strain diagram identifying upper and lower stress-whitening stress limits, σU and σLo*, respectively, as well as the visually observed stress-whitening stress. The modulus of elasticity, E, and the elastic limit stress, θ for the bulk adhesive tested are 3.3  105 and 3.5  103 psi (2276 and 24 MPa), respectively

for the bilinear behavior were found to occur slightly below the stress-whitening stress values. The equations they developed in this fashion are given as: ϵ¼ ϵ¼ ϵ¼


σ  E
σ  E

σ , E

0 σyt, prevails, then the hydrostatic pressure, p, also affects the yield process. A given stress tensor, σij, can be written as: σ ij ¼ p þ σ deviatoric , and the hydrostatic pressure, p, is defined as: p ¼ ð1=3Þ½σ 1 þ σ 2 þ σ 3 

(136)

Therefore, by using Eqs. 134–136, the von Mises yield criterion can be modified to the form: h i  2 σ yc  σ yt ½σ 1 þ σ 2 þ σ 3  þ ðσ 1  σ 2 Þ2 þ ðσ 2  σ 3 Þ2 þ ðσ 1  σ 3 Þ2 ¼ 2σ yc σ yt :

(137)

Conditions resulting in the creation of free volume, such as crazing (stress whitening) and cavitation relate to the magnitude of the hydrostatic stress. Level of the critical (or crazing) strain, εCR at such occurrences can be predicted using the relation: ϵCR ¼ C þ ðD=pÞ

(138)

where C and D are temperature and rate-dependent material constants, and p is the hydrostatic stress.

23.5

Conclusion

Constitutive modeling of adhesives and sealants include considerations for deformation theories and viscoelasticity with linearity and nonlinearity considerations, rubber elasticity, singularity methods, bulk adhesive as composite material, damage models, and the effects of cure and processing conditions on the mechanical behavior. However, the concept of the “interphase” must always be kept in mind for accurate predictions in mechanical behavior of surfaces and joints bonded with adhesives and sealants. Interfaces may constitute a weak link in the chain of load transfer in bonded joints. Also, the discontinuity of the material properties causes abrupt changes in stress distribution, as well as causing stress singularities at the edges of the interfaces. In most engineering joints the adherend surfaces have distinct topographies, which result in a collection of miniature joints in micron, and even nanoscale when bonded adhesively. If the interphase is not considered as a

23

Constitutive Adhesive and Sealant Models

661

discrete collection of individual chemical bonds, the methods of continuum mechanics can still be applied to this collection of miniature joints by assuming continuous or a combination of continuous/discontinuous interphase zones. Thus, the displacement at the interphase need not be continuous at every location. The analysis of a miniature joint contributing to the overall adhesion in a macro joint can be performed in a fashion similar to that for the macro joint itself, which is usually studied with the employment of the methods of linear or nonlinear elasticity, viscoelasticity, plasticity, fracture, damage, and/or failure mechanics.

References Agarwal BD, Broutman LJ (1990) Analysis and performance of fiber composites, 2nd edn. Wiley, New York, p 131 Agrawal R, Drzal LT (1989) J Adhes 29:63 Ahlonis JJ, MacKnight WJ, Shen M (1972) Introduction to polymer viscoelasticity. WileyInterscience, New York Ashton JE, Halpin JC, Petit PH (1969) Primer on composite materials. Analysis Technomic, Stamford, p 77 Barsoum RS (1989) J Adhes 29:149 Brinson HF (1974) In: Kausch H et al (eds) Deformation and fracture of high polymers. New York, Plenum Brinson HF, Renieri MP, Herakovich CT (1975) Fracture mechanics of composites. ASTM SPT 593:177 Budiansky B (1965) J Mech Phys Solids 13:223 Castro JM, Macosko CW, Perry SJ (1984) Polym Commun 25:82 Chase KW, Goldsmith W (1974) Exp Mech 14:20 Chow TS (1978a) J Polym Sci Polym Phys Ed 16:959 Chow TS (1978b) J Polym Sci Polym Phys Ed 16:967 Chow TS, Hermans JJ (1969) J Compos Mater 3:382 Darlington MW, Turner S (1978) Creep of engineering materials. Mechanical Engineering, London Davis JL (1964) J Polym Sci A 2:1311 Dempsey JP (1995) J Adhes Sci Technol 9:253 Dillard DA, Hiel C (1985) Proceedings of the 1985 SEM spring conference on experimental mechanics, Las Vegas, p 142 Eshelby JD (1959) Proc R Soc Lond A 252:561 Ferry JD (1961) Viscoelastic properties of polymers. Wiley, New York Findley WN, Peterson DB (1958) ASTM Proc 58:841 Flügge W (1975) Viscoelasticity, 2nd edn. Springer, Berlin Ganghoffer JF, Schultz J (1996) J Adhes Sci Technol 10:775 Garton A, Haldankar GS, Mclean PD (1989) J Adhes 29:13 Gittus J (1975) Creep, viscoelasticity and creep fracture in solids. Wiley, New York Gomatam RR (2002) Modeling fatigue behavior of electronically conductive adhesives. PhD dissertation, The University of Akron, Akron Gomatam RR, Sancaktar E (2004) J Adhes Sci Technol 18:1833 Gomatam RR, Sancaktar E (2006a) J Adhes Sci Technol 20:69 Gomatam RR, Sancaktar E (2006b) J Adhes Sci Technol 20:87 Good RJ, Gupta RK (1988) J Adhes 26:13 Green AE, Adkins JE (1960) Large elastic deformations and non-linear continuum mechanics. Oxford University Press, New York Groth HL (1988) Int J Adhes Adhes 8:107

662

E. Sancaktar

Hashin Z, Shtrikman S (1963) J Mech Phys Solids 11:127 Hata T, Gams M, Kojimer K (1965) Kobunshikagahn (Chem High Polym, Jpn) 22:160 Hata T, Ohsaka K, Yamada T, Nakamue K, Shibata N, Matsumoto T (1994) J Adhes 45:125 Hencky HZ (1924) Z Angew Math Mech 4:323 Hermans JJ (1967) Proc R Acad Amst B70:1 Hiel C, Cardon AH, Brinson HF (1983) The nonlinear viscoelastic response of resin matrix composite laminates. Virginia Polytechnic Institute and State University report no. VPI-E-83-6 Hill R (1964) J Mech Phys Solids 12:199 Hill R (1965) J Mech Phys Solids 13:213 Hilton HH (1975) In: Baer E (ed) Engineering design for plastics, Robert E. Krieger, Huntington, New York Jozavi H, Sancaktar E (1988) J Adhes 25:185 Jozavi H, Sancaktar E (1989a) J Adhes 27:143 Jozavi H, Sancaktar E (1989b) J Adhes 27:159 Jozavi H, Sancaktar E (1989c) J Adhes 29:233 Kaelble DH (1964) J Colloid Sci 19:413 Kaelble DH (1969) J Adhes l:102 Kelly A, Tyson WR (1965) Mech Phys Solids 13:329 Kerner EH (1956) Proc Phys Soc B69:808 Keuner VH, Knauss WG, Chai H (1982) Exp Mech 22:75 Knollman GC, Hartog JJ (1985) J Adhes 17:251 Koizumi S, Fuhue N, Matsunaga T (1970) Nihon Setchahu Kyohaishi. J Adhes Soc Jpn 6:437 Lipatov YS, Babich VF, Todosijchuk TT (1991) J Adhes 35:187 Liu DM (2000) J Mater Sci 35:5503 Ludwik PG (1909) Elemente der technologischen mechanik. Springer, Berlin, p 9 Ma W, Gomatam R, Fong R, Sancaktar E (2001) J Adhes Sci Technol 15:1533 McLellan DL (1969) Appl Polym Symp 12:137 Nakao K (1969) Preprint, symposium on mechanisms of adhesive failure, Tokyo, p 47 Nonaker Y (1968) Nihon Setchahu Kyokaishi (J Adhes Soc Jpn) 4:207 Peretz D, Weitsman Y (1982) J Rheol 26:245 Prandtl L (1925) Proceedings of the lst international congress on applied mechanics, Technisch Boekhandel en Dructerrg, Delft, p 43 Ramberg W, Osgood WR (1943) Description of stress-strain curves by three parameters. NACA TN 902, Washington, DC Ravi-Chandar K, Knauss WG (1984) Int J Fract 25:247 Reedy ED Jr, Guess TR (1995) J Adhes Sci Technol 9:237 Renieri MP, Herakovich CT, Brinson HF (1976) Rate and time dependent behavior of structural adhesives. Virginia Polytechnic Institute and State University, College of Engineering report no. VPI-E-76-7 Reuss E (1930) Z Angew Math Mech 10:266 Rochefert MA, Brinson HF (1983) Nonlinear viscoelastic characterization of structural adhesives. Virginia Polytechnic Institute and State University report no. VPI-E-83-26 Russel WB, Acrivos A (1973) Z Angew Math Phys 24:581 Sancaktar E (1981) Intl J Adhes Adhes 1:329 Sancaktar E (1985) Intl J Adhes Adhes 5:66 Sancaktar E (1991) J Adhes 34:211 Sancaktar E (1995) J Adhes Sci Technol 9:119 Sancaktar E (1996) Appl Mech Rev 49:S128 Sancaktar E, Beachtle D (1993) J Adhes 42:65 Sancaktar E, Brinson HF (1979) The viscoelastic shear behavior of a structural adhesive. Virginia Polytechnic Institute and State University, College of Engineering report no. VPI-E-79-14 Sancaktar E, Brinson HF (1980) In: Lee LH (ed) Adhesion and adsorption of polymers. Polymer science and technology series, vol 12-A. Plenum, New York, p 279

23

Constitutive Adhesive and Sealant Models

663

Sancaktar E, Dembosky SK (1986) J Adhes 19:287 Sancaktar E, Kumar S (2000) J Adhes Sci Technol 14:1265 Sancaktar E, Ma W (1992) J Adhes 38:131 Sancaktar E, Narayan K (1999) J Adhes 13:237 Sancaktar E, Padgilwar S (1982) J Mech Des 104:643 Sancaktar E, Schenck SC (1985) Ind Eng Chem Prod Res Dev 24:257 Sancaktar E, Wei Y (1998) In: van Ooij WJ, Anderson H (eds) Mittal Festschrift on adhesion science and technology. VSP, Utrecht, p 509 Sancaktar E, Zhang P (1990) Trans ASME J Mech Des 112:605 Sancaktar E, Jozavi H, Klein RM (1983) J Adhes 15:241 Sancaktar E, Schenck SC, Padgilwar S (1984) Ind Eng Chem Prod Res Dev 23:426 Sancaktar E, Jozavi H, El-Mahallawy AH, Cenci NA (1987) Polym Test 7:39 Sancaktar E, Turgut A, Guo F (1992) J Adhes 38:91 Sancaktar E, Babu SV, Zhang E, D’Couto GC, Lipshitz H (1995) J Adhes 50:103 Sawa T, Temma K, Nishigaya T, Ishikawa H (1995) J Adhes Sci Technol 9:215 Schapery RA (1966) Proceedings of the 5th U.S. national congress of applied mechanics, ASME, p 511 Schapery RA (1969a) Further development of a thermodynamic constitutive theory: stress formulation. Purdue University report no. 69-2 Schapery RA (1969b) Polym Eng Sci 9:295 Sharon G, Dodiuk H, Kenig S (1989) J Adhes 31:21 Sharpe LH (1972) J Adhes 4:51 Shaw SJ, Tod DA (1989) J Adhes 28:231 Sinclair JW (1992) J Adhes 38:219 Sokolovsky VV (1948) Prikl Mathematiha I Mehhanika 12:261 Struik DJ (1950) Classical differential geometry. Addison-Wesley, New York Tobolsky AV (1960) Properties and structure of polymers. Wiley, New York Turgut A, Sancaktar E (1992) J Adhes 38:111 Walpole LJ (1969) J Mech Phys Solids 17:235 Ward IM, Hadley DW (1993) Mechanical properties of solid polymers. Wiley, New York, p 206 Wei Y (1995) Electronically conductive adhesives: conduction mechanisms, mechanical behavior and durability. PhD dissertation, Clarkson University, Potsdam, New York Weitsman Y (1981) J Adhes 11:279 Wildemuth CR, Williams MC (1985) Rheol Acta 24(1):75 Wu TT (1966) Int J Solids Struct 2:1 Zabora RF, Clinton WW, Bell JE (1971) Adhesive property phenomena and test techniques. AFFDL-TR-71-68 Zhang MJ, Straight MR, Brinson HF (1985) Proceedings of the 1985 SEM spring conference on experimental mechanics, Las Vegas, p 205 Zhou J, Sancaktar E (2008) Chemorheology of epoxy/nickel conductive adhesives during processing and cure. J Adhes Sci Technol 22:957

Analytical Approach

24

Liyong Tong and Quantian Luo

Contents 24.1 24.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classical Analytical Analysis of Single-Lap Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.1 Adhesive Stress . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.2.2 Load Update and Bending Moment Factor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3 Other Analytical Solutions for Single-Lap Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.1 Adhesive Stress with the Effects of Transverse Shear and Geometric Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.2 Coupled Formulation for Load Update of Single-Lap Joints . . . . . . . . . . . . . . . . . 24.3.3 Numerical Comparisons . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.3.4 Analytical Approach for Adhesive Joints with Material Nonlinearity . . . . . . . . 24.4 Analytical Approach of Other Lap Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.1 Analytical Solutions of Symmetric Lap Joints Subjected to General Loadings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.4.2 Analytical Solutions of Asymmetric and Unbalanced Joints . . . . . . . . . . . . . . . . . . 24.5 Failure Criteria Based on Continuum Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.5.1 Shear and Peel Stresses of an Adhesive Joint with a Semi-infinitive Length . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.5.2 Failure Prediction Using the Criteria Based on Continuum Mechanics . . . . . . 24.6 Failure Criteria Based on Fracture Mechanics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.6.1 Failure Prediction Using Energy Release Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.6.2 Analytical Solutions of Energy Release Rates for Cohesive Failure and Debonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24.6.3 Failure Prediction Considering Plastic Deformation . . . . . . . . . . . . . . . . . . . . . . . . . . 24.7 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

666 668 668 673 674 674 677 679 682 685 685 686 691 691 692 692 693 694 697 698 698

L. Tong (*) · Q. Luo School of Aerospace, Mechanical and Mechatronic Engineering, The University of Sydney, Sydney, NSW, Australia e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_24

665

666

L. Tong and Q. Luo

Abstract

This chapter presents analytical approach for determining stress and strength of adhesively bonded joints. Selected applications of adhesively bonded joints are discussed first, and then mathematical models for stress analysis of these joints are outlined. Various closed-form solutions for adhesive stresses and edge bending moment for balanced single-lap joints are presented and compared. The method for finding analytical solutions for asymmetric and unbalanced adhesive joints is also discussed. Explicit expressions for mode I and mode II energy release rates for cohesive failure and interfacial debonding are presented for asymmetric joints with a semi-infinitive length subjected to general load combinations.

24.1

Introduction

Adhesive bonding technology is increasingly being used in practical structures due to its load-transferring performance. It can be used to effectively join similar and dissimilar thin sheets of materials via forming superb load-bearing structural joints (Adams et al. 1997; Tong and Steven 1999). It can behave superior to mechanical fastening in joining composite parts due to relatively smooth stress distribution compared to highly localized stress concentration in mechanically fastened joints. It is finding wider applications in many industries. For example, in aerospace and automotive industries, it is used to connect thin-walled structural parts; in civil engineering, it is used to reinforce concrete structures by bonding composite materials; and in intelligent structural systems, it is used to bond functional materials to host structures. Figure 1 depicts some of the typical joint configurations to join structural panel components. When adhesive bonding is used to repair or strengthen structural defects, such as cracks and ballistic damage, single-sided, double-sided, scarf and step repair configurations, as shown in Fig. 2, are usually used. Reinforced concrete (RC) beams with adhesively bonded fiber-reinforced polymer (FRP) as shown in Fig. 3a have also been widely used in civil engineering. In this case, the fiber-reinforced polymer plate plays the role of bridging the cracks in the concrete beams, thus slowing down or arresting the crack propagation. Figure 3b Fig. 1 Common joint configurations: (a) single-lap joint; (b) double-lap joint; (c) L-joint; (d) T-joint

a

b Adhesive

c

d

Adhesive

24

Analytical Approach

a

667

b

c

d

Adhesive

Adhesive

Fig. 2 Composite repair configurations: (a) single-strap joint; (b) double-strap joint; (c) scarf joint; (d) step joint

a

b

Adhesive

Cracks

Concrete Beam

PZT

PZT

PZT

PZT

PZT

PZT

Fiber reinforced polymer plate

Fig. 3 Adhesive bonding technology used in civil engineering and intelligent structure systems: (a) FRP-plated RC beam; (b) PZT smart plate

illustrates a piezoelectric (PZT) smart plate, in which PZT patches are adhesively bonded to the plate surface to control its structural performance. Figures 1, 2, and 3 illustrate that lap-type adhesive joints are common configurations used in practical structures. Although lap-type joints have the features of two-dimensional (2D) geometry, their stress analysis usually involves threedimensional (3D) stresses and strains; hence it is a 3D problem in nature. Numerical methods, such as finite element analysis (FEA) (see ▶ Chap. 25, “Numerical Approach: Finite Element Analysis”), have been used to solve 3D problems of adhesively bonded structures. However, simple and closed-form solutions are important in terms of identifying fundamental characteristics and key parameters. Analytical solutions can also be useful in preliminary design stage as they can offer in timely fashion, simple, quick, and meaningful answers. Therefore, considerable efforts have been devoted to the development of analytical approach for determining stress and strength of an adhesively bonded joint. Hence, this chapter focuses on presentation of basic mechanics models and analytical solutions for selected adhesively bonded joints.

668

L. Tong and Q. Luo

In order to derive closed-form explicit solutions for adhesively bonded joints/structures, it is important to develop appropriate mechanics models that take into account fundamental characteristics and key parameters. It is a compromise between the including of selected physical behavior and likelihood of admissible explicit solutions, and this balance plays a vital role in the development of analytical solution methods. A lap-type joint often consists of one or more adhesive layer(s) sandwiched between two or more beams/plates usually referred to as adherends. An adherend usually has a thickness smaller than the other in-plane dimension(s), and when the aspect ratio, that is, in-plane dimension-to-thickness ratio, is sufficiently large, it is reasonable to model the adherends as a beam or plate. As the theory for beam-like structures have been well established, adherends in most adhesively bonded structures can be simplified as beamor plate-like structures. An adhesive layer typically has a thickness of a fraction of a millimeter, for example, a bondline thickness of approximately 0.2 mm is common in secondarily bonded aerospace or automobile structures. The stress state in the adhesive layer is rather complex in 3D or 2D nature when the adherends are modeled as plates or beams, respectively. However, in practice, an adhesive is very thin compared to an adherend; therefore, the adhesive layer is generally modeled by an infinite number of springs or an interface by assuming that the only adhesive normal (thickness direction) and shear stresses exist, and they remain constant across the bondline thickness and only vary along the bondline length. This approximation leads to a mechanics model (referred as 1D model) that permits analytical solutions for stresses in an adhesively bonded joint. In this chapter, analytical models and solutions for adhesively bonded joints treated as beam-like structures are presented.

24.2

Classical Analytical Analysis of Single-Lap Joints

The simplest geometric configuration of adhesively bonded structures is probably the single-lap joint (SLJ), which has been widely used as a standard test specimen to calibrate adhesive properties and has been shown to be representative and challenging in identifying fundamental characteristics and key parameters in bonded lap joints.

24.2.1 Adhesive Stress Average Shear Stress Model This model treats the adherend as a rigid bar and the adhesive as continuous shear spring, which is based on the assumption that the adherends are so stiff that they do not deform and only the adhesive deforms as it is so soft, as shown in Fig. 4a. In this model, the adhesive shear stress uniformly distributes within the adhesive and is given by: τ ¼ F=ð2cÞ

(1)

24

Analytical Approach

669

a

b N1 + dN1

N1 F

t F

t

Undeformed N2

N2 + dN2

F F

dx

Deformed

Fig. 4 Stress analysis of the adhesive based on the one parameter elastic medium: (a) an average shear stress model; (b) free body diagrams of the shear lag model

where F is the applied tensile force per unit width and (2c) is the overlap length; the unit width of a beam has been assumed. In this chapter, the unit width and plane stress state are assumed for brevity. The results can be easily extended to plane strain state by changing the related parameters. Also, the composite adherends with the symmetric lay-ups are assumed. The results can be directly applied to joints with isotropic adherends.

Shear Lag Model Volkersen (1938) proposed a shear lag model, in which the adherend was modeled as a rod undergoing axial or longitudinal deformation only and the adhesive as a continuous shear spring. As shown in Fig. 4b, the equilibrium and constitutive equations for adherends 1 and 2 are: dN1 dN2 þ τ ¼ 0; τ ¼0 dx dx N i ¼ Adi

dui ði ¼ 1, 2Þ dx

(2) (3)

where Ad is the extensional stiffness and subscript i (i = 1, 2) denotes adherend 1 and 2. The adhesive shear stress is defined as: τ¼

Ga ð u2  u1 Þ ta

(4)

where Ga and ta are the shear modulus and thickness of the adhesive. The shear stress can be found by solving Eqs. 2, 3, and 4 and is given by:
  βv F coshβv x ð1  ψ Þsinhβv x  2 sinhβv c ð1 þ ψ Þcoshβv c sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð1 þ ψ ÞGa Ad1 and ψ ¼ where βv ¼ ta Ad1 Ad2 τ¼

(5)

670

L. Tong and Q. Luo

This solution clearly shows that the shear stress is not uniform along the bondline length direction and peaks at the adhesive ends. The maximum shear stress occurs at one end and is given by:
  βv F coshβv c ð1  ψ Þsinhβv c  when ψ  1 2 sinhβv c ð1 þ ψ Þcoshβv c

(6a)


  βv F coshβv c ð1  ψ Þsinhβv c þ ¼ when ψ  1 2 sinhβv c ð1 þ ψ Þcoshβv c

(6b)

τmax ¼ or τmax

For a relatively long overlap, the maximum shear stress can be approximately written as: ψβv F when Ad1 ¼ 6 Ad2 1þψ 1 ¼ βv F when Ad1 ¼ Ad2 2

τmax ¼ or : τmax

(6c)

In Eq. 6c, sinh βvc  cosh βvc is assumed, which is sufficiently accurate in most adhesive joints used in engineering.

Adhesive-Beam Model Goland and Reissner (1944) proposed an adhesive-beam model for the single-lap joint shown in Fig. 5a, in which a two-parameter elastic medium and Euler beams are used to model the adhesive and adherends, respectively. The free body diagrams of an infinitesimal element are shown in Fig. 6a, and the equilibrium equations of adherends 1 and 2 are: 8 dN > < 1 þ τ ¼ 0; dx > : dN2  τ ¼ 0; dx

dQ1 dM1 t1 þ σ ¼ 0; þ τ  Q1 ¼ 0 dx dx 2 dQ2 dM2 t2  σ ¼ 0; þ τ  Q2 ¼ 0 dx dx 2

(7)

where t1 and t2 are the thicknesses of adherends 1 and 2, t2=t1 is for the balanced joints, and τ and σ are the adhesive shear and peel stresses, defined by: τ¼


 Ga t1 dw1 t2 dw2 Ea þ ð u2  u1 Þ þ ; σ ¼ ðw2  w1 Þ ta 2 dx 2 dx ta

(8)

in which Ea is Young’s modulus of the adhesive. By using the Euler beam theory, Eqs. 7 and 8, the governing equations for adhesive stresses can be derived as:

24

Analytical Approach

671

c

a

c

l

Left outer adherend

F

O1

l

b

x3 z3

A

Overlap I z

O

F

Right outer adherend Tk

Mk

Mk Nk

x O2

Vk

Mk

Tk

c

B

II

Vk Qk Adherend 1 Adhesive Mk Nk

Qk

Adherend 2

d F

A

O1

I

B z

O

x

II

F O2

Fig. 5 A single-lap joint: (a) configuration; (b) deformation and forces in cross sections I and II of the deformed single-lap joint; (c) forces at the overlap edges; (d) SLJ with alignment tabs

d3 τ dτ d4 σ ¼ 0; 4 þ 4β4σ σ ¼ 0  β2c 3 dx dx dx

(9)

in which β2c

¼

αa β2τ ; βτ

pffiffiffi rffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffi 2 4 2Ea 8Ga 1 4Ad1 t21  ¼ ; βσ ¼ ; αa ¼ ð1 þ αk Þ; αk ¼ (10) 2 4 Ad1 ta D1 ta D1

where D1 is the bending stiffness of adherends 1 and 2 and αa and αk reflect the influences of composite lay-ups. For isotropic adherends, αa = 1 and αk = 3. Solving Eq. 9 with the boundary conditions shown in Fig. 5b, c yields: 8
> < Bσ1 ¼ sinh 2βσ c þ sin 2βσ c 2βσ ½Mk βσ ðsinh βσ c cos βσ c  cosh βσ c sin βσ cÞ þ V k cosh βσ c cos βσ c > > : Bσ4 ¼ sinh 2βσ c þ sin 2βσ c (12) As sinh βσc > > 1 and sinh βcc > > 1 in most cases, the maximum shear and peel stresses occur at the adhesive edges and can be simplified as: τmax


 Ft1 ðβc c þ αk Þ þ 2αk Mk ðβc c  1Þ Vk 2 ; σ max ¼ βσ Mk þ ¼ 8αa t1 c βσ

(13)

It is noted that the stress analysis of Goland and Reissner (1944) is for adhesively bonded isotropic joints, and an extension to composite joints is presented here.

Adhesive Models Based on Two-Dimensional Elastic Theory When an adhesive layer is relatively thick, the above adhesive model of assuming constant shear and peel strains may not be effective. In fact, shear and peel strains are not constant, and the axial strain may not be ignored even if the adhesive is very thin (e.g., < 2ua ¼ u2  u1 ; 2wa ¼ w2 þ w1 ; 2ϕa ¼ ϕ2 þ ϕ1 2N s ¼ N 2 þ N 1 ; 2Qs ¼ Q2  Q1 ; 2Ms ¼ M2  M1 ; 2V s ¼ V 2  V 1 > > : 2N a ¼ N 2  N 1 ; 2Qa ¼ Q2 þ Q1 ; 2Ma ¼ M2 þ M1 ; 2V a ¼ V 2 þ V 1 (18) where subscripts 1 and 2 indicate adherends 1 and 2; subscripts s and a denote the symmetric and anti-symmetric deformations; and the variables have the usual meanings used in Timoshenko beam theory. The equilibrium equations derived by considering the geometric nonlinearity of Fig. 6b are (Luo and Tong 2007, 2008):

24

Analytical Approach

8 dN > < 1 þ τ ¼ 0; dx > : dN2  τ ¼ 0; dx

675

dQ1 þ σ þ τϕ1 ¼ 0; dx dQ2  σ  τϕ2 ¼ 0; dx

dM1 t1 þ τ  Q1 ¼ N 1 ϕ1 dx 2 dM2 t1 þ τ  Q2 ¼ N 2 ϕ2 dx 2

(19)

When the Timoshenko beam is used to model the adherends, the adhesive shear and peel stresses can be expressed as: τ¼

2Ga  t1  2Ea ua þ ϕ a ; σ ¼ ws ta 2 ta

(20)

The governing equations taking into account the geometric nonlinearity of the overlap and the transverse shear stiffness of the adherends can be derived by using Eqs. 18, 19, and 20 and the constitutive equations of the Timoshenko beam, and they are given by (Luo and Tong 2008):
 8 3 d ua β2τ dua t1 d2 wa > > þ ¼0 < 3  4 dx 2 dx2 dx
 > d4 w α β2 dua t1 d2 wa β2 d2 wa > : 4a  k τ þ ¼0  k 2 2t1 dx 2 dx 2 dx2 dx

(21)

d 2 us d 4 ws d 2 ws ¼ 0; 4  4β2ng 2 þ 4β4nσ ws ¼ 0 2 dx dx dx

(22)

where β2ng


 β2 ¼ β2g þ k ; β4nσ ¼ 8

β4σ

þ

β2k β2g 2

! 1 ; βg ¼  2

rffiffiffiffiffiffiffiffiffiffiffi 2Ea Gk1 ta

(23)

in which Gk1 is the shear stiffness of adherends 1 and 2. When the boundary conditions are prescribed, analytical solutions of Eqs. 21 and 22 can be obtained. Luo and Tong (2008) derived the analytical solutions for given forces at overlap edges. For composite SLJs, the analytical solutions of Eq. 21 are:

ua ¼ Aa2 cosh βa1 x þ Aa4 cosh βa2 x þ Aa7 wa ¼ Ba1 sinh βa1 x þ Ba3 sinh βa2 x þ Ba6 x

(24)

where 2

β2a1 , β2a2

3 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 4 1 1 β ¼ 4β2c þ  β4c þ 1  β2 β 2 þ k 5 2 2αa c k 2 4 β2k

The analytical solutions of Eq. 22 for composite SLJs are:

(25)

676

L. Tong and Q. Luo

8 F > < us ¼ x 2Ad1      sinh β > s1 x sin β s2 x þ Bs4 cosh βs1 x cos β s2 x Δ < 0 : ws ¼ Bs1 þ þ þ þ ws ¼ Bs2 cosh βs1 x þ Bs4 cosh βs2 x Δ > 0

(26)

where 8 > Δ ¼ β4ng  β4nσ > > qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi > <  β βs1 ¼ β2nσ þ β2ng ; β2nσ  β2ng Δ < 0 s2 ¼ r rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi > qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi > pffiffiffi pffiffiffi > þ 2 4 4 þ > : βs1 ¼ 2 βng þ βng  βnσ ; βs2 ¼ 2 β2ng  β4ng  β4nσ Δ > 0

(27)

It can be shown that, for the case of thick adherend, thin adhesive and low transverse shear stiffness of the adherends, Δ > 0; otherwise, Δ < 0. In general, Δ < 0 for isotropic adherends made of steel and aluminum, and Δ > 0 for composite adherends due to the lower transverse shear stiffness. As ta ! 0, Δ > 0; it becomes a problem of interface fracture or delamination (Luo and Tong 2009b, c). Δ = 0 is a special case, whose solutions can be obtained by letting Δ ! 0 in the solutions for Δ < 0 or Δ > 0. By using Eqs. 20, 24, and 26, adhesive stresses can be directly obtained, and the details can be found in Luo and Tong (2008). In most cases, β2a1  β2c >> β2a2 , β2g >> β2k , and β2σ >> β2k . When these approximations are used, the analytical solutions can be simplified. The simplified adhesive stresses at the adhesive ends are: τmax ¼

½2αk Mk þ Ft1 ð1  r k Þβa1 c coth βa1 c αk ðFt1  2Mk Þβa2 c coth βa2 c þ (28) 8αa t1 c 8αa t1 c 8 " ! # > β > 2 s1 > > β Mk þ Vk Δ < σ β2σ " ! # (29) σ max ¼ þ þ > > β þ β > β2 M þ s1 s2 > Vk Δ>0 > k : σ 2β2σ

The adhesive stresses of a single-lap joint reach maximum at the edges, which are important in assessing the adhesive joint strength. The expressions of maximum adhesive stresses obtained using different analytical models are summarized in Table 1. β2 where r k ¼ k2 2βc It can be shown that when βk ! 0, the maximum adhesive stresses in nonlinear analysis degenerate to those in linear analysis. It is noted that different strategies can be used to obtain analytical solutions of adhesively bonded composite joints with and without considering geometrical nonlinearity. When the linear overlap is assumed, it is handy to express the governing equations in terms of the adhesive

24

Analytical Approach

677

Table 1 Comparisons of the maximum adhesive stresses Adherend model Average stress model Shear lag model Linear adhesiveEuler-beam model

Maximum peel stress   β2σ Mk þ Vβ k σ

h

  i β Δ 0

Nonlinear adhesiveTimoshenko-beam model

σ

Maximum shear stress F/(2c) βvF/2 ð2αk Mk þ Ft1 Þβa1 c αk ðFt1  2Mk Þ þ 8αa t1 c 8αa t1 c ½2αk Mk þ Ft1 ð1  r k Þβa1 c coth βa1 c 8αa t1 c αk ðFt1  2Mk Þβa2 c coth βa2 c þ 8αa t1 c

stresses. When geometrical nonlinearity is considered, it is convenient to write the governing equations in terms of the adherend displacements.

24.3.2 Coupled Formulation for Load Update of Single-Lap Joints One distinctive and challenging feature of a single-lap joint is its eccentric load path, which leads to the geometric nonlinearity that requires the load update. Goland and Reissner (1944) conducted uncoupled formulation for the load update and the adhesive stresses, respectively. In their formulation, the overlap was treated as a homogeneous beam for determining load update.

Coupled Formulation with Adhesive Shear Deformation Effect But No Overlap Nonlinearity Hart-Smith (1973) showed that the uncoupled formulation of Goland and Reissner (1944) is inconsistent with the force conditions of an individual adherend at the overlap ends. To overcome this deficiency and consider the influence of the adhesive deformation, Hart-Smith (1973) modeled each adherend of the overlap as an individual beam and derived the coupled formulation for load update and the adhesive shear stress. By using continuity conditions at intersection I, Hart-Smith (1973) obtained the edge moment factor and then simplified it as: k¼

1 2

1 þ βk c coth βk l þ ðβk cÞ =6

or k¼

1 1 þ βk c þ ðβk cÞ2 =6

when coth βk l  1 (30)

In this formulation (Hart-Smith 1973), the force conditions at intersections I and II are consistent; the influence of the adhesive shear strain is included, and the large deflections of the outer adherends are also modeled. However, the equilibrium equations in Eq. 7 and w1  (w1 + w2)/2 are used in this formulation. It is noted that Eq. 7 does not model the overlap geometric nonlinearity, and hence the large deflection effect of the overlap is ignored. Because w1  (w1 + w2)/2 is used, the effect of the adhesive peel strain on the edge moment factor is also neglected.

678

L. Tong and Q. Luo

Coupled Formulation with the Shear and Nonlinear Effects But No Peel Effect Oplinger (1994) treated the adherends as individual beams and considered the influence of the overlap deflection. The governing equations with the coupling of the adhesive shear strain and the overlap large deflection were derived, and then analytical solutions for SLJs were found. The edge moment factor was derived as:
 
  ta T h21 ta R3 1 þ þ R2 C2 þ 8R4 R C1 1 þ  C2 t T h22 t1 1  k¼
pffiffiffi

 T h21 ta T h21 1þ RC1 þ 8 1 þ R2 C1 R3 þ 8R4 t1 T h22 T h1

(31)

The parameter definitions in the above equation can be found in Oplinger (1994) and are not given here due to complexity. In Eq. 31, the overlap geometric nonlinearity and the adhesive shear, but not peel strain, are considered. In Eqs.30 and 31, the adherends are modeled as Euler beams, and thus the transverse shear stiffness is not modeled.

Fully Coupled Nonlinear Formulation Luo and Tong (2007, 2008) proposed fully coupled nonlinear formulation for isotropic SLJs based on the Euler beam theory and then extended to composite SLJs based on the Timoshenko beam theory. The displacements in Eqs. 24 and 26 are derived by taking into account geometric nonlinearity of the overlap and transverse shear stiffness of the adherends. The displacements of the left outer adherend of Fig. 5 can be derived as: u3 ¼

F Mk sinh βk x3 t1 þ ta þ x3 x3 þ uo1 ; w3 ¼  Ad1 F sinh βk l 2ð l þ c Þ

(32)

By using the continuity conditions at intersection I, unknowns Mk, uo1, and Ba6 can be determined. The adherend displacements, adhesive stresses, and the updating forces at the overlap ends can be obtained simultaneously. The simplified edge moment factor is given by: 

 βa1 cf ðβa2 cÞ  1 1 þ ðβk cÞ 8αa βa1 cð1 þ ta =t1 Þ   k¼ β cf ðβa2 cÞ þ αk 1 þ ðβk cÞ coth βk l þ ðβk cÞ2 αsM þ a1 8αa βa1 c 2

(33)

where f ðβa2 cÞ ¼

βa2 c coth βa2 c  1 ðβa2 cÞ2

; αsM ¼ αsM ¼

β s1 ðΔ < 0Þ or 2β2σ c

þ βþ s1 þ β s2 ð Δ > 0Þ 4β2σ c

(34)

24

Analytical Approach

679

For the isotropic adherends, αa = 1, αk = 3, and βa1  βc = βτ, Eq. 33 degenerates to that given in Luo and Tong (2007) for the isotropic SLJ. When the peel effect is neglected, Eq. 33 becomes: 

 βa1 cf ðβa2 cÞ  1 8αa βa1 cð1 þ ta =t1 Þ   k¼ β cf ðβa2 cÞ þ αk 1 þ ðβk cÞ coth βk l þ ðβk cÞ2 a1 8αa βa1 c 1 þ ðβk cÞ2

(34a)

In this case, the model is equivalent to that proposed by Oplinger (1994). In experiment, alignment tabs are often used to reduce peeling effects as shown in Fig. 5d (Adams et al. 1997; Tong and Steven 1999). In this case, the edge moment factor is (Luo and Tong 2016):   1 þ ðβ k cÞ 2 βa1 cf ðβa2 cÞ  1 þ ðβ k cÞ sinh βk l 8α β cð1 þ tc =tÞ  a a1  k¼ β cf ðβa2 cÞ þ αk 1 þ ðβk cÞcoth βk 1 þ ðβk cÞ2 αsM þ a1 8αa βa1 c

(34b)

When F is large, sinh(βkl) > > βkc; in this case, k calculated by Eq. 34b is almost the same as that by Eq. 33. Equation 33 includes all geometric parameters and material properties of composite single-lap joints except for the transverse normal stiffness. A good correlation with geometrically nonlinear FEA results has been illustrated in Luo and Tong (2007, 2008). In some cases, the following equations may also be used to predict the edge moment factor (Luo and Tong 2007): k¼

1 1 or k ¼ when coth βk l  1 1 þ βk c coth βk l 1 þ βk c

(35)

Equation 35 is simple and applicable for SLJs with a short overlap. Zhao et al. (2010) derived the second formula of Eq. 35 by considering overlap rotation and ignoring the overlap deformation. They showed that it was suitable for SLJs with a stiff overlap.

24.3.3 Numerical Comparisons Two examples of single-lap joints are used to illustrate the above formulations. One is an isotropic SLJ and the other is a composite SLJ. The data of the isotropic SLJ are E1 = 70 (GPa), v1 = 0.34, t1 = 1.6 (mm); Ea/E1 = 0.008, va = 0.4; and c/t1 = 32 and l/c = 5. For the composite SLJ, t1 = 1.6 (mm), ta = 0.2 mm, c/t1 = 32, l/c = 1.25; and Ea = 2.4 (GPa), va = 0.4. The ply thickness of composite adherends is 0.2 mm, and the ply material properties are EL = 138 (GPa), ET = 9.4 (GPa), vLT = 0.32,

680

L. Tong and Q. Luo

GLT = 6.7 (GPa); andv23 = 0.32, G23 = 3.56 (GPa); and the lay-up is [90/0/90/0]s. The numerical results are presented for the SLJs in plane strain.

Edge Moment Factor The edge moment factors (k) of the isotropic SLJ predicted by various analytical solutions are illustrated in Fig. 7. It is noted that Δ < 0 when the Timoshenko beam is used to model isotropic adherends of this SLJ. It has been shown that edge moment factor predicted by Eq. 33 correlates very well with that predicted by the geometrically nonlinear FE analyses (Luo and Tong 2007, 2008). Figure 7 indicates that the edge moment factor predicted by the formulation of Goland and Reissner (1944) is significantly larger than that predicted by Eq. 33 (Luo and Tong 2007, 2008) with an increase of the applied force due to ignoring influence of the adhesive deformation; the edge moment factor calculated by Oplinger’s formulation (1994) is obviously larger than that of Eq. 33 owing to neglecting the peel effect. The edge moment factor calculated using Eqs. 35 or 30 (Hart-Smith 1973) is considerably lower than that of Eq. 33. Equation 35 was given in Luo and Tong (2007), and the second formula in Eq. 35 was also derived by Zhao et al. (2010). This equation can be directly extended to composite SLJs. Figure 8 shows the edge moment factors of the composite SLJ calculated by using the theoretical formulations. It is noted Δ > 0 for this composite SLJ. It has also been illustrated that k calculated by Eq. 33 correlates well with that of the geometrically nonlinear FE analyses (Luo and Tong 2008) for the composite SLJs. Oplinger’s formulation (1994) is not easy to be extended to the laminated adherends, and thus its results are not given in this figure. Since k for Eq. 34a is equivalent to that of Oplinger’s model, the peel effect can be observed by comparing 0.9 0.8

Eq. 24.17

Edge moment factor k

0.7

Eq. 24.31 Eq. 24.33

0.6

Eq. 24.35 0.5

Eq. 24.30

0.4 0.3 0.2 0.1 0

0

1

2

3

4

5

6

7

b kc

Fig. 7 Edge moment factor of the isotropic SLJ predicted by the beam-adhesive models

8

24

Analytical Approach

681

0.6 Eq. 24.17 Eq. 24.33a

0.5 Edge moment factor k

Eq. 24.33 Eq. 24.35

0.4

Eq. 24.30 0.3 0.2 0.1 0

0

1

2

3

4

5

6

7

8

9

bk c

Fig. 8 Edge moment factor of the composite SLJ predicted by the beam-adhesive models

k for Eqs. 33 and 34a. It can be seen that the influence of adhesive peel strain on the edge moment factor is significant and should not be ignored.

Adhesive Stresses When the updating edge forces are determined, the relevant adhesive stresses can be calculated, and they are plotted in Figs. 9 and 10 for the composite SLJ. Figures 9 and 10 illustrate distributions of the shear and peel stresses for the composite SLJ predicted by the linear adhesive-beam model based on the Euler beam theory and the nonlinear adhesive-beam model based on the Timoshenko beam theory, respectively. In Figs. 9 and 10, the same edge moment factor at βkc = 9 calculated Eq. 33 is used for both models. Figure 9 shows that there are slight differences in values of shear stress predicted by the linear adhesive-beam model based on the Euler beam theory and the nonlinear adhesive-beam model based on the Timoshenko beam theory. However, there exist significant differences in the adhesive peel stress for the two models as shown in Fig. 10, including peak values and distribution patterns. Harris and Adams (1984) investigated single-lap composite joints using a geometrically nonlinear finite element analysis (NFEA). They observed a small discrepancy in shear stress but large differences in the adhesive peel stress and the edge moment by comparing numerical results of Goland and Reissner (1944) with those of the NFEA. Figures 8, 9, and 10 confirm their observations through analytical solutions, as Luo and Tong (2008) illustrated that the edge moment factor, shear, and peel stresses predicted by the nonlinear adhesive-beam model based on the Timoshenko beam theory correlate well with those of the NFEA using commercial software NASTRAN.

682

L. Tong and Q. Luo 140

Linear Euler beam model Nonlinear Timoshenko beam model

Shear stress (MPa)

120 100 80 60 40 20 0 –1

–0.95

–0.9

–0.85

–0.8

–0.75

–0.7

Overlap axis x (= x/c)

Fig. 9 Shear stress predicted by the linear adhesive-beam model based on the Euler beam theory and the nonlinear adhesive-beam model based on the Timoshenko beam theory 140

Linear Euler beam model 120

Nonlinear Timoshenko beam model

Peel stress (MPa)

100 80 60 40 20 0 –20 –40 –1

–0.95

–0.9

–0.85

–0.8

–0.75

–0.7

Overlap axis x (= x/c)

Fig. 10 Peel stress predicted by the linear adhesive-beam model based on the Euler beam theory and the nonlinear adhesive-beam model based on the Timoshenko beam theory

24.3.4 Analytical Approach for Adhesive Joints with Material Nonlinearity In design and strength analysis of adhesively bonded structures, nonlinear material properties need to be considered as soft or ductile adhesive materials can behave in

24

Analytical Approach

683

plastic regime. Owing to stress concentration and singular features, stresses near an edge of an adhesively bonded joint can reach beyond elastic limitation. Hence, material nonlinear analysis for adhesive joints is important. Analysis of material nonlinearity is usually required to be considered for all structural materials. However, due to low linear limitation of the ductile adhesive, stress–strain relationship of an adherend may still be within linear range when the adhesive is loaded beyond its linear limitation. Hence, only the adhesive material nonlinearity is considered here. For isotropic adherends that yield, Adams et al. (1997) have proposed a single design methodology (see ▶ Chap. 27, “Design Rules and Methods to Improve Joint Strength”). Hart-Smith (1973) conducted material nonlinear analysis for SLJs using the linearly elastic and perfectly plastic model to describe the adhesive shear stress–strain relationship and the linear material property for the adhesive peel stress and adherends. In his analysis (Hart-Smith 1973), the adhesive layer is divided into elastic and plastic regions as shown in Fig. 11. Grant and Teig (1976) considered material nonlinearity of adhesive by dividing the adhesive into multiregions, in which the shear lag model was used. The obtained governing equations were then solved numerically. Delale et al. (1981) used a strain energy function to define stress–strain relationships in conducting material nonlinear analysis for lap joints. Material nonlinear analysis of unbalanced lap joints subjected to general loadings was conducted by Bigwood and Crocombe (1990). In their analysis, the adherends were modeled as linearly elastic plates deforming in cylindrical bending, and plastic deformation theory was applied to the adhesive material. The stress–strain curve was described by using a continuous mathematical function. The adhesive yielding rule was assumed to obey the von Mises yield criterion. Bigwood and Crocombe (1990) derived a set of six simultaneous nonlinear differential equations in conjunction with the prescribed stress–strain function and then solved these nonlinear differential equations using finite difference method. The effect of the adhesive thickness on lap joint strength was also studied in Bigwood and Crocombe (1990). Yang et al. (2004) developed an analytical model to predict the adhesive stress and strain distributions. In their analysis, the adherends were modeled as a laminated anisotropic plate, and the adhesive constitutive equation was described using the elastic–perfectly plastic constitutive relationship. The von Mises yield

as

cs

cs

as Plastic zone x

F

F

Plastic zone z

Fig. 11 Elastoplastic analysis of shear lag joints

684

L. Tong and Q. Luo

criterion was also applied to the adhesive plastic deformation. They derived and numerically solved coupled differential equations of SLJs. In this analysis (Yang et al. 2004), the material nonlinear FEA was conducted for numerical comparison, and cohesive failure was studied. Lee and Kim (2007) presented a closedform analytical model for material nonlinear analysis of symmetric SLJs, in which the elastoplastic perfect material property for adhesive was considered. The analysis was conducted for independent shear and peel stresses in the adhesive. As discussed in the above brief review, material nonlinear analysis of adhesively bonded structures is very complicated, and analytical solutions are not admissible in general. Here, we only present analytical solutions of symmetric SLJs in tension with the elasto-perfectly plastic adhesive material using the shear lag model to explain the analytical approach for adhesively bonded joints with material nonlinearity. As shown in Fig. 11, the governing equations for elastic and plastic parts can be written as (Hart-Smith 1973; Grant and Teig 1976; Lee and Kim 2007): 8 2 d τ > > < 2 ¼ β2v τ  cs  x  cs dx : d2 γ 2τs > > : 2¼ cs  jxj  c Ad1 ta dx

(36)

where c = cs + as and γ is the adhesive shear strain and τs is the adhesive yield stress. The analytical solution of the shear stress in |x|  cs can be readily obtained: τ¼

βv ðF  2as τs Þcosh βv x 2sinh βv cs

(37)

When the maximum value of Eq. 37 is equal to τs, the plastic deformation will occur in the adhesive edge. The critical force F is equal to (2τs/βv). With an increase of applied load F, the plastic deformation will develop along the adhesive. The following equation can be obtained from Eq. 37 to determine the plastic zone length as: 2τs sinh βv ðc  as Þ ¼ βv ðF  2as τs Þcosh βv ðc  as Þ

(38)

When the plastic zone is small, sinh βv(c  as)  cosh βv(c  as). In this case, the plastic zone length as is approximately given by: as ¼

βv F  2τs 2βv τs

(39)

When as is determined, the stress distribution can be obtained: when |x|  (c  as), it is given by Eq. 37 and it is equal to τs in (c  as) < |x|  c.

24

Analytical Approach

24.4

685

Analytical Approach of Other Lap Joints

24.4.1 Analytical Solutions of Symmetric Lap Joints Subjected to General Loadings In Sects. 2 and 3, analytical solutions for single-lap joints are derived from the governing equations of Eqs. 9, 20, and 21 for the linear and nonlinear overlaps. The governing equations are also applicable to the other symmetric lap joints such as single-strap joints and cracked lap shear specimens.

Linear Overlap Based on the Euler Beam Theory The general solutions of Eq. 9 are (Goland and Reissner 1944):

τ ¼ A1 sinh βc x þ A2 cosh βc x þ A3 σ ¼ ðB1 sinh βσ x þ B2 cosh βσ xÞ sin βσ x þ ðB3 sinh βσ x þ B4 cosh βσ xÞ cos βσ x (40)

where Ai and Bi (i = 1, 2, 3; j = 1, 2, 3, 4) are integration constants determined by relevant boundary conditions. When the forces at both overlap edges are prescribed, the boundary conditions can be derived as: 8 c ð > > > > < τðxÞdx ¼ N 1I  N 1II ;

dτð cÞ β2τ ¼ ½t1 ðN 2  N 1 Þ  2αk ðM2 þ M1 Þx¼ c 8t1 dx

c > > > d2 σ ð cÞ d 3 σ ð cÞ > : ¼ 2β4σ ðM2  M1 Þjx¼ c ; ¼ 2β4σ ðV 2  V 1 Þjx¼ c 2 dx dx3

(41) By substituting Eqs. 40 into 41, the integration constants can be determined. Delale et al. (1981) presented analytical solutions of lap joints under tension, bending, and transverse shear, respectively.

Nonlinear Overlap Based on the Timoshenko Beam Theory The general solutions of Eqs. 20 and 21 are (Luo and Tong 2008):

ua ¼ Aa1 sinh βa1 x þ Aa2 cosh βa1 x þ Aa3 sinh βa2 x þ Aa4 cosh βa2 x þ Aa5 x2 þ Aa6 x þ Aa7 wa ¼ Ba1 sinh βa1 x þ Ba2 cosh βa1 x þ Ba3 sinh βa2 x þ Ba4 cosh βa2 x þ Ba5 x2 þ Ba6 x þ Ba7

(42) When Δ < 0, the analytical solutions are given by:

us ¼ A s1 x þ As2 

          ws ¼ B s1 sinh βs1 x þ Bs2 cosh β s1 x sin β s2 x þ Bs3 sinh β s1 x þ Bs4 cosh β s1 x cos βs2 x

(43a)

686

L. Tong and Q. Luo

When Δ > 0, the solutions are:

us ¼ As1 x þ As2 þ þ þ þ þ þ þ ws ¼ Bþ s1 sinh βs1 x þ Bs2 cosh βs1 x þ Bs3 sinh β s2 x þ Bs4 cosh βs2 x

(43b)

þ where the integration constants Aai and Bai (i = 1, 2,

, 7), Asi, and B sj or Bsj (i = 1, 2; j = 1, 2, 3, 4) can be determined by associated boundary conditions. When the forces at both overlap edges are given, the solutions can be found in Sect. 2 of Luo and Tong (2008). For the symmetric single-strap and lap shear joints, the analytical solutions can be obtained directly from the formulations of Luo and Tong (2008). For double-lap and double-strap joints, analytical analyses are attributed to those for asymmetric joints by using the variable transformation discussed in Luo and Tong (2002) for the host beam with the piezoelectric patches bonded on the top and bottom surfaces. The analysis for asymmetric joints will be presented in the next section.

24.4.2 Analytical Solutions of Asymmetric and Unbalanced Joints Analytical analysis of balanced single-lap joints gives us insight into the essences of stress analysis for adhesively bonded joints. It has been widely used for adhesive characterization and benchmarking numerical methods and experimental testing. In general, adherends are asymmetric and adhesive joints are unbalanced. Analytical approaches for unbalanced adhesive joints based on the Euler beam theory have been developed by a number of researchers (Hart-Smith 1973; Delale et al. 1981; Bigwood and Crocombe 1989; Luo and Tong 2002; Shahin et al. 2008). Luo and Tong (2009b, c) modeled the adherends as Timoshenko beams. Since the analytical procedure for the Euler beam is similar to that for the Timoshenko beam, the formulations are presented here for the Timoshenko beam only. Fully coupled nonlinear analytical solutions for asymmetric and unbalanced adhesive joints are very complicated. Nevertheless, when the updated loadings are given, linear analysis based on the Timoshenko beam theory can predict sufficiently accurate adhesive stresses except for the case of very large deflection. When the linear adhesive-beam model based on the Timoshenko beam theory is used, the distributions of adhesive shear and peel stresses are almost the same as those of the nonlinear model, and the values of the shear stress is slightly lower than that of the nonlinear analysis. Therefore, the linear adhesive-beam model based on the Timoshenko beam for stress analysis of asymmetric and unbalanced joints may be sufficiently accurate for engineering applications.

Governing Equations of Asymmetric and Unbalanced Linear Joints In this model of adhesively bonded joints with asymmetric and unbalanced adherends, the geometric nonlinearity is not included; the adherends are described

24

Analytical Approach

687

as the Timoshenko beam and a two-parameter elastic medium is used to model the adhesive. Also, the adhesive stresses should be deemed as those in the adhesive’s midplane. It should be noted that the predicted shear stress may be slightly lower due to ignoring the geometric nonlinearity. The equilibrium equations and adhesive stresses are shown in Eqs. 7 and 8, respectively. The constitutive equations for laminated adherends with asymmetric lay-ups are:
 dui dϕ dwi  Bi i ; Qi ¼ Gki  ϕi ; dx dx dx dui dϕi  Di ði ¼ 1, 2Þ M i ¼ Bi dx dx

N i ¼ Adi

(44)

where Bi (i = 1, 2) are the extension-bending coupling stiffness of the adherends. By substituting Eqs. 7 and 44 into Eq. 8, the governing equations in terms of shear and peel stresses can be derived as (Luo and Tong 2009b, c): 8 3 d τ dτ > > < 3  k1  k2 σ ¼ 0 dx dx > d4 σ d2 σ dτ > : 4  k5 2 þ k4 σ þ k3 ¼ 0 dx dx dx

(45)

where k1 ¼ kτ k1d ; k2 ¼ kτ k2d ; k3 ¼ kσ k2d ; k4 ¼ kσ k4d ; k5 ¼ kσ k5d ; kτ ¼ Ga =ta ; kσ ¼ Ea =ta

 8 D1 t 1 B1 D2 t2 B2 Ad1 Ad2 > > 4αa1 þ 4αa2  þ þ ; k4d ¼ < k1d ¼ Δ1
D1  Δ2
D2  Δ1 Δ2 t1 2B1 t2 2B2 1 1 > > : k2d ¼ Ad1  Ad2 þ þ  ; k5d ¼ Gk1 Gk2 2Δ1 t1 2Δ2 t2 Δi ¼ Adi Di  B2i ; αai ¼

1 þ αki Adi t2i ; αki ¼ 4 4Di

ði ¼ 1, 2Þ

(46)

(47)

(48)

Equation 45 is a set of differential equations with coupled shear and peel stresses. In the above equations, k1d and k4d reflect the extension, bending, and extensionbending coupling effects of the adherends; k2d indicates the unbalanced condition of the substrates that couples the shear and peel stresses; and k5d denote the shear stiffness. The shear and peel stresses become decoupled when k2d = 0, which is seen in the case of balanced joints with identical substrates. If k5d = 0, Eq. 45 becomes the governing equations based on the Euler beam theory. When the shear and peel stresses are solved analytically from Eq. 45, forces and displacements can be derived from Eqs. 7 and 44, respectively.

688

L. Tong and Q. Luo

Boundary Conditions The governing equations shown in Eq. 45 can be analytically solved without introducing any other assumption for the prescribed boundary conditions. When the forces at the adhesive joint ends I and II are given, the boundary conditions can be expressed as: 8 > > < x ¼ c : > > :x ¼ c :

dτ ¼ kτ H nI ; dx dτ ¼ kτ H nII ; dx

d2 σ  k5 σ ¼ kσ H mI ; dx2 2 d σ  k5 σ ¼ kσ H mII ; dx2

d3 σ dσ  k5 þ k3 τ ¼ kσ H qI 3 dx dx d3 σ dσ  k5 þ k3 τ ¼ kσ H qII dx dx3 (49)

where 

  

  8 D2 B2 t2 D1 B1 t1 Ad2 t2 B2 Ad1 t1 B1 > > H ¼ þ   þ þ  N N  M Mj1 nj j2 j1 j2 > > Δ1
2Δ1 2Δ 2Δ1 Δ1 > > Δ
2 2Δ2   2 Δ2 < Ad2 Ad1 B2 B1 Mj2  Mj1  N j2  N j1 H mj ¼  > Δ Δ Δ Δ 2 1  2 1 >
> > A A > > : H qj ¼  d2 Qj2  d1 Qj1 ðj ¼ I,II Þ Δ2 Δ1

(50) where Njk, Mjk, and Qjk ( j = I, II; k = 1, 2) are the force components in adherends 1 and 2 at the joint ends I and II.

Analytical Approaches Analytical solutions of the governing equations with coupled shear and peel stresses were discussed in detail in Luo and Tong (2002) for piezoelectric smart beams, in which the Euler beam theory was used. Stress analysis for adhesively bonded composite joints is to solve Eq. 45 with boundary conditions in Eq. 49. One method to solve this boundary value problem of the ordinary differential equations is elimination; it attempts to obtain the uncoupled equations with higherorder differentiations by eliminating the coupled terms (Humi and Miller 1988). In this method, the boundary conditions with the higher-order differentiations should also be derived. This technique has been used to solve the governing equations with the coupled shear and peel stresses for adhesive joints based on the Euler beam theory by Hart-Smith (1973), Delale et al. (1981), Bigwood and Crocombe (1989), and Shahin et al. (2008). In these formulations, the Euler beam is used to model the adherends. In the adhesive joint model of Delale et al. (1981), the adhesive axial deformation is also considered. Alternatively, the coupled equations can be directly solved (Humi and Miller 1988). The solutions are then substituted into the governing equations to find the relationships of the integration constants, and the boundary conditions can be directly used. To solve Eqs. 45 and 49, the general solutions of Eq. 45 can be obtained analytically, and then the relationships of the integration constants are

24

Analytical Approach

689

derived by substituting the general solutions into the first or second equation of Eq. 45. The integration constants can be determined by these relationships and the boundary conditions of Eq. 49. This approach is simpler than the elimination method in general. Luo and Tong (2009b, c) used this method for a semi-infinitive long bonded composite joint. Here, analytical solutions for the adhesive joint with a length of (2c) are introduced as follows.

General Solutions pffiffiffiffi pffiffiffiffi Substituting τ ¼ Ae kσ βt x and σ ¼ Be kσ βt x into Eq. 45 yields (Luo and Tong 2009b, c): λ3 þ a1 λ2 þ a2 λ þ a3 ¼ 0 or βt ¼ 0

(51)

where
 kτ k1d k4d kτ k1d k5d ¼ þ k5d ; a2 ¼ þ ; a3 λ¼ kσ kσ kσ

 kτ k1d k4d  k22d ¼ k2σ β2t ; a1

(52)

There are two different sets of solutions for Eq. 51, depending on: Δ ¼ Q3  R2 where Q ¼

a21  3a2 2a3  9a1 a2 þ 27a3 and R ¼ 1 9 54

(53)

where Δ is a function of ta; it may be larger than, equal to, or less than zero. It has been shown that Δ  0 as ta ! 0 (Luo and Tong 2009b). That is, when the adhesive layer is sufficiently thin, Δ  0; and Δ < 0 when the adhesive is relatively thick. Relying on the characteristic roots of the associated polynomial equation shown in Eq. 51, analytical solutions of Eq. 45 are given by (Luo and Tong 2009b, c): 8 τ ¼ A1 sinh β1 x þ A2 cosh β1 x þ ðA3 sinh β2 x þ A4 cosh β2 xÞ sin β3 x > < þðA5 sinh β2 x þ A6 cosh β2 xÞ cos β3 x þ A7 σ ¼ B1 sinh β1 x þ B2 cosh β1 x þ ðB3 sinh β2 x þ B4 cosh β2 xÞ sin β3 x > : þðB5 sinh β2 x þ B6 cosh β2 xÞ cos β3 x þ B7

ðΔ < 0Þ

(54) or 8 τ ¼ A1 sinh β1 x þ A2 cosh β1 x þ A3 sinh β2 x þ A4 cosh β2 x > < þA5 sinh β3 x þ A6 cosh β3 x þ A7 σ ¼ B1 sinh β1 x þ B2 cosh β1 x þ B3 sinh β2 x þ B4 cosh β2 x > : þB5 sinh β3 x þ B6 cosh β3 x þ B7

ðΔ  0Þ (55)

where calculation of βi (i = 1, 2, 3) for Eqs. 54 and 55 can be found in Luo and Tong (2009a, b); Ai and Bi (i = 1, 2, . . ., 7) are the integration constants.

690

L. Tong and Q. Luo

Determination of the Integration Constants There are 14 integration constants in Eqs. 54 or 55. By eliminating shear stress in Eq. 45, a 6 order differential equation with respect to peel stress can be obtained, and thus B7 = 0. The axial equilibrium equation of adherend 1 of the overlap is shown in Eq. 41. By substituting Eqs. 54 or 55 into the first or second equation in Eq. 45, six simple relationships of the integration constants can be derived. There are six independent equations in Eq. 49. Therefore, 13 independent equations can be used to determine the 13 integration constants without introducing additional assumptions. Directly solving the 13 simultaneous algebraic equations is too complicated in the context of obtaining the closed-form solutions. By using the method in Luo and Tong (2002), two sets of three simultaneous equations with three unknowns can be obtained, respectively. Therefore, explicit analytical solutions can be derived. The details can be found in Luo and Tong (2002) and are not discussed further here due to the space limitation. Geometrically Nonlinear Analysis for Unbalanced Single-Lap Joint Fully coupled nonlinear analysis for unbalanced SLJs was conducted by Li et al. (2015) with flexible interfaces being considered. Free body diagrams for the unbalanced SLJ are the same as those in Fig. 6b except that adherends 1 and 2 are not identical. By using the Timoshenko beam theory and the free body diagrams, the governing equations for the unbalanced SLJ with geometrical nonlinearity can be derived. The nonlinear terms in the governing equations were linearized by using following transformation (Jiang et al. 2016): N1 ¼

Ad1 P Ad2 P ; N1 ¼ Ad1 þ Ad2 Ad1 þ Ad2

(56)

where Ad1 and Ad2 are the extensional stiffness of adherends 1 and 2; P = N1 + N2. The linearized coupling differential equations are (Jiang et al. 2016): d6 w1 d 5 u1 d4 w1 d 3 u1 d2 w1 du1 þ a16 w1 ¼ f ðxÞ þ a þ a þ a þ a þ a15 11 12 13 14 4 3 2 6 5 dx dx dx dx dx dx 4 3 2 d w1 d u1 d w1 du1 þ b16 w1 ¼ gðxÞ þ b13 3 þ b14 þ b15 4 2 dx dx dx dx (57) By eliminating w1 from Eq. 57 or using the approach in Sect. 3.1, a 9 order ordinary differential equation can be obtained (Eq. 23 in Jiang et al. 2016). By letting λ ¼ β2t , a 4 order of polynomial equation is derived, and thus the analytical solutions for unbalanced geometrically nonlinear SLJs can be derived. Edge moment factors k1 and k2 are defined as: k1 ¼

2M1 2M2 ; k2 ¼ Pð t 1 þ t a Þ Pð t 1 þ t a Þ

(58)

24

Analytical Approach

691

where M1 and M2 are the undated bending moments at edges I and II. Other parameters in Eqs. 56–58 can be found in Jiang et al. (2016).

24.5

Failure Criteria Based on Continuum Mechanics

Failure of adhesively bonded structures can occur in adherends, adhesive, and interface. In this chapter, only failures in an adhesive and at interface between the adherend and adhesive are considered, referring to cohesive failure and debonding or delamination. The adhesive shear and peel stresses peak at adhesive edges, and the peak stresses increase to infinity with the adhesive thickness decreasing to zero. Therefore, the failure of the adhesive joint will be initiated from the adhesive edge generally. There are two types of failure criteria to assess adhesive joint strength, based on continuum mechanics and fracture mechanics, respectively. The failure criteria based on continuum mechanics can be summarized as (Huang et al. 2002): Maximum stress criterion : σ max < ½σ  and τmax < ½τ

(59a)

Von Mises criterion : ðσ von Þmax < ½σ 

(59b)

 Tsai-Wu criterion :

σ max ½σ 

2

 þ

τmax ½τ

2 20(t1 + ta + t2), the joint length may be approximately treated to be semi-infinitive. In this case, the adhesive stresses can be found, and the analytical solutions are given in Luo and Tong (2009c).

24.5.1 Shear and Peel Stresses of an Adhesive Joint with a Semiinfinitive Length When the length of an adhesively bonded composite joint is semi-infinitive, the shear and peel stresses can be expressed as (Luo and Tong 2009b, c):

692

L. Tong and Q. Luo



τ ¼ A2 eβ1 x þ eβ2 x ðA5 sin β3 x þ A6 cos β3 xÞ Δ0 σ ¼ B2 eβ1 x þ B4 eβ2 x þ B6 eβ3 x

(60a)

(60b)

By substituting Eqs. 60a or 60b into the first or second equation in Eq. 45, three relationships of the integration constants are obtained. By eliminating three unknowns from the three relationships and the three boundary conditions, a set of three-order algebraic equations can be derived for the coupled shear and peel stresses, which can be solved analytically. An explicit form of shear and peel stresses can be found, and the details are given in Luo and Tong (2009b, c).

24.5.2 Failure Prediction Using the Criteria Based on Continuum Mechanics When the maximum stresses are determined and the allowable stresses are calibrated, failure criteria based on continuum mechanics can be used to predict adhesive joint failure. When the failure criteria in Eqs. 59a and 59b are used to predict joint failure, it is found (Adams 1989; Gleich et al. 2001; da Silva et al. 2006) that the failure trend for different adhesive thickness predicted by the adhesive-beam models proposed by Goland and Reissner (1944) is different from that of the experimental results. In theory, the maximum stresses increase with decreasing adhesive thickness. However, failure load decreases with increasing the adhesive thickness when it is relatively thick, such as ta > 0.1 mm. It is believed that the peak stresses in the thicker adhesive allow yielding to spread, and the global yielding may develop before the maximum stress reaches the allowable value (Hart-Smith 1973; da Silva et al. 2009). When the global yielding occurs, the adhesive thickness effect on the failure load may be predicted by the material nonlinear analysis (see details in da Silva et al. 2009). Gleich et al. (2001) showed the partial agreement with this trend when the fracture mechanics approach was used. Due to complex and singular feature of stresses in the vicinity of an adhesive edge, failure criteria based on the damage zone model and fracture mechanics have been used to predict joint failure.

24.6

Failure Criteria Based on Fracture Mechanics

As the adhesive is normally thin in bonded structures, stress concentration occurs near adhesive edges. The adhesive edge stresses increase with a decrease in adhesive thickness. As the adhesive thickness approaches zero, the edge shear and peel stresses defined by Eqs. 8 or 10 will reach infinite. That is, stress singularity appears near the adhesive edges. The singular stress fields in a free edge of laminates and at interface of dissimilar materials have been widely studied.

24

Analytical Approach

693

Here, failure criteria based on energy release rates for adhesive joints will be discussed.

24.6.1 Failure Prediction Using Energy Release Rates Failure criteria based on fracture energy release rates can be written as (Fernlund et al. 1994): rffiffiffiffiffiffiffi
2
 GII GI GII 2 GT < GTc at φ ¼ A tan or þ

>

  ¼  þ G  I > > 2 D2 k1d Ad2 Ad1 D1 2 k1d k4d  k22d > > > rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2 <       k22d Q2 Q1 k5d M1 2  M D2  D1  kf k4d 1  k1d k4d D2  D1 > > > 

 > > > 1 N2 N1 1 t2 M2 t1 M1 2 > : GII ¼  þ  2k1d Ad2 Ad1 2 D2 D1

(66)

in which Nk, Mk and Qk (k = 1, 2) are the force components in adherends 1 and 2 at the joint ends I where the fracture pffiffiffi is initiated as shown in Fig. 12.

In Eq. 66, kf 1  kf  2 reflects the influence of interface deformation near pffiffiffi the crack-tip. As ta ! 0, kf ! 1; and when Δ = 0, kf ¼ 2. When the Euler beam theory is used to model the adherends, k5d = 0. Equation 66 is reduced to the energy release rates for debonding or delamination derived based on the Euler beam theory (Luo and Tong 2009a). If B1 = B2 and k2d = 0, the energy release rates can be simplified as: 8 2 32 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi > 2 > > Ad1 B1 Ad1 D1  B1 > 4 > ðQ 2  Q 1 Þ 5 > < GI ¼ 4 Ad1 D1  B2  ðM2  M1 Þ  Ad1 ðN 2 þ N 1 Þ  kf Ad1 Gk1 1 " #2 > > 2 > A D  B 2D  B t A t  2B d1 1 1 1 1 d1 1 1 > 1 > GII ¼

 ðN 2  N 1 Þ 

 ðM 2 þ M 1 Þ > : 16ð4αa1 D1 þ t1 B1 Þ Ad1 D1  B21 Ad1 D1  B21

(67) The energy release rate GI (GII = 0) for a double cantilever beam with crack length a subjected to a pair of load P is: 2 sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi32 kf Ad1 D1  B21 5 Ad1 ðPaÞ2 4  1þ GI ¼

or a Ad1 Gk1 4 Ad1 D1  B21 rffiffiffiffiffiffiffi2
k f D1 ðPaÞ2 1þ when B1 ¼ 0 GI ¼ D1 a Gk1

(68)

696

L. Tong and Q. Luo

Energy release rates in Eq. 65 are derived on the basis of adhesive edge stress by letting the edge adhesive thickness approach zero. Similar results are obtained in Bruno et al. (2003) and Wang and Harvey (2012) based on the beam theory. That is, these formulations are global ones. For an interface crack or debonding of an adhesive joint, local details (stress, force, or deformation) near a crack-tip should be taken into account. By using the beam theory and considering the local details, analytical formulations of the mode mixity ratios can be derived (Luo and Tong 2012; Harvey et al. 2014; Li et al. 2015). For the isotropic adherends, two simple formulations are given in (Luo and Tong 2012): 2

GII 3ð 1  r 2 Þ ¼ when M1 ¼ M2 ; N 1 ¼ N 2 ¼ Q1 ¼ Q2 ¼ 0 GI ð2 þ r þ 2r 2 Þ2

(69)

2

GI 3ð 1  r 2 Þ ¼ when N 1 ¼ N 2 ; M1 ¼ M2 ¼ Q1 ¼ Q2 ¼ 0 GII ð2 þ r þ 2r 2 Þ2

(70)

where r (=t1/t2) is the thickness ratio. The numerical studies in Harvey et al. (2014) and Li et al. (2015) show that Eqs. 69 and 70 are quite accurate when 0.1  r  10.

Numerical Comparison of Energy Release Rates for the Cracked Lap Shear Specimen An ASTM Round Robin for calculating energy release rates of cracked lap shear specimens (CLS) was reported by Johnson (1986). In this work, the energy release rates were calculated by nine groups for the two ASTM CLS specimens with four different debonded lengths (a = 53.54, 57.35, 76.4 and 152.6), one with symmetrical substrates (CLS-A: t1 = t2) and the other with asymmetrical substrates (CLS-B: t1 = 2 t2). The CLS-B specimen configurations are shown in Fig. 13. The material properties are E1 = E2 = 72.45 GPa, ν1 = ν2 = 0.33, Ea = 1.932, and νa = 0.4. It is indicated that the total energy release rate and mode ratio GI/GII predicted by the geometrically nonlinear finite element analysis (NFEA) were lab-to-lab consistent with the observed experimental behaviors (Johnson 1986). Fernlund et al. (1994) calculated the total energy release rate and mode ratio GI/GII for the CLS-A and CLS-B specimens using both local and global methods. The mode ratios for the CLS-A specimen predicted by both methods are almost the same as those of the NFEA, while the mode ratios for the CLS-B specimen predicted by both methods are considerably different from those of the NFEA. For the CLS-B specimen, the force components at the adhesive edge can be calculated using the formulations in Tong and Luo (2008). The adhesive edge stresses are then calculated using Eq. 57. The total energy release rate and mode ratio GI/GII are then obtained by using Eq. 60a or 60b. The total energy release rate for the CLS-B specimen predicted by Eqs. 57 and 60 is almost the same as that of the NFEA (Tong and Luo 2008). The mode ratios predicted by theoretical formulations and the NFEA are illustrated in Fig. 13. In Fig. 13, NFEA-1, NFEA-2, and NFEA-3

24

Analytical Approach

697

0.9 0.8 L (= 305 mm)

Mode ratio GI /GII

0.7

Global-Fernlund et al. (1994)

a

Local-Fernlund et al. (1994)

Adherend 1

0.6

NFEA-1-Johnson (1986)

Crack tip

0.5 F

Adherend 2

c

NFEA-2-Johnson (1986) NFEA-3-Johnson (1986)

Adhesive 0.4

Edge stress-Tong and Luo (2009c)

0.3 0.2 0.1 0

0

20

40

60

80

100

120

140

160

180

200

Crack length a (mm)

Fig. 13 The mode ratios of the CLS-B specimen predicted by theoretical formulations and the geometrically-nonlinear finite element analyses

are referred to the NFEA results in Johnson (1986); the mode ratios predicted by using the local and global methods are taken from Fernlund et al. (1994); and the mode ratios calculated by the adhesive edge stresses are from Tong and Luo (2008). It is noted that the coupling effect of interface shear and peel stresses are neglected in the global mode partition method in Williams (1988) and Fernlund et al. (1994). Since both local and global methods do not consider the adhesive layer, the mode ratios predicted by these methods are considerably different from that of the NFEA. The mode ratio predicted by the adhesive edge stresses based on the adhesive-beam model (Luo and Tong 2009b, c) correlates well with the NFEA given in Johnson (1986).

24.6.3 Failure Prediction Considering Plastic Deformation Fracture mechanics-based criteria have shown promising agreement with the experimental results for brittle adhesives (Adams 1989; Fernlund et al. 1994; Tong 1996; Van Tooren 2004; Luo and Tong 2009c). The energy release rate criterion can be applied to cohesive failure and interface fracture. When it is used, energy release rate and its critical value for each mode should be calculated and measured. When fracture mechanics-based criteria are used, fracture envelopes are normally set up to assess crack initiation and propagation. For ductile adhesives, plastic deformation occurs before structural failure and the cohesive zone model has been widely used to

698

L. Tong and Q. Luo

assess structural failure of adhesively bonded structures (Sheppard et al. 1998; Blackman et al. 2003; Liljedahl et al. 2006).

24.7

Summary

Analytical approaches for adhesively bonded structures are presented in this chapter. Stress analysis for adhesively bonded joints is conducted using the classical adhesive-beam model and the other adhesive-beam models. Closed-form solutions of symmetric joints are presented, and analytical procedures of asymmetric and unbalanced joints are discussed. Load update for single-lap joints is investigated in detail. Numerical results calculated using the classical and other formulations are illustrated and compared. It is shown that the nonlinear adhesive-beam model based on the Timoshenko beam theory provides enhanced results compared to the linear adhesive-beam model based on the Euler beam theory for adhesively bonded composite structures. Analytical solutions of energy release rates for cohesive failure and delamination are presented, and several failure criteria are reviewed and discussed. Acknowledgment The authors are grateful for the support of the Australian Research Council via a Discovery Projects grant (DP140104408).

References Adams RD (1989) Strength predictions for lap joints, especially with composite adherends: a review. J Adhes 30(1–4):219–242 Adams RD, Mallick V (1992) A method for the stress analysis of lap joints. J Adhes 38 (3–4):199–217 Adams RD, Comyn J, Wake WC (1997) Structural adhesive joints in engineering. Chapman and Hall, London Allman DJ (1977) A theory for the elastic stresses in adhesive bonded lap joints. Q J Mech Appl Math 30:415–436 Bigwood DA, Crocombe AD (1989) Elastic analysis and engineering design formulae for bonded joints. Int J Adhes Adhes 9(4):229–242 Bigwood DA, Crocombe AD (1990) Nonlinear adhesive bonded joint design analyses. Int J Adhes Adhes 10(1):31–41 Blackman BRK, Hadavinia H, Kinloch AJ, Williams JG (2003) The use of a cohesive zone model to study the fracture of fibre composites and adhesively-bonded joints. Int J Fract 119(1):25–46 Bruno D, Greco F, Lonetti P (2003) A coupled interface-multilayer approach for mixed mode delamination and contact analysis in laminated composites. Int J Solids Struct 40(26):7245–7268 Carpenter W (1980) Stresses in bonded connections using finite elements. Int J Numer Methods Eng 15(11):1659–1680 Carpenter WC (1991) A comparison of numerous lap joint theories for adhesively bonded joints. J Adhes 35(1):55–73 Chen D, Cheng S (1983) An analysis of adhesive-bonded single lap joints. J Appl Mech 50(1):109–115 Delale F, Erdogan F, Aydinoglu MN (1981) Stresses in adhesively bonded joints-a closed form solution. J Compos Mater 15:249–271

24

Analytical Approach

699

Fernlund G, Spelt JK (1991) Failure load prediction of structural adhesive joints, part 1: analytical method. Int J Adhes Adhes 11(4):213–220 Fernlund G, Papini M, McCammond D, Spelt JK (1994) Fracture load predictions for adhesive joints. Compos Sci Technol 51(4):587–600 Gleich DM, Van Tooren MJL, Beukers A (2001) Analysis and evaluation of bondline thickness effects on failure load in adhesively bonded structures. J Adhes Sci Technol 15(9):1091–1101 Goland M, Reissner E (1944) The stresses in cemented joints. J Appl Mech 11:A17–A27 Grant P, Teig IC (1976) Strength and stress analysis of bonded joints, British Aircraft Corporation Ltd. Military Aircraft Division, Report No SOR PJ 109 Harris JA, Adams RD (1984) Strength prediction of bonded single lap joints by non-linear finite element methods. Int J Adhes Adhes 4(2):65–78 Hart-Smith LJ (1973) Adhesive-bonded single-lap joints, NASA Langley research Center, NASA CR-112235 Harvey CM, Wood JD, Wang S, Watson A (2014) A novel method for the partition of mixed-mode fractures in 2D elastic laminated unidirectional composite beams. Compos Struct 116:589–594 Huang H, Yang CD, Tomblin JS, Harter P (2002) Stress and failure analyses of adhesive-bonded composite joints using ASTM D3165 specimens. J Compos Technol Res 24(2):93–104 Humi M, Miller W (1988) Second course in ordinary differential equations for scientists and engineers. Springer-Verlag, New York Hutchinson JW, Suo Z (1992) Mixed mode cracking in layered materials. Adv Appl Mech 29:63–191 Jiang Z, Wan S, Zhong Z, Li M (2016) Geometrically nonlinear analysis for unbalanced adhesively bonded single-lap joint based on flexible interface theory. Arch Appl Mech 86(7):1273–1294 Johnson WS (1986) Stress analysis of the cracked lap shear specimen-an ASTM round robin, NASA Technical Memorandum 89006 Lee J, Kim H (2007) Elasto-plastic analysis of adhesively bonded symmetric single lap joints under in-plane tension and edge moments. J Adhes 83(9):837–870 Li W, Cheng G, Wang D, Wu J (2015) A mixed mode partition method for delaminated beam structure. Eng Fract Mech 148:15–26 Liljedahl CDM, Crocombe AD, Wahab MA, Ashcroft IA (2006) Damage modelling of adhesively bonded joints. Int J Fract 141(1–2):147–161 Luo Q, Tong L (2002) Exact static solutions to piezoelectric smart beams including peel stresses–I: theoretical formulation. Int J Solids Struct 39(18):4677–4695 Luo Q, Tong L (2004) Linear and higher order displacement theories for adhesive bonded lap joints. Int J Solids Struct 41(22–23):6351–6381 Luo Q, Tong L (2007) Fully-coupled nonlinear analysis of single lap adhesive joints. Int J Solids Struct 44(7–8):2349–2370 Luo Q, Tong L (2008) Analytical solutions for adhesive composite joints considering large deflection and transverse shear deformation in adherends. Int J Solids Struct 45(22–23):5914–5935 Luo Q, Tong L (2009a) Calculation of energy release rates for cohesive and interlaminar delamination based on the classical beam-adhesive model. J Compos Mater 43(4):331–348 Luo Q, Tong L (2009b) Energy release rates for interlaminar delamination in laminates considering transverse shear effects. Compos Struct 89(2):235–244 Luo Q, Tong L (2009c) Fracture prediction of adhesively bonded structures using energy release rates. J Adhes Sci Technol 23(10):1415–1440 Luo Q, Tong L (2012) Analytic formulas of energy release rates for delamination using a global–local method. Int J Solids Struct 49(23–24):3335–3344 Luo Q, Tong L (2016) Solutions for clamped adhesively bonded single lap joint with movement of support end and its application to a carbon nanotube junction in tension. J Adhes 92(5):349–379 Ojalvo IU, Eidinoff HL (1978) Bond thickness upon stresses in single-lap adhesive joints. AIAA J 16(3):204–211 Oplinger DW (1994) Effects of adherend deflection on single lap joints. Int J Solids Struct 31(18):2565–2587

700

L. Tong and Q. Luo

Renton WJ, Vinson (1975) The efficient design of adhesive bonded joints. J Adhes 7(3):175–193 Shahin K, Kember G, Taheri F (2008) An asymptotic solution for evaluation of stresses in balanced and unbalanced adhesively bonded joints. Mech Adv Mater Struct 15(2):88–103 Sheppard A, Kelly D, Tong LY (1998) A damage zone model for the failure analysis of adhesively bonded joints. Int J Adhes Adhes 18(6):385–400 da Silva LFM, Rodrigues TNSS, Figueiredo MAV, de Moura MFSF, Chousal JAG (2006) Effect of adhesive type and thickness on the lap shear strength. J Adhes 82(11):1091–1115 da Silva LFM, das Neves PJC, Adams RD, Spelt JK (2009) Analytical models of adhesively bonded joints-part II: comparative study. Int J Adhes Adhes 29(3):331–341 Srinivas S (1975) Analysis of bonded joints, NASA TN D-7855, Apr 1975 Tong L (1996) Bond strength for adhesive-bonded single-lap joints. Acta Mech 117(1–4):101–113 Tong L, Luo Q (2008) Chapter 2 Analysis of cracked lap shear (CLS) joints. In: da LFM S, Oechsner A (eds) Modelling of adhesively bonded joints. Springer, Heidelberg Tong L, Steven GP (1999) Analysis and design of structural bonded joints. Kluwer Academic, Boston. 1999 Van Tooren MJL (2004) Experimental verification of a stress singularity model to predict the effect of bondline thickness on joint strength. J Adhes Sci Technol., 2004 18(4):395–412 Volkersen O (1938) “Die Nietkraftverteilung in Zugbeanspruchten Nietverbindungen mit Konstanten Laschenquerschnitten” (The rivet load distribution in lap-joints with members of constant thickness subjected to tension). Luftfahrtforschung 15:41–47 Wang S, Harvey C (2012) A theory of one-dimensional fracture. Compos Struct 94(2):758–767 Williams JG (1988) On the calculation of energy release rates for cracked laminates. Int J Fract 36(2):101–119 Yang C, Pang SS (1993) Stress–strain analysis of adhesive-bonded single-lap composite joints under cylindrical bending. Compos Eng 3(11):1051–1063 Yang CD, Huang H, Tomblin JS, Sun WJ (2004) Elastic-plastic model of adhesive-bonded singlelap composite joints. J Compos Mater 38(4):293–309 Yousefsani SA, Tahani M (2013) Analytical solutions for adhesively bonded composite single-lap joints under mechanical loadings using full layerwise theory. Int J Adhes Adhes 43:32–41 Zhao B, Lu ZH (2009) A two-dimensional approach of single-lap adhesive bonded joints. Mech Adv Mater Struct 16(2):130–159 Zhao X, Adams RD, da Silva LFM (2010) A new method for the determination of bending moments in single lap joints. Int J Adhes Adhes 30(2):63–71

Numerical Approach: Finite Element Analysis

25

Ian A. Ashcroft and Aamir Mubashar

Contents 25.1

25.2

25.3

25.4

25.5

25.6

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.1.1 Brief History of the Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.1.2 Recent Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.1.3 Applications of Finite Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.1.4 Comparison with Other Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.2.1 Fundamentals of the Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.2.2 Natural Coordinates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.2.3 Isoparametric Elements and Numerical Integration . . . . . . . . . . . . . . . . . . . . . . . . . . . Nonlinear Finite Element Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.1 Sources of Nonlinearity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.3.2 Solution Algorithms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermal Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.4.2 Heat Conduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dynamic Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5.2 Free Vibration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.5.3 Dynamic Response Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Application of the Finite Element Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6.1 Preprocessing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6.2 Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.6.3 Post-processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

702 703 704 707 708 708 709 711 712 714 714 715 716 716 716 718 718 719 720 720 720 724 725

I. A. Ashcroft (*) Faculty of Engineering, University of Nottingham, Nottingham, UK e-mail: [email protected] A. Mubashar Mechanical Engineering, Middle Eastern Technical University, Northern Cyprus Campus, Akara, Turkey e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_25

701

702

I. A. Ashcroft and A. Mubashar

25.7

726 726 727 730 732 737 738

Application of FEA to Adhesive Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7.1 General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7.2 Stress Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7.3 Failure Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.7.4 Advanced Analysis Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25.8 Summary and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

Finite element analysis (FEA) is a widely used computer-based method of numerically solving a range of boundary problems. In the method, a continuum is subdivided into a number of well-defined elements that are joined at nodes, a process known as discretization. A continuous field parameter, such as displacement or temperature, is now characterized by its value at the nodes, with the values between the nodes determined from polynomial interpolation. The nodal values are determined by the solution of an array of simultaneous equations using computational matrix methods, and the accuracy of the results is dependent on the discretization, the accuracy of the assumed interpolation form, and the accuracy of the computation solution methods. The current popularity of the method is based on its ability to model many classes of problem regardless of geometry, boundary conditions, and loading. Modeling the behavior of adhesive joints is complicated by a number of factors, including the complex geometry, the complex material behavior, and the environmental sensitivity. FEA is currently the only technique that can comprehensively address the challenges of modeling bonded joints under realistic operating conditions. However, a reliable and robust method of using FEA to model failure in bonded joints is still to be developed.

25.1

Introduction

In its 50-year life, the finite element method (FEM) has progressed from a specialist technique of the aerospace stress analyst to a ubiquitous tool of advanced engineering analysis, widely taught in undergraduate programs and the foundation of a billiondollar industry. It is based on two basic concepts. The first is the representation of a continuum as a number of simple, well-defined sub-elements with a finite number of degrees of freedom. The second is the solution of differential equations by the weak (or generalized) method through the application of numerical integration techniques. This reduces a potentially unsolvable differential equation to a series of simple, linear simultaneous equations, which can be solved through the application of matrix methods. The methods at the core of the finite element method were proposed in the nineteenth century, but it was only with the development of microprocessor computers that the solution of the large number of simultaneous equations required to represent complex problems became practical. The power of the method is in its simplicity and generality, being unbounded by geometry, loads, boundary conditions, or class of problem. However, it should be remembered that both the discretization process and

25

Numerical Approach: Finite Element Analysis

703

the numerical solution methods used in the finite element method involve errors, and, hence, the FEM will always present an approximate solution to a problem. In most cases, these errors can be minimized and are usually more than compensated for by the reduction in errors associated with simplifying a problem to a level that enables an exact solution to be determined. The mathematical basis of the FEM is now well established and its role as a practical engineering tool widely recognized. Further development, application, and usage of the method are envisaged in the foreseeable future. Although numerous analytical methods for determining stresses in bonded joints have been proposed, they generally require significant simplification of geometry, boundary conditions, and material behavior; even then, the more advanced analyses are exceedingly complex, sometimes requiring numerical methods of solution. The application of FEM to bonded joints, with its capability of handling complex geometry, material behavior, and boundary conditions, as well as nonlinearity and multiphysics modeling, is potentially very attractive. However, a number of cautions should be observed. Firstly, the thickness of the adhesive layer with respect to other dimensions is usually small, raising problems of creating a suitable mesh with a reasonable number of elements, while avoiding excessive element distortion. Secondly, adhesive material behavior can be exceedingly complex and will vary considerably with environment, presenting problems in adopting a suitable material constitutive law. Thirdly, the interface between the adhesive and adherend is difficult to model accurately, and at certain points, theoretical singularities exist at the bi-material junction, presenting difficulties in applying standard failure criteria. These issues can all be addressed to some degree to provide effective analyses of bonded joints, and it is envisaged that further advances in these areas will be made and that the application of FEA to bonded joints will continue to increase. However, despite all the advantages, it should still be remembered that the FEM is an approximate technique, and the integrity of results is highly dependent on the skills of the analyst. This chapter is subdivided into a number of sections. This section provides a general background to the development and application of the FEM. Section 2 provides the fundamentals of the method, and Sect. 3 discusses some of the practical aspects of finite element modeling. Section 4 discusses application of the FEM to adhesively bonded joints, and Sect. 5 provides conclusions from the chapter and indicates potential future developments in the application of FEA to bonded joints.

25.1.1 Brief History of the Finite Element Method The finite element method is a numerical analysis procedure that provides an approximate solution to problems in various fields of engineering. It is based on the matrix methods of structural analysis of the 1920s and 1930s. Advances in the aerospace industry and the development of computers in the 1950s and 1960s saw further development and computerization of these methods. Interest in delta wings leads to the requirement for a two-dimensional (2D) element of arbitrary shape. In 1956, the stiffness of a general 2D triangular element was derived (Turner et al. 1956), and in 1959, the direct stiffness method as a general, computer-based method

704

I. A. Ashcroft and A. Mubashar

of structural analysis was proposed (Turner 1959). This was the direct precursor of nearly all current commercial finite element analysis (FEA) methods. The term finite element was introduced by Clough (1960), and the first widely used finite element program was written by Wilson in the early 1960s (Fish and Belytschko 2007). The applicability of the method to general field applications was described by Zienkiewicz and Cheung (1967), and by the end of the 1960s, finite element analysis (FEA) was being applied to a wide variety of boundary problems, such as heat transfer, fluid flow, dynamics, free and forced vibration, and electrical and magnetic fields. The mathematical foundations of the finite element method were laid down in the 1970s (Champion and Ensminger 1988). In the 1980s, generalized software packages designed to run on powerful computers were developed, and improved techniques for analyzing nonlinear problems were established. The personal computer (PC) was introduced in the late 1970s, and by the 1990s, the PC became sufficiently powerful to be practical for FEA. Commercial FEA software for PCs was developed and became widely used. FEA is now used in practically all fields of engineering analysis and is the clearly favored technique for many industries, including aerospace, automotive, nuclear, electronics, and defense. Several software vendors offer extremely powerful, general purpose, finite element software which can solve a wide range of problems.

25.1.2 Recent Developments Finite element modeling was once the domain of a few high-technology industries and academic communities. An increase in the sophistication of the method and a widening of the user base has been accompanied by a shift from in-house software codes to powerful, user-friendly commercial software. The market is now dominated by a relatively small number of general commercial software packages and a number of more specific codes satisfying niche markets. Most of the commercial packages are capable of user modification, such as the introduction of user-written elements or material constitutive models. A recent trend in the commercially available software is the availability of multiphysics simulations. The finite element method can be applied to many classes of problems such as structural, heat transfer, material transport (such as diffusion), fluid mechanics, and electromagnetic. Multi-physics analysis is concerned with the coupling of different analysis classes, such as thermo-structural, fluid-chemical, electrothermal, fluid-structural, hygro-thermo-structural, etc. In multi-physics modeling, the nature of the coupling must be considered. For example, in a fully coupled thermomechanical analysis, both the effect of the thermal field on the structural response and the effect of the structural response on the thermal field are represented. This is implemented by partitioning the structural and field analyses using common time integration. The thermal and structural analyses are then performed either concurrently or sequentially, and communication is achieved via a common data transfer file. For strongly coupled systems, data transfer should be on the iterative level, but for weakly coupled systems, an incremental data transfer is usually sufficient. There may also be

25

Numerical Approach: Finite Element Analysis

705

situations that a single software or type of analysis is not able to solve for all the physics inherent in certain complex structures ad systems. Such situations include problems involving fluid and structural domain interaction, system-level rigid body dynamics with inclusion of flexible parts, or inclusion of hydraulic/pneumatic control in a finite element-based simulation. In such cases, multiple solvers can be linked together by using software that acts as controller for all the involved solvers. The control software facilitates the transfer of information between the solvers during the simulation process and also provides mapping between field variables. The information exchange between solvers is generally linear, where one solver completes a job and the results are provided to the next solver. Such a linkage enables optimization and design space exploration studies involving various fields that fully interact with each other. However, the computational requirements of running these studies are generally high as multiple iterations are required to find an optimized solution. The limitations of computational requirements have been mostly met by the use of high performance computing clusters (HPCs). HPCs provide a multi-node, multicore configuration where a master node schedules the jobs received from multiple users. Installation and maintenance of HPCs is a costly business, and many companies are now moving toward cloud computing. In cloud computing, Internet-based services may be used to access remote compute clusters for the solution of large finite element-based problems. The access to cloud-based computing resources can be requirement based, which may help in reducing the cost of an analysis. The current issues with the use of cloud computing are related to the availability of high-speed Internet and data security. In a drive to make the finite element method more accessible to users of various levels of experience, packaging of complex finite element analysis in the form of apps is becoming increasingly popular. A few commercial finite element software (such as COMSOL) are now providing this as a built-in capability and providing server applications for deploying the apps across the enterprise. The advantage of this approach is that complex finite element analysis models, which take a considerable time, effort, and level of expertise to setup, are developed and validated by experts in the field. These validated finite element models are then packaged as apps, allowing predetermined parameters to be changed in a controlled fashion, thus enabling users less experienced in finite element theory to carry out various types of studies. Another area of current development is in damage and failure modeling. It is impossible for a linear finite element analysis to predict failure in a structure. However, in nonlinear analysis, it is possible to implement a failure model, and increasingly complex failure models are now supplied as standard features of commercial FEA software. It should be noted, however, that even though the failure model may look complex, the method of implementation within FEA is usually fairly straightforward. In most cases, this involves utilization of the results from the various increments of a nonlinear analysis. These are processed via some failure model to determine whether failure or damage has occurred in any of the elements. The properties of the failed or damaged elements are then modified, usually by control of the element stiffness matrix, removal of elements, or affecting nodal connectivity/element interaction. One of the simplest failure models is that based on plastic yielding. In this case, progressive yielding of the structure is modeled until a “limit state” or state of “global

706

I. A. Ashcroft and A. Mubashar

yielding” is reached. Fracture mechanics principles can also be used within an FEA. This can be based on either the stress intensity factor or energy approaches. An extension to this approach is cohesive zone modeling (CZM), which also allows for damage of a material ahead of a crack. The CZM technique allows the modeling of damage and failure in ductile (Dugdale 1960; Barenblatt 1962) and quasi brittle (Hillerborg et al. 1976) materials. The technique uses a layer of cohesive elements to represent a potential crack path or interface. The virtual crack closure technique (VCCT) for the determination of strain energy release rate was proposed by Rybicki and Kanninen on 1977 but has only recently been implemented in major commercially available finite element software. Another method for modeling cracks in materials is the extended finite element method (XFEM). The XFEM uses enriched shape functions to represent a discontinuous displacement field (Mohammadi 2008). The main advantage of XFEM is that the crack may initiate at any point in the material and propagate based on loading conditions. No remeshing is required as the crack can grow within an element and need not follow element boundaries. In order to overcome the individual drawbacks of CZM and XFEM methods, when applied to the adhesive joints, a combination of the two methods has been used (e.g., Mubashar et al. 2014; Stuparu et al. 2016). In such adhesive joint models, the XFEM is used to simulate cohesive damage and failure within the adhesive layer, whereas CZM is used to represent interfacial damage and crack growth. This enables crack initiation and propagation that can transfer from a cohesive crack to an interfacial crack and vice versa. However, such highly nonlinear models suffer from non-convergence problems, and some form of viscous stabilization is generally used to obtain a converged solution. Neumayer et al. (2016) attempted to develop a cohesive element that is able to predict cohesive as well as interfacial failure in the adhesive layer. The advantage of such a formulation is that a single layer of cohesive elements would be able to correctly predict the type of failure in an adhesive joint. The reduction in the finite element model size comes from using structural elements such as beams or shell elements for adherends. The proposed finite element formulation was used to determine the strength of single lap adhesive joints where it provided good prediction of failure strength of the joint under cohesive failure. However, for interfacial failure, the strength of the adhesive joints was underestimated by the proposed cohesive element. Another methodology to alleviate problems associated with continuum meshes during large deformation simulations is to use meshless methods. One such method is smoothed particle hydrodynamics (SPH), which is generally used for highly nonlinear solid mechanics problems. In this method, a material continuum is approximated by discretization involving particles rather than elements, which allows for large deformations and crack propagation along arbitrary paths. The SPH method is a particlebased representation of the partial differential equations of a continuum. The particles are not directly connected to each other but are affected by neighboring particles. This means that the value of field variable at a point also includes contributions from the neighboring elements, where the number of neighboring elements contributing to a particle is based on a kernel function. The value of a field variable is determined by an interpolation scheme. Displacement is the field variable in the case of a structural analysis, as in the finite element method. A few attempts have been made to apply this method to adhesive joints (e.g., Tsai et al. 2014; Mubashar and Ashcroft 2015).

25

Numerical Approach: Finite Element Analysis

707

Continuum damage mechanics-based models are now available and implemented in commercial finite element software. Damage may be regarded as the growth of microvoids or microcracks in a material (Lemaitre and Desmorat 2005). Continuum damage mechanics was introduced in the late 1950s (Kachanov 1958) and is able to predict the reduced load bearing capacity of a material before failure. Damage is related to the growth of microcracks and microvoids in a material. Damage in a material is generally defined by a damage variable that has constitutive equations for evolution. The damage variable may be defined such that it is based on some physical characteristic of the material or is a representation of the effects of damage, e.g., reduction in stiffness. Damage may be isotropic or anisotropic and has also been defined for fatigue and creep loading conditions. The damage model combined with a failure criterion can be used in conjunction with plasticity models, providing a complete definition of material behavior.

25.1.3 Applications of Finite Element Analysis Owing to the general nature of the finite element method, it can be used to solve boundary value problems in many fields. The initial applications of the finite element method were in aircraft structures, but it is now widely applied to the solution of many other engineering problems. These include the stress and thermal analysis of components, crash analysis of vehicles and aircraft, seismic analysis of civil engineering structures, nuclear engineering, fluid flow analysis, soil structure interaction, steady-state and transient behavior of electronic devices, and electromagnetics. The use of composites is increasing in modern structures, such as aircrafts and automobiles, and FEA provides a convenient method to determine the performance and durability of composites. The fabrication of different types of composite lay-up can be explored at design stage to avoid problems during manufacturing. FEA can also be used in the optimization of structural components where automatic mesh generation makes the process highly automated. Full vehicle body noise and vibration analysis are performed to improve the performance of modern automobiles. Virtual crash testing of aircraft and drop testing of electronic devices such as mobile phones and laptops may now be conducted using commercially available FEA software. Applications in offshore structures include ship design and reliability studies of oil and gas equipment. Manufacturing industry is also taking advantage of the capabilities of finite element method where finite element method-based simulation models are used to develop better cutting tools, estimate the best machining conditions, predict residual stresses, and determine the shape of a part after forming processes. A recent area of application is in biomedical engineering where FEA is used to determine the performance of implants such as heart valves and the expansion of stents during angioplasty. Prosthesis designs may be analyzed using FEA so they can be adapted to a particular patient. Other biomedical applications involve modeling of bone, biological soft tissues, blood flow, and blood vessels (Kojić et al. 2008). Some examples of finite element models, including a heart stent, chassis for multi-axle

708

I. A. Ashcroft and A. Mubashar

vehicle, an adhesively bonded composite T-Joint, and two-dimensional orthogonal machining, are shown in Fig. 1.

25.1.4 Comparison with Other Methods Other analysis methods that use discretization include the finite difference method, the boundary element method, and the finite volume method. However, FEA is by far the most commonly used method in structural mechanics. The finite difference method approximates differential equations using difference equations. The method works well for two-dimensional problems but becomes cumbersome for regions with irregular boundaries (Segerlind 1984). Another difference between the finite element and finite difference methods is that in the finite difference method, the field variable is only computed at specific points, while in the finite element method, the variation of a field variable within a finite element is available from the assumed interpolation function (Hutton 2003). Thus the derivatives of a field variable can be directly determined in finite element method as opposed to the finite difference method where only data concerning the field variable is available. The boundary element method is also not general in terms of structural shapes (MacNeal 1994).

25.2

The Finite Element Method

This section provides a condensed view of the fundamentals of the finite element method, in which an attempt has been made to summarize the underlying theory behind the main classes of analysis that may usefully be used in analyzing Fig. 1 Examples of finite element models: (a) a heart stent, (b) chassis of a multiaxle vehicle, (c) composite T-joint, and (d) machining of material

25

Numerical Approach: Finite Element Analysis

709

adhesively bonded joints. One of the many textbooks on finite element analysis, such as Cook (1995) or Zienkiewicz and Taylor (2000), should be referred to for a more complete treatment of the subject.

25.2.1 Fundamentals of the Finite Element Method In the finite element method (FEM), a continuous structure is considered as a number of smaller elements joined at nodes. Each node has a limited number of degrees of freedom (dof). Hence, the continuum is now represented by a finite number of dof and determined by the number of elements, the number of nodes per element, and the number of dof per node. Over each element, a field quantity, such as displacement in structural analysis, is interpolated, usually adopting a polynomial form, from values of the field quantity at the nodes. By joining the elements, the field quantity is interpolated over the whole structure by an array of polynomial expressions. A set of simultaneous equations is formulated in which the primary unknowns are the values of the field quantity at the nodes. There are a number of variants of the finite element method depending on the formulation and solution of the problem. The most common variant is the stiffness method in which the displacement compatibility conditions are satisfied and the equations of equilibrium are solved to yield the unknown nodal displacements. The governing equation using matrix symbolism is of the following form: ½F ¼ ½K ½u

(1)

where [u] is a vector of unknown values of the field quantity at the nodes, [K] is a matrix of known constants, and [F] is a vector of known loads. In stress analysis, [u] is the displacement vector, [F] is the vector of applied loads, and [K] is the stiffness matrix. The most widely used methods in finite element formulation are the variational method and the method of weighted residuals. The variational approach involves the solution of a governing partial differentiation equation, such as the equilibrium equation in an elasticity problem, by determining the conditions that make some quantity (or functional) stationary (i.e., either maximum or minimum). In elasticity problems, the functional used is the total potential energy of the structure. First, an approximate form for the solution is assumed. This must be admissible, that is it must satisfy internal compatibility and essential boundary conditions. The stationary potential energy theorem is then used to determine the optimum value for the constants. This theorem states that for all admissible configurations of a conservative system, those that satisfy the equations of equilibrium make the potential energy stationary with respect to small admissible variations of displacement, i.e., equilibrium configurations are defined when @Πp ¼0 @ai

(2)

710

I. A. Ashcroft and A. Mubashar

where Πp is potential and ai is the ith dof. Πp ¼ U þ Ω

(3)

where U is the strain energy and Ω is the potential of the load system. The strain energy in a stressed body is given by ð

1 2

U¼

fegT ½EfegdV

(4)

where [E] is the elastic constant matrix, which defines the relationship between stress {σ} and strain {e}, i.e., fσ g ¼ ½Efeg

(5)

The potential of the load system is Ω ¼  f ue g T f Fe g

(6)

The total potential of the system is therefore Πp ¼ U þ Ω ¼

1 2

ð fegT ½EfegdV  fue gT fFe g

(7)

A shape function, [N], is used to relate the vector of displacements in the element, {u}, to the vector of nodal displacements, {ue}. fug ¼ ½N fue g

(8)

The strain vector can then be determined by partial differentiation of the displacement vector, giving  fe g ¼

du dx





 dN ¼ fue g ¼ ½ B fu e g dx

(9)

where [B] is the strain-displacement matrix. Substituting (9) into (7), Πp ¼

fue gT 2

ð ½BT ½E½BdV fue g  fue gT fFe g

(10)

Applying the stationary potential energy theorem, ð @Πp ¼ 0 ¼ ½BT ½E½BdV fue g  fFe g @ fue g ð fFe g ¼ ½BT ½E½BdV fue g

(11) (12)

25

Numerical Approach: Finite Element Analysis

711

and as {Fe} = [Ke]{ue}: ð ½K e  ¼ ½BT ½E½BdV

(13)

Equation 13 can be solved by numerical integration for each element.

25.2.2 Natural Coordinates It can be seen that integration of the [B] matrix is necessary for element formulation via the variational approach. A numerical integration method, such as Gauss quadrature, is used in which the real element shape is mapped onto an idealized element shape with well-defined boundaries and hence integration limits. This is achieved by transforming from the physical coordinate system (e.g., xy) to a natural coordinate system (e.g., ξη) as shown in Fig. 2. The shape functions can now be derived in terms of the natural coordinates. However, it is necessary to transform the variables to the physical coordinates in order to determine [B]. This can be achieved using the chain rule. For the 2D case in Fig. 2,      @x @N @y þ @ξ @y @ξ       @N @N @x @N @y ¼ þ @η @x @η @y @η @N ¼ @ξ



@N @x

(14) (15)

In matrix form, this can be rewritten as

h

y

x3,y3

−1,1

1,1

x4,y4 mapping x

x1,y1 x2,y2

Physical coordinates

−1,−1 x

Fig. 2 Transformation from physical to natural coordinates

1,−1

Natural coordinates

712

I. A. Ashcroft and A. Mubashar

2 3 3 @N @N 7 6 @x 7 1 6 6 @ξ 7 6 7 4 @N 5 ¼ ½J  4 @N 5 2

@y

(16)

@η

where the Jacobian matrix [J] is given by 2

@x 6 @ξ ½J  ¼ 6 4 @x @η

3 @y @ξ 7 7 @y 5

(17)

@η

The Jacobian can be regarded as the scale factor that relates the physical length with the reference length. Element distortion affects the Jacobian determinant and introduces errors into the numerical integration. Thus element distortion needs to be controlled to obtain acceptable integration results.

25.2.3 Isoparametric Elements and Numerical Integration In isoparametric elements, the same interpolation functions are used to define the variation of the primary variable across the element and the shape of the element. In other words, [N] is used to determine the element shape from the nodal coordinates [xe] (hence the term shape function, which takes on added significance for isoparametric elements) and to determine element displacement from nodal displacements. x ¼ ½N ½xe T

(18)

u ¼ ½N ½ue T

(19)

In the case of quadratic or higher interpolations, this has the advantage that the distribution of the primary variable is represented more closely and also that the element sides can be curved. The element stiffness matrix requires the strain-displacement relationship to be known, e.g., for the one-dimensional case: ex ¼

du ¼ dx



 d ½ N  fue g dx

(20)

where d dξ d ¼ dx dx dξ

(21)

25

Numerical Approach: Finite Element Analysis

713

As [N] is expressed in terms of natural coordinates, the chain rule for differentiation must be used. First, the Jacobian is calculated as J¼

dx d ¼ ½N fxe g dξ dξ

(22)

And hence the [B] matrix is calculated as ½B ¼

1 d ½N  J dξ

(23)

The stiffness matrix [k] is now obtained using ð ð ½k ¼ ½BT ½E½Bdx ¼ ½BT ½E½BJdξ

(24)

The only issue left is that Eq. 24 cannot generally be solved algebraically. Instead a numerical integration scheme (such as Gauss Quadrature) has to be used to determine the stiffness matrix. An integral having arbitrary limits can be transformed so that its limits are from 1 to 1, e.g., xð2

I¼

f ðxÞdx

(25)

φðξÞ dξ

(26)

x1

can also be expressed as ð1 I¼ 1

and therefore, as in all isoparametric formulations in FEA, the integral can be evaluated between the same limits 1 and 1. In general, the integral can be computed by sampling from a number of points and multiplying by an appropriate value or weighting factor W. This leads to the following generalization for Gauss quadrature: ð1 I¼

φðξÞdξ  W 1 φ1 þ W 2 φ2 þ . . . þ W n φn

(27)

1

where n is the number of sampling (Gauss) points. As the number of gauss points increases, more accurate integration results are obtained in general. On the other hand, using too many Gauss points would require more computational resources, and the results may not improve.

714

I. A. Ashcroft and A. Mubashar

The element stiffness matrices are summed to form the global stiffness matrix, as given in Eq. 1, which will be a symmetrical square matrix with the order of the numbers of dof in the finite element model. Sophisticated computational techniques have been developed that enable the solution of these equations for large assemblages of elements, thereby, allowing complex structures to be modeled to a high degree of accuracy. Once all the nodal degrees of freedom have been determined, it is a relatively simple procedure to determine strains in the elements from Eq. 9 and then to calculate stresses using Eq. 5.

25.3

Nonlinear Finite Element Analysis

25.3.1 Sources of Nonlinearity In a linear analysis, response is directly proportional to load. This is applicable if displacements and rotations are small, stress is directly proportional to strain, and loads maintain their original directions as the structure deforms. In a linear analysis, equilibrium reactions are written in terms of the original configuration, and displacements are obtained by solving a single set of equilibrium equations. Nonlinearity typically arises because of material or geometric nonlinearities. Nonlinearity makes the problem more difficult because the geometry, support conditions, and material properties required for the equilibrium equations are not known until the solution is known. The solution cannot be obtained in a single step, and some sort of iterative solution must be used, together with a relevant convergence test. In a nonlinear analysis, the principle of superposition cannot be applied, and a separate analysis is required for each load case. Geometric nonlinearity arises when deformations are large enough to significantly alter the way load is applied or resisted by the structure. Examples include follower loads, membrane stiffening, and snap-through effects. Finite element analysis of large deflection problems is carried out using the same elements as used in small deflection analysis, but extra attention must be paid to errors due to excessive element distortion. Nonlinear solution algorithms update the stiffness matrix in the solution process, ensuring equilibrium equations based on the deformed configuration are generated. Contact nonlinearity may also be considered a form of geometric nonlinearity. Most real problems involve contacting bodies; however, in many cases, this is circumvented in FEA by the use of simplifying boundary conditions. However, in cases where the interaction of two or more bodies is of interest, a contact analysis must be set up. It should be noted that if contact is not set up, the two contacting bodies will simply pass through each other. Numerous methods of incorporating contact in FEA are available but most involve two main steps. The first is the detection of contact, and the second is the imposition of forces to prevent interpenetration of the contacting bodies. In order to detect penetration, it is usual to first identify the surfaces that are likely to make contact. As a nonlinear analysis

25

Numerical Approach: Finite Element Analysis

715

proceeds, checks are then made to assess contact. Once contact has been made, the penetration distance is calculated and appropriate forces applied to prevent interpenetration. From this analysis, the point of contact, contact force, deformed shape of contacting bodies, and contact stresses can all be calculated. The results may be influenced by the mesh in the contact region, the nonlinear controls, and the contact algorithms used. Material nonlinearity may be hyperelastic or elastoplastic. The difference between the behavior of an elastic and elastoplastic material is seen on unloading as in the former case, the unloading path coincides with the loading path, whereas in the latter case, a different unloading path results in permanent deformation when the load has been removed. Elastoplastic behavior is characterized by a linear region up to the yield point, after which softening behavior is seen. Hyperelastic materials such as elastomers exhibit nonlinear elastic response for even large strains.

25.3.2 Solution Algorithms Numerical methods are incapable of solving nonlinear equations explicitly, and the actual behavior must be approximated by a sequence of linear steps. Incremental and iterative methods are available to solve a system of nonlinear equations. In the incremental method, the response is approximated by dividing the solution into a number of linear increments and updating the stiffness at each increment. The incremental method may underestimate the nonlinear behavior, and a progressive divergence from the actual response may be observed. A better approximation would be obtained by decreasing the size of the increments, but for a controlled reduction in error, an iterative method is required. Iterative methods utilize the secant stiffness, which is defined as F/u. One of the commonly used iterative techniques is the Newton-Raphson method. The first stage in this method is the same as the first increment in the incremental method. After the increment, as the stiffness as a function of displacement is known, the internal forces can be calculated from: r ¼ ku

(28)

where k is the secant stiffness at the end of increment. The difference between the applied force and the resisting force can then be calculated from: e¼Fr

(29)

This force imbalance can then be used to drive the displacement toward the correct value by using an equilibrium iteration method at constant load. The process can be repeated until a suitably small imbalance is attained. The Newton-Raphson method may be computationally expensive in a multidof problem because a new global stiffness matrix is used in each iterative step.

716

I. A. Ashcroft and A. Mubashar

In the modified Newton-Raphson method, the same global stiffness matrix is used in all the iterative steps within an increment. This method requires more iterations to achieve convergence, but each iteration is computed far more quickly. Appropriate selection of convergence criteria is essential to obtain sufficient accuracy in a reasonable time. Convergence is also based on the solution algorithm. Convergence failure is usually defined when the convergence tolerance is not reached in a specified number of iterations. In this case, it is common to reduce the incremental load step.

25.4

Thermal Analysis

25.4.1 Introduction Thermal analysis is primarily concerned with the calculation of temperature distribution in a body and the magnitude and direction of heat flow. There are many similarities between the formulation and solution of thermal and structural problems in FEA. The global FE equation for a steady-state thermal problem is ½K T ½T  ¼ ½Q

(30)

where [KT] is a matrix of material constants including thermal conductivity, [T] is a column matrix of nodal temperatures, and [Q] is a column matrix of thermal loads. Thermal finite elements are assembled in a similar way as structural finite elements. In contrast to stress analysis, however, thermal analysis is a scalar problem as temperature is nondirectional. Also, a thermal finite element node has only one degree of freedom, namely, temperature. In a thermal analysis, the temperature field (like the displacement field in structural analysis) is continuous, but the temperature gradients (like strains in structural analysis) are not continuous between elements, and higher mesh refinement is required in regions where temperature gradient changes rapidly.

25.4.2 Heat Conduction Fourier heat conduction equation for an isotropic material: f x ¼ k

@T @T or qx ¼ kA @x @x

(31)

where fx is heat flux per unit area (Wm2), k is the thermal conductivity (Wm1 K1), qx is the rate of heat flow (W), and A is the cross-sectional area (m2). The negative sign indicates that heat flows in the opposite direction to the increase in temperature. This can be extended to a general anisotropic case:

25

Numerical Approach: Finite Element Analysis

8 9 2 kx vm

(17) (18) (19)

It can be seen that the effective traction (te) is a function of the effective opening displacement (ve) and is reversible until the critical effective opening displacement (vc) has been reached. The other parameters that define the failure behavior are the area under the traction-separation curve, which corresponds to the strain energy release rate (GC) and the maximum effective opening displacement (vm). The values for the moisture-dependent cohesive zone law parameters were derived from experimental work done by Liljedahl et al. (2005, 2007) using mixed mode flexure (MMF) tests, the moisture-dependent mechanical properties of the adhesive, and calibrated values using the load displacement plots from aluminum/adhesive FM73 (Cytec Engineered Materials Ltd.) single-lap joints. Moisture-dependent mechanical properties were also used for the carbon fiber-reinforced polymer (CFRP) adherend, together with a moisture-dependent Tsai-Wu failure criterion, although failure in the CFRP did not occur in the models, in agreement with the experiments. The predicted effect of 50
C, 95% r.h. conditioning of the joint shown in Fig. 20 using the CZM discussed above is illustrated in Fig. 22. In Fig. 22a it can be seen that the failure load is predicted to decrease with conditioning time. A comparison of the predicted failure load with that measured experimentally can be seen in Fig. 22b. For exposure times up to 12 weeks, the CZM predictions show good agreement with the experimental failure loads. At 26 weeks, the failure load is underestimated and is within the large experimental error at 52 weeks. It is unknown whether the

31

Effect of Water and Mechanical Stress on Durability

909

a 25

7075/T6 Aluminium 4

IM7/8552

IM7/8552 Unidirectional CFRP

2

7075/ T6 Aluminium 87.5

12.5

87.5

b

Y X Z

Fig. 20 (a) Dimensions of CFRP/Al double lap joint (b) typical FE mesh (Jumbo 2007)

experimental increase in strength at 26 weeks and subsequent decrease at 52 weeks are “real” effects or experimental scatter. The latter is more likely; in which case the predictions are probably as good as can be expected. Liu et al. (2016) used a similar approach for modeling the adhesive joint strength degradation in double-lap shear joints with carbon/epoxy composite adherends. The bilinear traction-separation law was degraded for the moisture-conditioned adhesive joints. Different degradation factors were used for the adhesive layer and the adhesive-adherend interface, where the interfacial layer was degraded using an extra factor. The degradation factors were determined using Eq. 20.:

910

I. A. Ashcroft et al.

te

Fig. 21 Bilinear cohesive zone model

Gc

nm ne

nc

a

b

25,000 2D:Dry 2D:26 weeks

24

2D:52 weeks

22

15,000

Load, kN

Load, N

20,000

26

10,000

Experimental FE CZM Prediction

20 18 16 14

5,000

12 0

10 0

0.1

0.2 0.3 Displacement, mm

0.4

0

10

20 30 40 Exposure time, weeks

50

60

Fig. 22 Effect of aging at 50
C, 95% r.h. on (a) the load-displacement and (b) the failure load of a CFRP/Al/epoxy DLJ using CZM (Jumbo 2007)

rffiffiffiffiffiffiffiffiffi σH f TH ¼ T0 σ

(20)

where TH and T0 are critical tractions with and without moisture absorption and σ H and σ are ultimate tensile strengths of the adhesive with and without moisture absorption. f is the degrading factor used for the interface. Jumbo (2007) also used the Gurson-Tvergaard-Needleman (GTN) CDM to predict failure in the same joints (see ▶ Chap. 23, “Constitutive Adhesive and Sealant Models”). This model was proposed by Gurson (1977) to model damage in elastic-plastic materials incorporating void formation, growth, and coalescence, leading to crack formation and failure, and was later extended by Needleman and Tvergaard (1984). This model is potentially applicable to damage prediction in rubber-toughened polymers, as the rubber particles are sites for cavity nucleation. The yield condition is defined as
2
 h i σ q2 σ kk  F¼ þ 2q1 f cosh  1 þ ð q1 f  Þ 2 ¼ 0 σy 2σ y

(21)

31

Effect of Water and Mechanical Stress on Durability

911

where σ is the equivalent stress, σ y is the yield stress of the matrix material, and σ kk is the hydrostatic stress. The parameters q1 and q2 were introduced by Tvergaard to account for the effect of void interactions on the stress distribution between cavities but are usually derived from finite element calibration of experimental curves. Damage evolution is measured by the void volume fraction f which increases due to void nucleation and growth. Void nucleation is controlled by a normal distribution of stress or plastic strain, and the modified model replaces the void volume fraction f with the modified void volume fraction f*. A rapid decrease in the load-carrying capacity of the material if void coalescence occurs is linked to f*, such that f ¼ f

if f  f c

(22)

where f ¼ fc þ




 f u  f c ðf  f c Þ fF  fc

(23)

where fc is the critical void volume fraction and fF is the void volume at failure, and for most metals, f u ¼ 1=q1 . The existing value of the void volume fraction changes due to the growth of existing voids and the nucleation of new voids and is characterized by f_ ¼ f_growth þ f_nucleation

(24)

The growth of voids is given by p f_growth ¼ ð1  f Þe_kk

(25)

For modeling purposes, plastic strain-controlled nucleation of void has been used and is given by "
 # fN 1 epm  en 2 p _f nucleation ¼ p ffiffiffiffiffi exp  e_m 2 S S 2π

(26)

where fN is the void volume fraction of forming particles, en is the mean strain at which void nucleationP occurs, S is the standard deviation, σ n is the mean stress for void nucleation, and ¼ σ þ 13 σ kk. When the material reaches 90% of the void volume fraction at failure, fF, the material is considered to have failed, with the stiffness and of the failed element reduced to zero. The property requirements for the model are • Young’s modulus and Poisson’s ratio  • Moisture-dependent effective shear-hardening curves σ o epo . The initial, critical, and failure volume fractions ( f, fc, and fF)

912

I. A. Ashcroft et al.

• The void interaction parameters, q1 and q2 • The volume fraction of nucleating particles fN • The standard deviation and mean strain for void nucleation (S and en) The direct experimental determination of some of these constants is difficult, leading to their determination by an iterative process using FE analysis. Hence, in Jumbo (2007), some of the model parameters were determined from experimental data and others by calibrations from FE analysis of a representative joint. The results are shown in Fig. 23. It can be seen that the 2D model slightly overpredicts the failure load but that a good agreement is seen with the experimental data with the 3D model. Comparing Figs. 22b and 23, it can be seen that the 3D Gurson model agrees well with the CZM.

31.5

Summary and Conclusions

It is often seen that the strength of adhesive joints decreases on aging in natural environments. This is usually because of the effect of environmental moisture, which is absorbed by polymeric adhesives and affects both the mechanical behavior of the adhesive and the interface between the adhesive and adherend. These effects are usually increased by increasing temperature or the application of stress. The effect on the interface is potentially the most damaging; however, this can usually be mitigated by the application of a suitable surface treatment prior to bonding. In this case it is often seen that the effects of absorbed moisture are largely reversible. The effect of environment on adhesive joints is commonly investigated experimentally through accelerated aging tests in which joints are exposed to high temperature and humidity. This can provide a useful comparison of different systems but cannot be used to predict the performance of joints under actual service conditions. Currently, the best method of predicting the effect of environmental aging on the strength of bonded joints is through coupled hygro-mechanical finite element 30 25 Failure load, kN

Fig. 23 Effect of aging at 50
C, 95% r.h. on the failure load of a CFRP/Al/epoxy DLJ predicted using the Gurson CDM (Jumbo 2007)

20 15 Experimental

10

2D Gurson 3D Gurson

5 0 0

10

20 30 40 Exposure time, weeks

50

60

31

Effect of Water and Mechanical Stress on Durability

913

analysis. The first stage in this method is to determine the moisture distribution in the joint as a function of time from a diffusion analysis. This is used to determine the spatial distribution of mechanical properties and the applied hygroscopic strain. A mechanical analysis is then used to determine the stresses in the joint as a function of aging time, and an appropriate failure criterion can be used to predict the residual strength of the joint as a function of aging time. This method is now well developed and has been validated for simple joints; however, further development and validation are required. It would also be useful to develop a software tool based on this method for the nonspecialist designer.

31.6

Further Information

The most comprehensive treatment of wet durability is Kinloch (1983). In this, Minford (1983) has compared wet durabilities of joints with phenolic and epoxidebased adhesives. Joints in aluminum have been reviewed by Brewis (1983), and Mahoon (1983) has dealt with titanium and Brockmann (1983) with steel. Reviews which have provided updates are those by Critchlow and Brewis (1995, 1996) which, respectively, deal with the surface treatment of titanium and aluminum and that by Armstrong (1997). A comprehensive comparison of accelerated aging and natural aging has been made by Ashcroft et al. (2001). Modeling the environmentally induced degradation of adhesively bonded joints has been reviewed by Crocombe et al. (2008). The effects of temperature and moisture on adhesives and adhesive joints have recently been reviewed by Viana et al. (2016b).

References Abdel Wahab MM, Ashcroft IA et al (2001) J Adhes 77:43 Armstrong KB (1997) Int J Adhes Adhes 17:89 Ashcroft IA, Digby RP et al (2001) J Adhes 75:175 Barbosa AQ, da Silva LFM et al (2015) J Adhes Sci Technol 29:1714 Berens AR, Hopfenberg HB (1978) Polymer 19:489 Brewis DM (1983) In: Kinloch AJ (ed) Durability of structural adhesives. Applied Science Publishers, London, p 215. Chapter 5 Brewis DM, Comyn J et al (1980a) Polymer 21:1477 Brewis DM, Comyn J et al (1980b) Int J Adhes Adhes 1:35 Brewis DM, Comyn J et al (1980c) Polymer 21:134 Brewis DM, Comyn J et al (1981) Poly Engg Sci 21:797 Brewis DM, Comyn J et al (1987) Int J Adhes Adhes 7:30 Brewis DM, Comyn J et al (1990) Int J Adhes Adhes 10:247 Briskham P, Smith G (2000) Int J Adhes Adhes 20:33 Brockmann W (1983) In: Kinloch AJ (ed) Durability of structural adhesives. Applied Science Publishers, London, p 281. Chapter 7 Butt RI, Cotter JL (1976) J Adhes 8:11 Carter FG, Kibler KG (1978) J Compos Mater 12:118 Comyn J (1983) In: Kinloch AJ (ed) Durability of structural adhesives. Applied Science Publishers, London, p 85. Chapter 3

914

I. A. Ashcroft et al.

Comyn J (2005) In: Adams RD (ed) Adhesive bonding: science, technology and applications. Woodhead Publishing Ltd., Cambridge, UK, p 123. Chapter 6 Comyn J, Brewis DM et al (1987) J Adhes 21:59 Crank J (1975) The mathematics of diffusion. Oxford Science Publications, Oxford Critchlow GW, Brewis DM (1995) Int J Adhes Adhes 15:161 Critchlow GW, Brewis DM (1996) Int J Adhes Adhes 16:255 Critchlow GW, Ashcroft IA et al (2006) Anodising aluminium alloy, United Kingdom: Patent No GB 3421959A Crocombe AD, Ashcroft IA et al (2008) In: da Silva LFM, Ochsner A (eds) Modelling of adhesively bonded joints. Springer, Berlin, p 225. Chapter 8 Davis RE, Fay PA (1993) Int J Adhes Adhes 13:97 Fay PA, Maddison A (1990) Int J Adhes Adhes 10:179 Fernandes RL, de Moura MFSF et al (2016) Int J Adhes Adhes 68:30 Gledhill RA, Kinloch AJ (1974) J Adhes 6:315 Gledhill RA, Kinloch AJ et al (1980) J Adhes 11:3 Gurson AL (1977) J Eng Mater Technol. Trans ASME 99:2 Han X, Crocombe AD et al (2014) J Adhes 90:420 Heshmati M, Haghani R et al (2016) Compos Part B 92:447 Hua Y, Crocombe AD et al (2007) J Adhes Sci Technol 21:179 Jumbo FS (2007) Modelling residual stresses and environmental degradation in adhesively bonded joints. PhD thesis, Loughborough University, Loughborough Kinloch AJ (ed) (1983) Durability of structural adhesives. Applied Science Publishers, London Korta J, Mlyniec A et al (2015) Compos Part B 79:621 Krishnan P, Abdul Majid MS et al (2016) Compos Struct 148:1 Liljedahl CDM, Crocombe AD et al (2005) J Adhes Sci Technol 19:525 Liljedahl CDM, Crocombe AD et al (2007) Int J Adhes Adhes 27:505 Liu S, Cheng X et al (2016) Compos Part B 91:431 Loh WK, Crocombe AD et al (2003) J Adhes 79:1135 Loh WK, Crocombe AD et al (2005) Int J Adhes Adhes 25:1 Mahoon A (1983) In: Kinloch AJ (ed) Durability of structural adhesives. Applied Science Publishers, London, p 255. Chapter 6 Minford JD (1983) In: Kinloch AJ (ed) Durability of structural adhesives. Applied Science Publishers, London, p 135. Chapter 4 Mubashar A, Ashcroft IA et al (2009a) J Adhes 85:711 Mubashar A, Ashcroft IA et al (2009b) Int J Adhes Adhes 29:751 Mubashar A, Ashcroft IA et al (2011) Eng Fract Mech 78:2746 Needleman A, Tvergaard V (1984) J Mech Phys Solids 32:461 Parker BM (1988) J Adhes 26:131 Sadigh S, Ali M (2016) J Fail Anal Prev 16:1134 Viana G, Costa M et al (2016a) J Adhes 93:95 Viana G, Costa M et al (2016b) Proc I Mech E Part L 0:1

Adhesives in Space Environment

32

Sabine Dagras, Julien Eck, Claire Tonon, and Denis Lavielle

Contents 32.1 32.2 32.3 32.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesives Used in Space Programs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Environment Specification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.1 Adhesives Testing Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.4.2 Test Facilities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.5 Impact of Space Environment on Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.5.1 Degradation Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.5.2 Irradiation Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

916 917 920 923 923 924 929 929 932 938 939

Abstract

Adhesives are widely used on a spacecraft for several reasons. They allow a significant mass reduction and, using an adhesive allows. The combination of several properties (as the mechanical fixation and some electrical function). One of their main drawbacks is that they are organics and thus they react with space environment. Among all the components of the space environment which could lead to material degradations, the three most important are the energetic charged particles (protons and electrons particles with different fluxes and energies), the electromagnetic radiation from the direct solar flux (and most specifically in the Ultra Violet wave length) and the Atomic Oxygen (ATOX) coming from the photodissociation of the oxygen molecules of the atmosphere by the solar S. Dagras (*) · J. Eck · C. Tonon · D. Lavielle Airbus Defence and Space, Toulouse, France e-mail: [email protected]; [email protected]; [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_32

915

916

S. Dagras et al.

electromagnetic radiation. These factors lead to physicochemical degradation of the adhesives, which can affect their functional properties. All the adhesives used in spacecraft have to be validated on ground before launch and their behavior in space environment must be evaluated. Test facilities are developed to simulate charged particles, UV, and ATOX constraints. Even if it is not possible to be 100% representative of the in-orbit conditions, tests, methodologies and specifications exist to correctly evaluate the degradation and to take into account the adhesives end of life properties, in the spacecraft design. Atoms and molecules along the charged particles trajectory can be either excited or ionized. The primary effects are generally deformation, embrittlement, and discoloration, which impact the mechanical integrity of the adhesive. The modification of mechanical properties is generally the result of the competition between chain scissions and crosslinking in the macromolecular chains. The main UV effect on adhesives is generally a degradation of their optical properties (e.g., solar cell cover glass). Indeed, during the interaction of a photon with an atom or a molecule of the material, the photon can either be absorbed or diffused with possible loss of energy. These photochemical effects may result in a change of color of the material, leading to an increase of solar absorptance. Atomic Oxygen might also have detrimental effects on adhesives, mainly erosion of surfaces and oxidation.

32.1

Introduction

Spacecrafts and space systems in general significantly evolved till the first systems launch into orbit in the years 1960. The missions evolved and it is now required to have more and more power and function accessible for the telecommunication spacecraft as for the earth observation or scientific missions. These missions’ new developments and needs lead to a significant increase of the spacecraft sizes and masses. Due to launchers constraints, it is now require optimizing as much as possible the satellites design and all the parts and equipments are reduced and accommodated to limit their mass. Taking into account these constraints, it is not possible any more to have spacecraft mainly composed of metallic materials which strongly behave the specific space environment. Polymers and composite materials uses in space systems were thus more and more developed and current satellites holds various organics materials for various applications. Among all these organics materials, the adhesives are widely used including for external applications on the satellites and can thus be fully exposed to space environment. In parallel of this mass reduction objective, space missions’ durations are extended and satellites shall now withstand more than 15 years in orbit without degradation of their functionality. The reliability of all the satellites parts have to be demonstrated on ground before launch. Among all the parameters to be evaluated, the behavior of organic materials (including adhesives) to space environment is a driver in the spacecraft design. Indeed, the materials in orbit are submitted to various

32

Adhesives in Space Environment

917

Fig. 1 Illustration of the satellites evolution. First spacecraft in orbit, Sputnik (left) and a recent telecommunication spacecraft (right)

constraints such as radiations of charged particles, ultraviolet, Atomic Oxygen. . . (Fig. 1). The following chapter address the behavior of adhesives in space environment and is thus composed as follows. In the first part of this chapter, a description of the typical uses and needs related to adhesives on spacecraft is given. Then, the specificities of the space environment are explained. Once in orbit, the adhesives have to withstand constraints as particles radiations, UV, and Atomic Oxygen (ATOX) that are not present on Earth. Depending on the mission, the location and the adhesive use on the spacecraft, these constraints can vary a lot. Before spacecraft launch, the adhesives have to be validated and tested on ground with conditions as representative as possible of the space environment. In this chapter, the testing methodologies are detailed. Finally, the impact of this space environment on adhesives and the associated degradations are developed. The degradation mechanisms and the effects of the irradiation on the adhesive macroscopic properties are presented.

32.2

Adhesives Used in Space Programs

Adhesives are widely used in space applications for many different applications. This material family combines several properties which are really important for space programs: – They allow durable bonding of dissimilar materials (e.g., polymer/metal, composite/metal, etc.) – They allow significant mass reduction for mechanical fixation compared to traditional screws or rivets. This mass reduction is generally a driver for a domain where each gram avoided is important. – Some adhesive families allow a use in an extended temperature range, which can be of high interest for external areas of the spacecraft. On spacecraft solar

918

S. Dagras et al.

Fig. 2 Adhesives used for space applications should have low-outgassing levels. They shall withstand space environment harsh constraints: extreme temperatures, high vacuum, UV and radiation environment, and/or Atomic Oxygen

arrays, temperature excursions can reach
180  C/+180  C as illustrated in Fig. 2. – The combination of several properties (mechanical fixation and electrical grounding as an example) is finally an advantage for spacecraft’s design optimization.

The main use of adhesives on a space instrument or a spacecraft is the mechanical fixation. Adhesives shall withstand important loads due to the spacecraft launch (vibration and acoustics constraints). Once in-orbit, the most significant mechanical loads are shocks due to deployment of subassemblies (e.g., solar arrays, antennas). Another critical parameter is the large variation of materials thermal expansions which can be observed due to the important temperature ranges (typically between
180  C/+180  C for external appendices). This might induce thermomechanical deformations. Therefore, impact of space environment on mechanical properties is mandatory to assess the acceptability for the use of adhesives on spacecraft structures. The mechanical design of structural adhesive joints is driven by mechanical loads during on-ground tests and launch. The ageing during operational in-orbit lifetime is driven by thermomechanical loads, as no vibration loads comparable to the launch loads exist. In addition to its mechanical function, an adhesive could also be used for its electrical or thermal properties. Indeed, adhesives loaded with conductive fillers allow the grounding of most of the spacecraft parts. On the contrary, adhesives could

32

Adhesives in Space Environment

919

Fig. 3 Examples of adhesive use on spacecrafts: bonded stand-off for MLI (multilayers insulation) blankets mechanical installation on spacecraft structure (left), optical surface reflectors bonded on spacecraft structure for thermal control purpose (middle), electronic PCB proceeds through “touch up” after completing frame bonding, which uses a new automated adhesive dispensing process (NASA.gov image) (right)

be also be used as varnish or conformal coating for electrical insulation purposes (see Fig. 3). When applied for the bonding of thermal active parts, as heatpipes, thermal conductivity of the adhesive is a driver to guarantee the thermal efficiency of the component onto the structure. The same advantage is foreseen in electronic assembly, as the ability of thermally conductive adhesives helps dissipating heat as more and more power dissipating components are mounted on PCB. Finally, the optical properties of some adhesives are used on space systems. The most relevant example consists in transparent adhesives used for the bonding of the solar cells cover glasses. The evolution of the transmission properties of the adhesive (yellowing due to interaction between adhesive and radiations and UV) is a driver in the solar arrays design and sizing. Even if adhesives are advantageous for some applications, they also present drawbacks for space applications. As they are organic products, their properties are modified during the spacecraft in-orbit life, due to exposure to space environment (and more specifically radiation and UV). Moreover, the adhesives are submitted to vacuum once in orbit and their volatile components outgas. This can be a major constraint for missions where the outgassed components could lead to the molecular contamination of the instruments optics, degrade their optical properties, and reduce the instruments performance. Therefore, the validation on ground of adhesive behavior to such constraints is performed through dedicated and specific test campaigns. Several adhesive chemical families are used in spaces programs. Among all of them, the most used are the epoxies for their mechanical properties, silicones for their good behavior to thermal environment, acrylics (generally used as tapes), and polyurethanes. As space industry represents a relatively reduced commercial amount compared to other industrial sectors, it is not possible to develop specific adhesives for each application. Most of the adhesives used on a spacecraft are therefore commercial products nevertheless, for some really specific applications, developments are performed to cover a dedicated mission and the adhesive formulation can be adapted to be more resistant to space environment.

920

32.3

S. Dagras et al.

Environment Specification

The space environment represents a harsh medium for all spacecraft. Proper assessment of the potential effects is an essential part of the engineering process leading to the qualification of any element of the spacecraft (Space weather effect catalogue). As defined in ECSS Standard (ECSS-Q-ST-10-04C), the natural space environment of a given item corresponds of environmental conditions defined by the external physical world for the given mission. The main components of the space environment that can lead to material degradations are: – – – – – –

Meteoroids and debris Galactic cosmic rays (GCR) Solar energetic particles (SEP) Photons, from the γ rays to the radio frequency The solar wind Atomic Oxygen

Whereas debris and micrometeorites can induce instantaneously damages in the materials, the remaining constraints can be assimilated as ageing factors. In that latter cases, the cumulative effect of these constraints lead to progressive degradation of the materials all along the mission duration. In this chapter, we will focus on the effects of UV, particle radiations, and Atomic Oxygen (ATOX). A spacecraft receives electromagnetic radiation originating from the direct solar flux. The solar spectrum can be divided as follows in several parts depending on the wavelength range considered: – – – – –

Soft X-rays or XUV (0.1 nm–10 nm) Extreme ultraviolet or EUV (10 nm–121 nm) Ultraviolet or UV (100 nm–400 nm) Visible or VIS (380 nm–760 nm) Infrared or IR (0.70 μm–1 mm)

In general, for material testing, the spectrum used for solar simulation is limited to the UV region because, although this band is responsible for only 1% of the total solar irradiance, it is assumed that the major degradation of the materials is due to these photons. The second type of radiation leading to potential degradation of the materials is particle radiation. Energetic charged particles are encountered in interplanetary space and in the magnetospheres of moons or planets. There are different types of particles, coming from different origins. The main ones are: – The cosmic rays, entering the solar system from the interstellar medium and composed of protons, electrons, and fully ionized nuclei. Their energy remains in the GeV to TeV range.

32

Adhesives in Space Environment

921

– The solar energetic particles, originating from Sun flaring or from shock waves associated with coronal mass ejections, consist of protons, electrons, and heavy ions with energies from a few tens of keV to GeV ranges. – The solar wind, which is part of the Sun’s outer atmosphere expanding outwards and composed mainly of protons, electrons, alphas with an energy range between few eV and few keV.

solar wind

e magnetosheath

bo w

Fig. 4 Radiation belts illustration (https://www. britannica.com/science/VanAllen-radiation-belt)

sh oc kw av

Although the energy of the cosmic rays and the solar energetic particles are much higher than the ones of the solar wind or of the particles within the radiation belts, their flux are relatively low and thus their contribution to the material degradation is negligible. Conversely, this type of radiation (cosmic rays and solar energetic particles) are a major concern for electrical components as it can induce so called single event that could lead to their dysfunction or their destruction. Geomagnetic fields of the planets or the moons influence the energetic particles and can also induce specific radiation environments. The particles coming from the interplanetary medium are trapped in the magnetic field and then accelerated along the field lines depending on their polarization. Thus, radiations belts are created around the planets or moons. Around the Earth, these radiation belts are also called Van Allen belts. They are composed of an inner and an outer belt and extend from typically 100 up to 65,000 km. These belts are mainly constituted of protons (with energies up to hundreds of MeV) and electrons (energies up to few MeV) coming from the solar wind. As illustrated in Figs. 4 and 5, depending on its orbit (Low-Earth Orbit, MediumEarth Orbit, Geostationary-Earth Orbit), the satellite will be submitted to different particles flux and energies with the Van Allen belts.

use topa gne a m magnetosphere

Van Allen radiation belts

neutral sheet

Earth

ck ho ws bo

magnetosphere

v wa

e

solar wind

ma

g ne

magnetosheath

topau se

922

S. Dagras et al.

Fig. 5 Space radiations belts for the satellites different orbits (https://www.nasa.gov/mission_ pages/sunearth/news/gallery/20130228-radiationbelts.html)

Whereas the LEO missions, characterized by a spacecraft altitude below typically 1000 km, are located below the inner belt and are submitted to relatively low particle radiations, GPS or GEO missions are inside the outer belt and submitted to higher particle fluxes, leading to significant dose levels compared to LEO ones. Additionally, in the outer belt, the fluxes of low-energy particles are higher than in the inner belt, leading to an increase of the deposited dose within the external materials, directly exposed to space environment. In the case of interplanetary missions, specific radiation environments shall also be considered. As an example, in the case of Jovian missions, during the cruising phase between the planets, the spacecraft will be mostly submitted to solar wind but also to radiations belts of Jupiter or its moons (Ganymede, Europa, etc.), characterized by important fluxes of highly energetic electrons. Missions towards the Sun are also exposed to very specific environments. The probes are exposed to very high flux of solar wind particles together with high temperatures, leading to potential synergistic effects. Another component of the space environment able to induce degradation of the materials is Atomic Oxygen. These atoms are generated during the photodissociation of the oxygen molecules of the atmosphere by the solar electromagnetic radiation (mainly short wavelengths, λ < 243 nm), leading to the breaking of the diatomic bonds (NASA-HDBK-6024). These atoms are very reactive and can interact with the spacecraft materials through oxidation mechanisms.

32

Adhesives in Space Environment

923

Atomic Oxygen is present around Earth but can also be produced in other planetary environments, such as the Mars orbital environment, where oxygen is present.

32.4

Testing

32.4.1 Adhesives Testing Methodology As a general rule, any technology used on a spacecraft has to be qualified with regards to functional and environmental constraints. For adhesives, this rule is also applicable. Therefore, for each adhesive application, a test plan is set up to finalize the different functional and environmental stresses that need to be verified on ground to validate further space application. The main applications for adhesives in space are either structural (up to a certain extent for general purpose adhesives) or electrical for grounding adhesives or optical (to bond glasses or transparent materials) as already mentioned in ▶ Chap. 4, “Wetting of Solids.” Due to the stringent space environment, many environmental stresses are to be considered both before launch (long-term storage) and after launch for the mission (vacuum, radiation, temperature cycling, etc.). All these stresses are mission dependent and are analyzed on a case by case basis even if general guidelines are driven by ECSS standards (ECSS-Q-ST-70-06C). The adhesives qualifications generally consist in two levels: – A test campaign on samples. Samples are processed in order to evaluate the adhesive properties evolution before and after accelerated ageing. – An overall validation at equipment and system level. Specific tests campaign could be performed on new development on spacecraft to qualify the bonding against vibrations and mechanical constraints representative of the launch campaign. (e.g., vibration, shock, etc.). This addition of some other stresses can reveal any potential weak points in adhesive use. In all cases, functional studies (mechanical, thermal, electrical, etc.) are performed to determine the limits of the tests to be performed on test samples to validate the mission profile. However, in terms of risk mitigation, the major part of the qualification is performed at sample level. The general principle of the test campaign on adhesive samples is described in Fig. 6. The test flow-chart is composed of different branches: – One set of reference samples (not experiencing any stress) – Several sets of stressed samples that combine conditions to simulate the mission profile

924

S. Dagras et al.

Fig. 6 Test campaign typical flow chart at adhesive samples level

A typical example of such a test program is represented in Fig. 6 where the central branch simulates storage conditions and the right branch cumulated storage and mission profile. At the end of the program, both reference samples and stressed samples are functionally tested and final values are compared to reference ones. The characterization tests are defined to be representative of the relevant properties. Mechanical tests as shear test or peeling can be performed if the adhesive is used for parts fixation. Electrical or optical measurements could also be done. In function of the need, functional tests could be performed which aim to validate the macroscopic use of the adhesive. More detailed characterizations tests could be performed to understand the degradation phenomena which occurred in the material (DSC, RMN, etc.). Generally, the test plan includes a success criteria based on an acceptable functional change before and after submission to environmental tests (typically 20% decrease for a mechanical strength decrease). Among all the constraints seen by an adhesive on a spacecraft, the specific behavior to radiations, ATOX and UV has to be analyzed and validated by tests on ground.

32.4.2 Test Facilities Radiations For the specific radiation validation, as mentioned in Sect. 5, the best approach is to simulate as close as possible on ground the expected radiation mission constraint. For particle radiation, particle accelerators (protons, electrons) are generally used to irradiate the samples. Parameters like particles, flux, vacuum conditions (to remove potential oxygen related bleaching) are key parameters to consider.

32

Adhesives in Space Environment

925

Fig. 7 Example of calculated real total ionizing dose (TID) depth curve and ground irradiation dose profiles

Also, when characterizing materials in space, it is important to simulate the total ionizing dose profile expected in the material. This dose is expressed in Grays or in rads (1 Gray = 100 rads): 1 Gray corresponds to the absorption of 1 Joule per kg of matter. As explained in Sect. 5, in space, the resulting dose profile is due to a combination of different particles (nature, energy, and flux). It is not possible to reproduce on ground the exact combination but an appropriate choice of several particles with dedicated energies allows approaching the flight conditions as illustrated in Fig. 7. In low-earth orbit missions, the spacecraft external dose is generally quite low, in the range of 5  105 Gy on surface materials (typical order of magnitude). However, for geostationary missions, surface dose as high as 1  109 Gy can be obtained and dose in the range of a few MGy can be achieved below a small shielding. In adhesives directly exposed to external environment, the dose level decreases quickly when penetrating within the materials. The use of Co60 (Cobalt 60 – gamma rays) exposure leads (due to gamma photons/matter interaction) to flat dose profile energy deposition. In that case, as the dose deposited within the material thickness is homogeneous, it is not possible to reproduce a dose profile and a dose level is applied, leading to an under/overexposure of the tested sample depending on the set parameters. For bulk materials (and low radiation levels), Co60 exposure might be used provided it is representative of the expected in orbit conditions (heavily shielded material with typical aluminum shielding thickness higher than 5 mm, and dose contribution mainly related to Bremsstrahlung effect). To simulate the dose profile on ground, particles accelerators like Van de Graaff are generally used with energies lying from a few tens of keV to a few MeV for

926

S. Dagras et al.

protons and few hundreds of keV to a few MeV (typically 3 to 5) for electrons. Ion implanters can also be used for low energy protons (up to a few hundred keV). Generally, homogeneous irradiation surfaces are of the order of 50 mm  50 mm up to 200 mm  200 mm. Fluxes used in the radiation tests are limited to a typical range of 11012 particles/ cm2/s. When the flux is too high, a temperature increase can occur on the samples and change drastically the radiation behavior with regards to future mission conditions. This phenomenon is particularly true for organic materials as this temperature increase lead to the overrun of the glass transition. Generally, fluxes in orbit are a few orders of magnitude lower and radiation exposure last up to 15 years instead of a few hours or days for on-ground simulation. Therefore, fluxes shall be a compromise between representativeness and test duration. On some accelerators, secondary vacuum conditions are achievable or use of flowing Nitrogen gas is also acceptable (leading to absence of oxygen). For low energy protons (up to a few hundred keV), irradiations under secondary vacuum conditions are mandatory to avoid any interaction of particles with air molecules.

UV In general, the spectrum used for solar simulation is limited to the UV region, because it is assumed that the major degradation is due to these photons (ECSS-QST-70-06C)]. For UV irradiation, one constraint is to simulate as close as possible the external atmosphere solar spectrum. Indeed, for organic materials like adhesives, specific wavelength, specific to extra-atmospheric solar spectrum could have a specific effect on materials degradation (low wavelength and high energy). Xe-based lamp can generally be used in this case. The xenon lamp spectrum differs from all solar spectra because of the intense line output in the 800–1100 nm region. The use of specific filters allows the reduction of the mismatch and provides the closest spectral match to solar spectra. Another constraint is to limit the acceleration factor. If not controlled, temperature as high as 200  C can be reached quickly (especially under vacuum) at samples level and lead to thermal degradation not representative of real in orbit degradation. Finally, vacuum conditions during exposure are important to make sure that no specific gas absorption will occur (decreasing the spectrum part at low end) and allowing oxygen residues interaction with material surface. To do a space simulation in a reasonable time, it is required to accelerate the irradiations as fast as possible without modifying the degradation mechanisms which could induce wrong results. This problem is that of the reciprocity principle of equivalent sun hours (ESH). When using a nonfiltered Xenon light source, an acceleration factor limitation to 3 is recommended (ASTM E512) in order to limit the thermal charge on the materials.

32

Adhesives in Space Environment

927

In order to apply a higher acceleration factor, Xenon sources with specific filters can be used. These filters allow the removal of the visible and infrared part of the spectrum (see Fig. 8). Thus samples are only exposed to UV radiations leading to a thermal charge reduction by a factor of 10 (see Fig. 9).

Energy

UV Simulation Thekaekara

200

300 400 Wavelength (nm)

Fig. 8 Example of one UV simulation compared to Thekaekara spectrum

IRRADIANCE (Wm−2nm−1) (ASTM CURVE ONLY)

4 ASTM E490 3 ORIEL SOLAR SIMULATOR

2

1

0 250

750

1250 WAVELENGTH (nm)

1750

Fig. 9 Typical output spectral distribution of AM 0 Simulators normalized to the ASTM E490 standard spectrum by matching total power density from 250–2500 nm (https://www.newport.com/ n/solar-simulator-spectral-irradiance-data)

928

S. Dagras et al.

Degradation Estimation The solar constant is defined in (ECSS-Q-ST-10-04C) as the radiation that falls on a unit area of surface normal to the line from the Sun, per unit time, outside the atmosphere, at one astronomical unit (1 AU = average Earth-Sun distance). In the case of ultraviolet radiations, it is estimated that the degradation is a function of the solar illumination in the UV (wavelengths typically between 200 and 400 nm), expressed in W.m
2. This illumination is also usually expressed in space technology by the number of equivalent sun hours (ESH), which, for GEO environment, is equal to the UV part of the solar spectrum integrated over 1 h: 1 ESH ¼ 108  3600 ¼ 3:888:105 J:m
2

(1)

where 108 W.m
2 represent the integrated flux on the UV range of the solar spectrum. Equation 1: Definition of an ESH (Equivalent Sun Hours) As an example, over 1 year (8760 h), GEO satellite materials, depending on their location on the spacecraft, can be exposed up to 2500 ESH per year. Note that simulation tests in laboratory specify that the tested material will have to be irradiated with the same number of UV ESH and the same spectral distribution as the material in the specified mission conditions. It has to be noticed that the solar spectral distribution is very difficult to completely reproduce in laboratory. The difference can conduct to bad interpretation of the importance and/or nature of the degradation.

ATOX The Atomic Oxygen Simulators that are used around the world, depending on the energy of the Atomic Oxygen fluxes, can be divided into thermal and hyperthermal. The basic physical processes used for those sources are supersonic expansion, laser detonation, ion neutralization, electron-simulated desorption, and photo-dissociation (Kleiman et al. 2003). One of the most representative technologies for materials testing applications was shown to be the laser-sustained discharge sources, where lasers are used to produce a high-temperature plasma that is subsequently expanded in a free jet or supersonic nozzle to produce a high-velocity neutral beam. Such sources have been demonstrated to produce beams of the desired velocity of 8 km.s-1 at flux levels of 1017–1018 cm.s
2,i.e., exhibited both the appropriate energy for LEO simulation and high flux required for accelerated testing. The cost and limited availability of materials test time in these high-quality LEO environment simulators has generated considerable interest in the use of thermal energy oxygen atoms produced by plasma ashers for materials testing (Banks et al. 2013). This type of facility can be operated easily at low cost. However, the main drawback is the lack of representativeness with major differences in terms of variations in species, energies, thermal and radiation exposures, all of which may result in different reactions and erosion rates.

32

Adhesives in Space Environment

929

However, due to the variety and complexity of the differences between the LEO Atomic Oxygen environment and ground laboratory simulations, differences exist which have not allowed the development of a simple relationship or proportionality between the LEO and ground laboratory erosion yields. Two of the most significant differences between most ground laboratory and LEO environments are the energy and directionality of the Atomic Oxygen.

Specific Test Facilities Some laboratories have developed for decades a specific test facility that combines particle irradiation (electrons/protons) and UV irradiation under secondary vacuum. In addition, in-situ measurements like thermo-optical measurements can be performed before re-exposure of the samples to ambient atmosphere. Typical exposure areas are here again around 50 mm  50 mm up to 200 mm  200 mm. Cooling systems (coupled to sample back surface to limit temperature increase) are used for sample temperature regulation. Acceleration factors in the range of 10–50 are generally proposed and exposure of a few years in orbit can be simulated in a few days or weeks. In terms of acceleration, it is important to note that the acceleration factors for UV and particle irradiation, even if tested in the same facility, are about different values. However, the same rule applies to not overstress the samples by a too high acceleration factor leading to too high temperature increase of the samples that would generate appearance of nonrepresentative ageing mechanisms.

32.5

Impact of Space Environment on Adhesives

32.5.1 Degradation Mechanisms In the case of polymers, radiation effects have been extensively studied (Makhlis 1975; Paillous 1993). The difference must be done between charged particles (electrons, protons, heavy ions), which energy (from 1 kev to several MeV) is sufficient to excite and ionize molecules and UV photons but is insufficient to ionize them (excepted for far UV). Indeed, ionization occurs when the incident energy of the particle is sufficient to overcome the binding energy of electrons in atoms or molecules, thus creating ions. On another hand, excitation occurs when the incident energy is not sufficient to ionize molecules but can create excited intermediate electronic states (triplets, singulets, color centers), by the excitation of the molecule and the transfer of an electron from a fundamental state to an excited state. This can be summarized for a molecule AB by: – Excitation AB ! AB* – Ionization AB ! AB+ + e-

930

S. Dagras et al.

We present here both effects of ultraviolet and charged particles, as these components of space environment contribute to the degradation of adhesives. The effect of Atomic Oxygen (ATOX) are also presented. An important point to be considered is the thickness of the degraded zone for each parameter. In the case of ultraviolet radiations, the effect depends on the absorption of the material: – For low absorbent materials in the considered wavelengths, the degradation can apply at high thickness, – For high absorbent materials, the damage will be limited to more superficial effects. The difficulty to evaluate the degraded zone is that the degradation generally appears as a change of the absorption properties of the material, the material becoming more absorbent. The consequence is a limitation of the degraded thickness as the irradiation persists. In the case of charged particles, electrons and protons have to be distinguished, as their stopping power in matter is different. Depth penetration in matter, and thus degraded thickness, is far more important for electrons than for protons: a few microns for protons to compare with several hundred of microns to millimeters for electrons. In this section, the degradation of the polymer matrices due to radiations is mainly considered. Nevertheless, as a function of the specific use of the adhesive, the degradation of the loads can also be important.

Ultraviolet During the interaction of a photon with an atom or a molecule of the material, the photon can either be absorbed or diffused with possible loss of energy. In space environment, the solar electromagnetic spectrum covers wavelengths from 10
4 to 104 μm, with a spectral distribution close to a black body at 5762 K. The total energetic flux (called the solar constant) is 1366 W.m
2 (ASTM E490). However, the degradation of materials is mainly due to the ultraviolet radiation, with wavelengths lower than 0.4 μm. The energetic flux in this range is 118 W.m
2. Two phenomena can be distinguished, depending on the wavelength: – 0.2 < λ < 0.4 μm: the energy of ultraviolet photons is between 3.1 eV at 0.4 μm and 6.2 eV at 0.2 μm. This energy is not sufficient to ionize molecules but can generate excitation. – λ < 0.2 μm: the energy of the photons is sufficient to ionize molecules, leading for example to scissions or crosslinking in polymer chains. These photochemical effects may result in a change of color of the material, increase of solar absorptance and/or loss of mechanical properties.

32

Adhesives in Space Environment

931

Charged Particles The energy of charged particles in space environment is usually between 1 keV and several MeV. Atoms and molecules along their trajectory can be either excited or ionized. In the case of an ionization, the secondary electron, liberated during the primary interaction between the charged particle and the matter, can have sufficient energy to leave the site of the primary interaction and produce new excitations and ionizations along its trajectory, before being thermalized and absorbed at a few hundred angstroms from the primary interaction site (see Fig. 10). The reactions held on the track of the secondary electron are often grouped and form clusters, responsible for the chemical modification of the material. If the secondary electron is not sufficiently energetic, it can be recaptured by the parent ion to form an excited molecule or a free radical. A free radical can recombine with the material to form the initial material or can diffuse around the interaction site, leading to the formation of new chemical products. This can be particularly important for contamination aspects. Indeed, the materials outgassing is controlled on space programs as it can degrade the nearest instruments (such as telescope) properties. Consequently, if we consider that the nature of the created species only depends on the interaction with secondary electrons, then, it is independent of the nature of the radiation. On the contrary, the spatial distribution of the created species will depend on the radiation. In the case of protons, the primary track is linear and the defects are distributed in dense clusters very close to the track. In the case of “δ-ray”

Ion clusters

Track of primary particle isolated ionizations and spurs track of low-energy secondary electron positive ion free electron excited molecule

Fig. 10 Distribution of ions and excited molecules along the trajectory of a rapid electron (Clegg and Collyer 1991)

932

S. Dagras et al.

electrons, the effects will be more dispersed because of the scattering of the particles in matter. Moreover, the difference of mass between electrons and protons induces different stopping powers and penetration depths in matter. In the case of charged particles, it is generally estimated in a first approximation that the degradation is a function of the quantity of reactional intermediates (excited molecules, free radicals, etc.), then a function of the dose of energy which is absorbed in the material.

ATOX Atomic Oxygen can react with polymers, carbon, and many metals to form oxygen bonds with atoms on the exposed surface. For most polymers, hydrogen abstraction, oxygen addition, or oxygen insertion can occur. With continued Atomic Oxygen exposure, all oxygen interaction pathways eventually lead to volatile oxidation products. This results in gradual erosion of hydrocarbon or halocarbon material. Silicone materials, a family of commonly used spacecraft materials, are an exception to polymeric materials that erode away with Atomic Oxygen exposure. With Atomic Oxygen exposure, oxidation reactions cause the surface of silicones to oxidize and volatilize their hydrocarbon content and convert to a hardened silica-based surface layer, which is resistant to Atomic Oxygen erosion. ATOX effects can be important on space materials and it is thus important to know it. However, this will not be discussed in details as adhesives are generally shielded from direct exposure to Atomic Oxygen, and they are not affected by this space environment factor.

32.5.2 Irradiation Effects Mechanical Properties Change in mechanical properties is particularly critical in the case of polymers. Both UV and particles can damage polymers. The primary effects are generally deformation, embrittlement, and discoloration, which impact the mechanical integrity and the thermal equilibrium of the material. These effects are determined by measurement of changes in lap shear strength, tensile strength, peel test or fatigue tests following irradiation. The modification of mechanical properties is the result of the competition between chain scissions and crosslinking in the macromolecular chains. Scissions are ruptures in the macromolecular chain and lead to a diminution of the elastic and the resistance limits and an increase of the elongation capability. On another hand, crosslinking consists in the formation of new transversal bonds in the macromolecular chain: the polymer becomes two- or three-dimensional. This contributes to an increase of the resistance limit but a diminution of flexibility, the polymer becoming more fragile. The crosslinking is illustrated on silicone rubbers through several studies (Zhang et al. 2006; Jochem et al. 2013). After 200 keV proton irradiation up to a fluence of 1015 p+/cm2 or 400 keV electron irradiation up to a fluence of 3.1015 e
/cm2, the network densifies, which leads to an increase of the glass

32

Adhesives in Space Environment

933

Fig. 11 Cracks in an epoxy adhesive foil after irradiation

Fig. 12 Illustration of the mechanical properties degradation (shear strength) on an acrylic adhesive before and after electrons irradiation

Acrylic tape Average shearing strength (N) 140 120 100 80 60 40 20 0 Reference

Irradiated

transition temperature of a few degrees, an increase in tensile strength and an increase of the modulus. All these changes reflect the increase of the macromolecular chain rigidity after irradiation. Even if protons penetrate less than electrons, the effects of the irradiation are similar and can be detected on the bulk material properties (see Figs. 11 and 12). At higher doses, the competition between crosslinking and chain scission is in favor of the second one, resulting in a degradation of the polymer. As the crosslinks density decreases, the tensile strength also decreases and surface cracks are initiated as a result of the backbone chain length reduction (Fig. 13). Few studies are available on epoxies used as adhesives, as they are generally protected from direct environment exposure and they are known to be very resistant to radiation from a mechanical point of view. When epoxies are used as a matrix for

934

S. Dagras et al.

Fig. 13 Evolution of properties of a silicone rubber versus proton irradiation fluence: tensile strength (left), crosslinking density (middle), glass transition temperature (right) (Zhang et al. 2006)

composites in structural applications, plasticization of the polymers at elevated temperatures increases residual deformations in the material structure, while at low temperatures, embrittlement of the polymer, repeated from cycle to cycle, multiplies the effect of microcracking and changes of the composite’s properties. Depending on the delta-CTE between the matrix and the fiber, the additional microcracking due to radiation degradation may be of second order compared to the effect of thermal cycling only as described in Fig. 14 (Paillous and Pailler 1994).

Optical and Thermo-Optical Properties The macroscopic effect of the degradation of polymers is a change of color of the material, which is coming from the chemical reactions in organic molecules (bond breakage or crosslinking). It is known that UV radiation can darken optical adhesives and coatings. In this regard, silicones are superior to epoxies. An example is given in Fig. 15 (Fisher et al. 2013). The UV irradiation of several silicone adhesives used for bonding solar cells on panels show a coloration of the glue, which is favored by the creation of crosslinks. Charged particles may also lead to discoloration of adhesives. Depending on the particles energy, the effect can be observed in the bulk material (Fig. 16).

32

Adhesives in Space Environment

935 14

14 12 10 8 Epoxy A

6

Epoxy B

4 2 0

0

20

40

60

Number of thermal cycles in a typical temperature range (-100°C/+100°C)

Microcracks density (cracks.cm-1)

Microcracks density (cracks.cm-1)

16

12 10 8 Epoxy A

6

Epoxy B

4 2 0

0

20

40

60

Equivalent me in space (years)

Fig. 14 Example of typical microcracking evolution: effect of thermal cycling (left) and effect of a combined thermal cycling and GEO-simulated environment (right) on two different epoxy adhesives. In this example, the Epoxy A is affected by thermal cycling and radiations while the Epoxy B is only affected by radiative environment

Fig. 15 Silicone glues used for the assembly of solar cells: Picture of sandwiched samples (coverglass/glue/coverglass) after UV exposure under vacuum using the ESTEC SUV facility, (a) Elastosil S 690, (b) Elastosil S 695, (c) Dow Corning DC 93–500 (Fisher et al. 2013)

An increase of the solar absorptivity can also be observed for polymer adhesive tapes, which can be critical if they are used for satellite thermal control purposes where specific optical properties are needed. Generally, this change is mainly attributed to the polymer substrate, with a slight contribution of the adhesive. As for example, the comparison between 3 M™ 92 (kapton ®/silicone) and Permacel P-224 (kapton ®/acrylic) after combined UV and charged particles irradiation shows a slightly higher degradation of the silicone tape, due to the difference in the absorption properties of the adhesive (Marco et al. 2009) (see Fig. 17).

936

S. Dagras et al.

Fig. 16 Silicone bulk adhesive evidence of color change before and after charged particles irradiation (strong darkening)

0.7

Solar absorptivity

Fig. 17 Solar absorptivity evolution of adhesive tapes after the simulation of 8 years in geostationary environment (Marco et al. 2009). 3 M™ 92 (kapton ®/silicone) and Permacel P-224 (kapton ®/ acrylic)

0.6 0.5 Adhesive film P224 0.4

Adhesive film 3M92

0.3 0

2

4 Duration (year)

6

8

Electrical Properties Changes in the electric conductivity have already been observed on materials after irradiation. This phenomenon results from the interaction of the charged particles of the space environment with materials and is very complex. It depends on the initial electrical conductivity (in surface and in volume), the secondary emission, and the photoemission of the material. In some cases, the accumulation of charges at the surface of the material can be followed by electric discharges and electromagnetic emission, which alters the material locally and can affect the good functioning of electronic equipments. Epoxy adhesives have generally a good behavior with respect to radiation environment. The irradiation at high-radiation doses has however an effect on the bulk conductivity of epoxies. On Scotch-Weld™ DP490, a decrease of the bulk conductivity can be observed from 10
14 Ω
1.m
1 for epoxy pristine samples to 2.10
15 Ω
1.m
1 after 400 keV electrons irradiation (up to 106 Gray) (Siegel and Stewart 1969). Indeed, at such dose level, high energy radiation can induce physicochemical ageing which tends to lower ionic transport and therefore the overall bulk conductivity.

32

Adhesives in Space Environment

937

Fig. 18 Arrhenius plot of σ DC as measured by surface potential decay method on irradiated silicone samples: without fillers (open symbols ○, □, Δ) and with fillers (filled symbols ●, ■, ~). Corresponding values of temperature and σ DC are displayed on the upper x-axis and right-hand y-axis, respectively. Arrhenius fits of data are represented by dashed lines (Lewandowski et al. 2016)

Silicone adhesives irradiated in similar conditions (400 keV electrons, up to 1.4.106 Gray) also reveals a decrease in the electrical conductivity (Lewandowski et al. 2016). The behavior of the adhesive is affected by the presence of the inorganic fillers in the formulation, which contribute to enhance the degradation (Fig. 18). A suggested mechanism is the crosslinking in the silicone network and at filler-matrix interfaces. The densification of SiO3 crosslinks in the polymer network after irradiation leads to an increase in resistivity. On another hand, if the polymer is charged with inorganic fillers, the SiO4 crosslinks generated at filler-matrix interface become the most restrictive barrier to the charge transport, so that the increase of resistivity is higher for charged adhesive than for the neat one.

Outgassing Once in space environment, adhesives are under vacuum, so they exhibit outgassing of volatile and sometimes condensable components, which can become contaminants for critical surfaces. Irradiation can affect the outgassing, as the degradation of the polymer leads to gas generation. Indeed, gas yields are good indicators of radiation resistance and can help in understanding degradation mechanisms. VUV radiation and protons can increase the outgassing rate of adhesive materials. A study dedicated to the vacuum UV photolysis of a polydimethylsiloxane (PDMS) polymer has indicated the breakdown of the Si-CH3 bond as the most probable reaction occurring and resulting in the formation of methane and hydrogen as volatile split-off compounds (Roggero 2015). Another problem is the liberation of contaminant species when scissions of chains occur. These sources of contamination can affect all the spacecraft and have to be taken into account in contamination studies.

938

S. Dagras et al.

Fig. 19 Bubbles creation on an acrylic transfer tape due to irradiation (before irradiation on left/ after irradiation on right).

In the case of very high dose rates, volatile species created during irradiation might not have time to diffuse out of the material. The internal constraints generated by the gas accumulation can then lead to unexpected phenomena. Figure 19 presents an example of acrylic adhesive degradation after exposure to particles irradiation with a high flux. Outgassing phenomena is highlighted as bubbles are visible below the tape.

32.6

Conclusion

Various properties of the adhesives are used on satellites. Indeed, they can be used for optical, thermal, electrical reasons in addition to their more common fixation use. To fulfill all these functions, several materials families are applied as silicones, epoxies, acrylics, or polyurethanes. In function of their location on a spacecraft and depending on the mission, they can be submitted to different constraints and, among all of them, the specific space environment constraints (charged particles, UV, or Atomic Oxygen). It is important to know and to understand the physicochemical degradations in the adhesive materials due to this environment to evaluate the EOL (end of life) properties and to be able to correctly design the satellites parts. It is not possible to reproduce the exact space environment on ground; nevertheless, specific test facilities were developed to be able to test as representatively as possible the different materials (including adhesives) and to characterize their degradation. As an example, some of the facilities enable to cumulate UV and different charged particles irradiation under vacuum. Various effects of space environment can be observed on adhesives in function of the constraint type. The reaction of a material to charged particles, UV and Atomic Oxygen are different and several properties can be affected. The photochemical effect induced by the interaction of a photon with the material while submitted to UV irradiation may result in a change of color of the adhesive and/or loss of mechanical properties.

32

Adhesives in Space Environment

939

While submitted to charged particles irradiation (protons or electrons), the adhesives atoms and molecules can be either excited or ionized. The result consists generally in a first step in a cross linking of the macromolecules and finally in a chain scission while submitted to an increasing ionization dose level. The overall mechanical properties of the adhesives are generally impacted with an increase of the adhesive stiffness associated to an additional brittleness. Atomic Oxygen can also react with polymers and, for the sensitive materials, a surface erosion is generally the main effect. Adhesives are used on spacecraft for various applications. Depending on the satellite mission, the use and location of the adhesive, the undergone environment can be very different, leading to various material degradations mechanisms. In the future, two different ways forward for adhesives on the space industry could be considered: – Specific developments and formulation of adhesives used on scientific programs for dedicated applications (functionalization of the adhesives). These specific developments would take into account the behavior of the organic compound with regards to space constraints. – Using more commercial adhesives not specifically developed to withstand the space constraints (charged particles, Atomic Oxygen, or UV). In this case, there will not have specific developments but an adaptation of the spacecraft design to take into account the adhesive limitations and its degradation due to the environment. In both cases, it is important to know and to understand the behavior of adhesives in space environment and the physicochemical phenomenon associated. The tests performed on ground in dedicated facilities and the R&D studies on this really specific domain are thus a major point in the future of space programs.

References ASTM E-490 Standard. Solar constant and zero air mass solar spectral irradiance tables ASTM E512 Standard. Practice for combined simulated space environment testing of thermal control materials with electromagnetic and particulate radiation. American Society for Testing and Materials, Philadelphia ASTM.E.490A Standard. Solar constant and zero air mass solar spectral irradiance tables Roggero A (2015) Analyse du vieillissement d’un adhésif silicone en environnement spatial: influence sur le comportement électrique », CIRIMAT/ONERA. PhD thesis Banks BA, Dill GC, Loftus RJ, de Groh KK, Miller SK (2013) Comparison of hyperthermal ground laboratory atomic oxygen erosion yields with those in low earth orbit. NASA/TM—2013216613 Clegg DW, Collyer AA (1991) Irradiation effects on polymers. Elsevier Applied Science, London/New York ECSS-Q-ST-10-04C (2008) Space engineering: critical item control ECSS-Q-ST-70-06C (2008) Space product assurance: particle and UV radiation testing for space materials

940

S. Dagras et al.

Fisher HR et al (2013) Degradation mechanism of silicone glues under UV irradiation and options for designing materials with increased stability. Polym Degrad Stab 98:720–726 Fornes E et al (1981) The effects of electron and gamma radiation on epoxy based materials. 3rd annual technical review, Nov 16–19 https://www.britannica.com/science/Van-Allen-radiation-belt https://www.nasa.gov/mission_pages/sunearth/news/gallery/20130228-radiationbelts.html https://www.newport.com/n/solar-simulator-spectral-irradiance-data. Newport light sources. Technical note: “solar simulator spectral irradiance data” Jochem H et al (2013) Effects of 400 keV electrons flux on two space grade silicone rubbers. Mater Chem Phys 141:189–194 Kleiman J, Iskanderova Z, Gudimenko Y, Horodetsky S (2003) Atomic oxygen beam sources: a ciritcal overview. Proceedings of the 9th international symposium on materials in a space environment, Noordwijk, 16–20 June 2003 Lewandowski S et al (2016) Particle flux effects on physicochemical polymer degradations. J Spacecr Rocket 53(6):1146–1151. https://doi.org/10.2514/1.A33500 Makhlis FA (1975) Radiation physics and chemistry of polymers. J Chem Educ 53(2):138. https:// doi.org/10.1021/ed053pA138.1 Marco J, Remaury S, Tonon C (2009) Eight years GEO ground testing of thermal control coatings. ISMSE 11 – unpublished NASA-HDBK-6024 (2014) Spacecraft polymers atomic oxygen durability handbook Paillous A (1993) Radiation damage to surface and structure materials. In: The behavior of systems in the space environment, pp 383–405. https://doi.org/10.1007/978-94-011-2048-7_17 Paillous A, Pailler C (1994) Degradation of multiply polymer-matrix composites induced by space environment. Composites 25(4):287–295 Siegel S, Stewart T (1969) Vacuum-ultraviolet photolysis of polydimethylsiloxane. Gas yields and energy transfer. J Phys Chem 73:823–828 “Space weather effect catalogue” ESTEC/Contract No.14069/99/NL/SB; ESA Space Weather Study (ESWS); WP 310 Range of space weather and effects (ESWS-FMI-RP-0001) Tavlet M, Hominal L (1997) Shear tests on adhesive for magnet collars for the LHC. Cryogenics 38(1998):47–50 Tonon C (2000) PhD thesis, ENSAE Van de Voorde M, Restat C (1972) Selection guide to organic materials for nuclear engineering. CERN technical report, pp 72–77 Zhang L et al (2006) Effect of 200 keV proton irradiation on the properties of methyl silicone rubber. Radiat Phys Chem 75:350–355

Fatigue Load Conditions

33

Ian A. Ashcroft

Contents 33.1 33.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2.1 Fatigue Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2.2 Fatigue Initiation and Propagation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.2.3 Fatigue Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3 Factors Affecting Fatigue Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3.1 Load Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.3.2 Environmental Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.4 Prediction Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.4.1 Total-Life Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.4.2 Phenomenological Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.4.3 Fracture Mechanics Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.4.4 Damage Mechanics Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.5 Creep-Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.6 Impact Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33.7 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

942 943 943 945 946 947 947 948 950 950 955 959 964 966 970 971 972

Abstract

Fatigue involves the failure of materials under cyclic loading, where the maximum load can be significantly lower than that required to cause static failure. Polymeric adhesives, as with most materials, are susceptible to fatigue failure, and hence, fatigue should be accounted for when designing bonded structures subjected to cyclic loading. Adhesive joints have potentially good fatigue resistance compared with other joining methods; however, they are also susceptible to accelerated fatigue failure due to the combined actions of fatigue with I. A. Ashcroft (*) Faculty of Engineering, University of Nottingham, Nottingham, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_33

941

942

I. A. Ashcroft

environmental ageing and/or viscoelastic creep. In this chapter, the effect of the environment and various fatigue loading parameters on the fatigue behavior of adhesively bonded joints is discussed before describing the main methods of characterizing and predicting fatigue. Traditionally, fatigue behavior has been characterized through the use of experimentally derived stress–life plots, and fracture mechanics-based progressive crack growth methods have also been widely discussed. In more recent years, damage mechanics-based progressive modeling methods have been proposed that have the advantage of predicting both initiation and crack progression phases of fatigue and have also been shown to be readily adapted to the prediction of variable amplitude fatigue and combined fatigue-environmental ageing. The chapter finishes with descriptions of two special cases of fatigue: creep-fatigue and impact fatigue, which have been shown to be extremely detrimental to the fatigue life of bonded joints under certain conditions.

33.1

Introduction

In engineering, fatigue relates to the failure of a structure under cyclic loading, generally, at a significantly lower load than that required for quasistatic failure. Some form of fatigue loading is present in most engineering structures, for example, aircraft, ships, cars, buildings, and bridges, and is also seen in many nonengineering applications, such as sports equipment, furniture, and even human parts, such as knees and elbows. Fatigue is a particularly dangerous phenomenon as it can result in sudden, catastrophic failure after many years, or decades, of safe service. This is because a long period can be spent in the initiation phase of fatigue damage, in which there are little or no outward signs of damage. Damage can accelerate unstably once a critical degree of damage has been attained, leading to rapid failure of the structure. Fatigue damage can be initiated or accelerated by many factors, such as accidental impact, overloading, corrosion, surface damage, and abrasion. The ubiquitous nature and potentially disastrous effect of fatigue in engineering structures means that in applications where cyclic loading is significant, the designer must attempt to design against fatigue failure. Unfortunately, the stochastic nature of fatigue damage means it is difficult to predict accurately. This difficulty is compounded by the fact that the in-service loading and environment of many engineering components, which can significantly affect the fatigue process, are seldom known to a great degree of accuracy. Hence, it is generally difficult to design against fatigue failure without resorting to large safety factors, and hence incurring the large structural inefficiencies associated with overdesign, with subsequent cost and performance consequences. An alternative to designing against fatigue failure is to monitor parts for fatigue damage at prescribed intervals and remove them from service before damage reaches a critical point. This is a useful and efficient method of preventing fatigue failure where crack growth is stable and can be easily monitored. However, in some cases, such as is often the case with adhesive joints, monitoring fatigue damage is difficult, for example, if the damage initiation is in an

33

Fatigue Load Conditions

943

inaccessible location or the critical crack size before rapid fracture is small. In this case, an alternative strategy is required to combat fatigue, such as automatic replacement of parts after the design life or a “safe fail” design. One of the benefits of bonded joints is that stresses are more uniformly distributed than in riveted or bolted joints. However, stress concentrations will, of course, still exist in bonded joints, as discussed in detail in ▶ Chaps. 24, “Analytical Approach,” and ▶ 25, “Numerical Approach: Finite Element Analysis.” Another potential advantage is that the bonding process does not explicitly weaken the adherends as the high temperatures and phase transformations involved in welding and the hole drilling required for rivets and bolts can. Although, it should be noted that there are likely to be stress concentrations in the adherend in the joint area which may initiate adherend fatigue failure, such as is seen in aluminum adherend single-lap joints at high cycles and in the delamination of composite adherends. In general, adhesively bonded joints perform well in fatigue compared with bolted and welded joints; however, a number of potential problems for adhesive joints subjected to fatigue should also be recognized. Both the adhesive and the interfacial region between adhesive and adherend are potentially sensitive to the environment and this will affect the fatigue resistance of the joint. Adhesives can also be susceptible to creep under certain conditions, and combined with fatigue this can lead to accelerated failure. In a bonded joint, failure can occur in the adhesive, in the adherend, or in the interfacial region between the two, and the relative fatigue resistance of the various components is dependent on many factors, such as geometry, environment, and loading, and may vary as damage progresses. This complex failure process means that it can be difficult to accurately predict fatigue failure in adhesively bonded joints under real service conditions. Section 2 discusses some of the main factors when considering fatigue, with particular reference to issues applicable to bonded joint. Section 3 discusses how various loading parameters and environmental conditions affect the fatigue behavior of bonded joints. Section 4 presents the main methods of characterizing and predicting the response of bonded joints to fatigue loading. Finally, there are sections on the special cases of creep-fatigue and impact fatigue (Sects. 5 and 6), before the final summary and conclusion.

33.2

General Considerations

33.2.1 Fatigue Loading In fatigue, the load varies with time, and the load spectrum is usually characterized in terms of peaks and troughs in the varying load. A fatigue cycle is defined as the time between adjacent peaks, and the fatigue frequency is the number of cycles in a unit time, for example, Hz (cycles per second). A fatigue spectrum can be characterized in terms of the applied load or displacement for simple samples, but in more complex structures, it is often more useful to consider the cyclic stress (or some other

944

I. A. Ashcroft

Fig. 1 Constant amplitude, constant frequency, stresscontrolled, sinusoidal waveform

One cycle smax

Stress

sa

sm

Δs

sa smin

Time

parameter that can be related to failure, such as strain or strain energy release rate) in areas of possible fatigue failure. It is common in laboratory experiments to represent fatigue as a constant amplitude, sinusoidal waveform, as illustrated in Fig. 1. This is termed constant amplitude fatigue (CAF) and the waveform may be defined by the frequency and two stress parameters, such as the maximum stress (σ max) and the stress amplitude (σ a). Other parameters defined in Fig. 1 are the mean stress (σ m), the minimum stress (σ min), and the stress range (Δσ). In most engineering applications, however, it is likely that frequency, amplitude, mean, and waveform will vary with time. This is called variable amplitude fatigue (VAF). The first step in analyzing components subjected to VAF is the characterization of fatigue spectra experienced by the component in service. It is possible to characterize the in-service load spectra through simulation or experimentation. In some cases, a typical spectrum may be repeated throughout the structure’s life and will be similar for each structure manufactured, for example, the takeoff and landing sequences of a commercial passenger aircraft or the run-up sequence of power generation equipment. However, for many applications, e.g., private cars, the spectra can vary considerably from unit to unit. In this case, a “worst-case scenario” or statistical approach may have to be taken. Once the in-service load spectrum has been generated, it can be used in simulations and/or testing. In order to accelerate a fatigue test program, it is generally preferable to reduce the spectrum. In metals, where the materials may be rate insensitive over a large range, an easy method of accelerating tests is to compress the spectrum by testing at high frequency, although any time-dependent effects, such as corrosion or creep, will not be accurately represented in such a test. Hence, when devising accelerated service simulation tests for adhesive joints, care must be taken

33

Fatigue Load Conditions

945

Fig. 2 Example of a loadcontrolled, variable amplitude waveform

8

Load, kN

6

4

2

0

2

4

6

8

10

Time, s

when introducing acceleration techniques that any influential time-dependent or load sequencing effects are retained in the spectrum. Another way of reducing the spectrum is to remove cycles from the spectrum that do not contribute to the fatigue damage. Figure 2 shows a reduced fatigue spectrum used to represent the fatigue loading on an aircraft wing. It can be seen that this is a load-controlled fatigue spectrum that includes changes in the load amplitude and mean, but maintains a constant frequency.

33.2.2 Fatigue Initiation and Propagation The fatigue life of a structure is often divided into initiation and propagation phases. In adhesively bonded joints, the differentiation between these two phases, and even if there really are two such phases, remains a contentious issue. At the predictive modeling level, a distinction can be made between how a propagating crack is analyzed and how the number of cycles before a macrocrack has formed can be predicted, and this can pragmatically be used to differentiate between the initiation and propagation phases. Fatigue initiation in adhesives is a complex and little-understood process. Commercial adhesives are multicomponent materials, with filler particles, carrier mats, and toughening particles typically added, and failure may involve many mechanisms, including matrix microcracking (initiation, growth, and coalescence), filler particle fracture or debonding, cavitation of rubber toughening particles, and debonding of carrier mat fibers. An additional difficulty in characterizing the fatigue initiation process in bonded joints is that initial damage tends to be internal. This makes nondestructive characterization difficult, whereas with destructive characterization methods, it is difficult to avoid sectioning artifacts. In most cases, a purely mechanistic definition of the initiation and propagation phases is not achievable, and a more pragmatic approach must be taken. In terms of in-service inspection,

946

I. A. Ashcroft

initiation may be linked to the detectability of flaws, the end of the initiation life being indicated by the first detection of a crack using whatever technique is being deployed. For the stress analyst, a useful differentiation is to treat the fatigue damage as an initiation phase until a sufficient crack has formed that further growth can be predicted using fracture mechanics. In finite element-based damage mechanics, the initiation phase can be defined as the period before complete damage of an element. Mechanistically, there is also a blurring between the definitions of “damage” and “cracking” in a bonded joint. For example, a region considered as damaged rather than cracked will often contain microcracking, and in a cracked joint, only an idealized version of the main macrocrack is usually considered, whereas this will almost certainly be accompanied by other types of damage in a process zone ahead and around the main crack. In finite element modeling, damage is often represented by reducing the material properties (the continuum damage mechanics approach), usually stiffness, of an element. Cracking is usually modeled by detaching elements at nodes, after which fracture mechanics methods can be used to model crack propagation. In cohesive zone modeling, both damage and crack growth are represented by using specialized elements to join adjacent continuum elements. This is discussed in detail in ▶ Chap. 25, “Numerical Approach: Finite Element Analysis.”

33.2.3 Fatigue Testing Mechanical testing of bonded joints can range from inexpensive coupon tests through the testing of structural elements to the testing of full prototypes, which may be extremely expensive. In all cases, however, fatigue testing will be considerably lengthier and more costly than quasistatic testing. Coupon tests can fulfil a number of roles. Single material tests may be used to generate material property data, while joints can be tested to compare material systems or joint geometries, evaluate performance over a range of loading and environmental conditions, generate design data, or provide validation data for predictive models. Coupon samples used in the fatigue testing of bonded joints are similar to those used in quasistatic testing, as discussed in ▶ Chaps. 19, “Failure Strength Tests,” and ▶ 20, “Fracture Tests.” Simple lap joints, such as single- and double-lap joints, are generally used to generate stress–life (S–N ) curves, and standard fracture mechanics tests, such as the double cantilever beam, are used to generate fatigue crack growth curves. The fatigue testing of adhesive lap joints is covered by the standards BS EN ISO 9664:1995 and ASTM D316699. In the former, it recommends that at least four samples should be tested at three different stress amplitude values for a given stress mean, such that failure occurs between 104 and 106 cycles. This standard also advises on statistical analysis of the data. In general, fatigue data exhibits greater scatter than quasistatic data and this needs to be taken into account when using safety factors with fatigue data. Further advice on the application of statistics to fatigue data can be found in BS 3518-5:1966.

33

Fatigue Load Conditions

33.3

947

Factors Affecting Fatigue Behavior

33.3.1 Load Factors Load factors that will affect the fatigue behavior of adhesively bonded joints include the amplitude, mean, and minimum stresses and frequency. In common with most materials, fatigue life in a bonded joint will tend to decrease if the stress amplitude or the mean stress increases. The maximum stress is also important. If this is greater than the yield stress, then low cycle fatigue (LCF) will occur, greatly reducing the fatigue life. In some cases, a critical stress, called an endurance limit, may exist, below which fatigue failure will not occur. As adhesives tend to be viscoelastic or viscoplastic in nature, then a nonzero mean stress can lead to progressive creep of the joint over time. This is exacerbated at low frequencies, where time under load may become as significant as number of cycles in defining failure. At high frequencies, hysteretic heating may lead to premature failure or the high strain rates involved may induce brittle fracture. High strain rates are also seen in impact fatigue, in addition to dynamic effects, which, as discussed in Sect. 6, can be extremely detrimental to adhesives. Most bonded joints are designed for tensile loading, and if the joint is subjected to accidental compressive loading, then buckling may occur, from which high peel forces will arise, leading to rapid fracture. In variable amplitude fatigue, load interaction effects have been observed to both accelerate and retard the rate of fatigue damage, in different materials. Fatigue crack growth rate retardation is probably the more commonly reported phenomenon. For example, it is generally reported that overloads retard fatigue crack propagation in ductile metals. Proposed mechanisms to account for this phenomenon include the effect of compressive residual stresses in the vicinity of the crack tip, crack tip closure effects, and crack tip blunting. Although neglecting such beneficial effects by using a noninteractive lifetime predictive methodology can remove the opportunity to achieve a lighter structure, at least, the design errs on the side of safety. However, if the load interactions cause crack growth acceleration, the structure under investigation can fail much earlier than predicted using CAF data and a noninteractive prediction methodology. Although most published work indicates retardation behavior after an overload, there is also work in the literature reporting crack growth acceleration for both metals and composites. These studies report several different mechanisms accounting for the acceleration behavior. For example, Nisitani and Nakamura (1982) studied the crack growth behavior of steel specimens tested under a spectrum composed of a very small number of overloads and a very large number of cycles below or near the fatigue limit. They observed that the application of a linear cumulative damage rule resulted in extremely nonconservative predictions of fatigue life. A possible explanation was that although understress cycles cannot initiate a crack, they can potentially contribute to the fatigue damage created by overloads. Farrow (1989) found that the fatigue life of composite laminates subjected to small block loading was shorter than that for laminates subjected to large block loadings when the blocks had different mean stress levels. He called this phenomenon the “cycle mix effect.” Schaff and

948

I. A. Ashcroft

Davidson (1997a, b) reviewed the study made by Farrow and developed a strengthbased wearout model. They suggested that the cycle mix effect occurred during the transition from one CAF stage to another having a higher mean stress value, although they did not discuss the mechanisms behind the strength degradation during this transition. Erpolat et al. (2004a) observed a similar crack acceleration effect when VAF testing bonded CFRP double-lap joints.

33.3.2 Environmental Factors It is well known that adhesives and adhesion can be adversely affected by factors in the natural environment effects, and that this is a topic of considerable complexity (see ▶ Chap. 31, “Effect of Water and Mechanical Stress on Durability”). When environmental effects are combined with fatigue testing, we have an added complexity, owing to the introduction of coupled time-dependent effects. The main effects of environmental exposure can be classified as those affecting the adhesive, those affecting the adherend, and those affecting the interface (or interphase) between the two. In terms of the adhesive, an increase in temperature or the absorption of moisture generally results in plasticization of the adhesive, with an accompanying reduction in modulus and failure load. However, strain to failure and fracture toughness may increase. This can affect fatigue initiation and propagation in a number of ways. The plasticization will tend to reduce stress concentrations, although stresses may now be significant over a larger area; hence, the resistance to brittle fatigue failure may increase, but the resistance to creep-fatigue may decrease. These effects will become more significant close to the glass transition temperature (Tg), and similarity can be seen between the effects of absorbed moisture, increased temperature, and decreased test rate (or frequency). Some of these issues are illustrated by the results shown in Table 1 (data from Ashcroft et al. 2001a, b). This table shows the fatigue limits for bonded CFRP lap-strap and double-lap joints. In the case of the lap-strap joints, it can be seen that temperature has little effect on Table 1 Effect of environment on the fatigue limit for bonded CFRP-epoxy lap-strap and doublelap joints (Data from Ashcroft et al. 2001a, b) Sample Lap-strap joint

Preconditioning Vacuum desiccator

45  C/85% RH

Double-lap joint

Vacuum desiccator

Test conditions
50  C/ambient 22  C/ambient 90  C/ambient 90  C/97% RH 22  C/95% RH 90  C/ambient 90  C/97% RH
50  C/ambient 22  C/ambient 90  C/ambient

Fatigue limit (kN) 14 15 14 7 15 5 5 10 10 3.3

33

Fatigue Load Conditions

949

the fatigue threshold of those sample stored and tested in nominally dry conditions. However, samples tested in hot-wet conditions experience a significant reduction in the fatigue threshold. Samples were also conditioned under high humidity conditions until saturation. Samples subsequently tested wet at 22  C had no change in the fatigue threshold compared to those tested dry, whereas samples tested at 90  C, whether wet or dry, experienced a large reduction in the fatigue threshold. Interpretation of these results is complicated by the fact that complex mixed mode failure paths were observed. In order to explain these results, the effect of temperature and moisture on the mechanical behavior of the adhesive must be considered. The stress to failure and modulus of the adhesive decreases as the temperature increases, but the strain to failure and total strain energy density at failure increases. The competing effects of these different trends conspire to maintain the fatigue limit at a relatively constant value between 22  C and 90  C when stored and tested dry. Differential thermomechanical analysis (DTMA) was carried out on samples saturated to different moisture levels, and it was seen that for every 1% of moisture absorbed, the glass transition point of the adhesive decreased by approximately 15  C. As the Tg of this adhesive is approximately 130  C, the saturated adhesive tested at 22  C would not be expected to behave markedly differently from those tested dry at 22  C and 90  C. However, the saturated sample tested at 90  C is in the glass transition temperature range as the adhesive is capable of absorbing 3–4% of moisture. It is clearly less capable of resisting stress under these conditions, and hence the fatigue limit is greatly reduced. If the lap-strap results are now compared with the results from testing double-lap joint manufactured using the same materials, it can be seen in Table 1 that whereas the lap-strap joints are relatively temperature insensitive over the test range when dry, the double-lap joints experience a large decrease in fatigue resistance as temperature is increased from 22  C to 90  C. This can be attributed to creep-enhanced failure at the higher temperature in the double-lap joint, which experiences accumulated creep because the joint remains under a tensile load throughout the fatigue testing. This phenomenon is prevented in the case of the lap-strap joints because the CFRP strap adherend spanning the loading points remains elastic. The combined effects of fatigue and creep are discussed further in Sect. 5. Moisture can also affect the interface between the adhesive and adherend, and this can significantly affect the fatigue behavior of a bonded joint. For example, Little (1999) carried out fatigue tests on aluminum single-lap joints bonded with the same adhesive as that in the experiments discussed above. He found that with chromic acid-etched (CAE) adherends, there was little difference in the fatigue threshold of those samples tested dry and those tested immersed in distilled water at 28  C. However, samples with grit-blasted and degreased (GBD) adherends exhibited a significantly lower fatigue threshold when tested wet, and the locus of failure changed from cohesive failure in the adhesive to failure in the interfacial region. Datla et al. (2011a) also investigated the combined effects of temperature and humidity on the fatigue behavior of adhesive joints with pretreated aluminum adherends. In their case, an asymmetric double cantilever beam was used and it was found that for their system, the joint degradation was mainly influenced by

950

I. A. Ashcroft

elevated temperature at high crack growth rates and elevated moisture at low crack growth rates. They also investigated the effect of a cyclic ageing environment for this joint, with intermittent salt spray on these joints (Datla et al. 2011b) and found superior fatigue performance compared with joints subjected to constant humidity ageing. They attributed this to salt water environment causing lower water concentrations in the adhesive. For further reading on this subject, Costa et al. (2017) have recently reviewed the published literature regarding the effect of environmental conditions and ageing on fatigue performance.

33.4

Prediction Methods

The ability to predict the fatigue behavior of bonded joints is potentially useful for a number of reasons. Firstly, it can be used to support the design of bonded structures, to ensure that fatigue failure is not likely to occur in service, and to aid in the design of efficient fatigue-resistant joints, resulting in safer, cheaper, and higher performance structures. Analytical or computational predictive modeling can help in these objectives; however, the current state of confidence in such modeling means that in most cases, it must be accompanied by a complementary testing program. Modeling can also be helpful in developing and understanding of the mechanisms involved in fatigue failure. This can be achieved through comparing the results from carefully designed experimental tests with the predicted results from progressive damage models. Finally, predictive modeling can be used to support the in-service monitoring and re-lifeing of structures. The main goals in the modeling of fatigue are to predict the time (or number of cycles) for a certain event to occur (such as macrocrack formation, critical extent of damage, or complete failure) or to predict the rate of change of a fatigue-related parameter, such as crack length or “damage.” The various methods of doing this are presented in this section, and the approach used is to introduce and describe each of the main methods that have been used to date, together with one or two examples of their application to bonded joints.

33.4.1 Total-Life Methods In the total-life approach, the number of cycles to failure (Nf) is plotted as a function of a load-related variable, such as stress or strain amplitude. Where the loading is low enough that the deformation is predominantly elastic, a stress variable (S) is usually chosen and the resultant plot is termed an S–N curve, or Wöhler plot, and this is known as the stress–life approach. Under these conditions, a long fatigue life is often seen, and hence this is sometimes termed “high cycle fatigue” (HCF). The S–N data is either plotted as a log-linear or a log-log plot and a characteristic equation can be obtained by empirical curve fitting. The constants in the curve-fitted equations are dependent on many factors, including material, geometry, surface condition, environment, and mean stress. Caution should be exercised when trying to apply S–N

33

Fatigue Load Conditions

951

14 22°C 90°C 22°C (unbroken) 90°C (unbroken)

Load amplitude, kN

12 10 8 6 4 2 0 1

10

100

1,000 10,000 Cycles to failure

100,000 1,000,000

Fig. 3 Load–life curve for CFRP-epoxy double-lap joints (Data from Ashcroft et al. 2001b)

data beyond the samples and conditions used to generate the data. The standard stress–life method gives no indication of the progression of damage, although in some cases, the onset of cracking is indicated on the plot in addition to the complete failure, hence allowing the initiation and propagation phases to be differentiated. The above factors mean that the S–N curve is of rather limited use in predicting fatigue behavior; however, it is still useful as a design tool and in fatigue modeling as a source of validation data. A further limitation in the application of S–N curves to fatigue prediction in bonded joints is that there is no unique relation between the easily determined average shear stress in the adhesive layer and the maximum stress. For this reason, load rather than stress is often used in total-life plots for bonded joints and these are known as L–N curves. A typical L–N curve for epoxy-bonded CFRP double-lap joints is shown in Fig. 3. It can be seen that the L–N curve can be divided into a number of different regions: a low cycle fatigue (LCF) region below approximately 1000 cycles, a high cycle fatigue (HCF) region between approximately 1000 and 100,000 cycles, and an endurance limit region above approximately 100,000 cycles. Fatigue life depends not only on the load amplitude but also on the mean load as either increasing the mean or increasing the amplitude tends to result in a reduction of the fatigue life. This is illustrated in Fig. 4a, which shows results from fatigue testing steel-epoxy single-lap joints at different load amplitudes and R-ratios (maximum load/minimum load), where an increasing value of R indicates an increasing mean for a given load range. It can be seen that either increasing the mean or increasing the amplitude results in a reduction of the fatigue life. The relationship between amplitude and mean on the fatigue life can be illustrated in a constant-life diagram in which combinations of mean and amplitude are plotted for a given fatigue life. A constant-life diagram plotted from the data in Fig. 4a is shown in Fig. 4b. This

952

I. A. Ashcroft

a

Load amplitude, kN

4.5 4

R = 0.1

3.5

R = 0.5 R = 0.75

3 2.5 2 1.5 1 0.5 0 100

1,000

10,000

100,000

1,000,000

Cycles to failure

b 3

Load amplitude, kN

2.5 2 1.5 1 10,000 cycles 0.5

100,000 cycles

0 0

2

4 6 Mean load, kN

8

10

Fig. 4 (a) Effect of R-ratio on fatigue life. (b) Constant-life diagram (Data from Crocombe and Richardson 1999)

relationship can be represented by the Goodman (linear) or Gerber (parabolic) relationships (Dowling 1999). It can be seen that in this case, a linear relationship provides a reasonable fit to the experimental data. In some cases, efforts have been made to differentiate between the initiation and propagation phases in the S–N behavior of bonded joints. Shenoy et al. (2009a) used a combination of back-face strain measurements and sectioning of partially fatigued joints to measure damage and crack growth as a function of number of fatigue cycles. It was seen from the sectioned joints that there could be extensive internal damage in the joint without external signs of cracking; therefore, determination of an initiation phase from external observations alone is likely to lead to an

Fatigue Load Conditions No.of cycles to failure, Nf

Failure curve

FG

Fatigue load

Fig. 5 Enhanced load–life curve for adhesively bonded single-lap joints, showing regions of crack initiation (CI), stable crack growth (SCG), and fast crack growth (FG). The broken lines show back-face strain as a function of number of cycles (Shenoy et al. 2009a)

953

Normalized backface strain

33

SCG

CI No. of cycles, N

overestimation of the percentage of the fatigue life sent in initiation. Shenoy et al. (2009a) identified three regions in the fatigue life of an aluminum/epoxy single-lap joint, as illustrated in Fig. 5. An initiation period (CI) in which damage starts to accumulate, but a macrocrack has not yet formed, a stable crack growth (SCG) region in which a macrocrack has formed and is growing slowly, and a fast crack growth region (FCG), which leads to rapid failure of the joint. It was seen that the percentage of life spent in each region varied with the fatigue load. At low loads, the fatigue life was dominated by crack initiation, whereas crack growth dominated at high loads. The broken lines in Fig. 5 show how the back-face strain signal varies as a function of fatigue cycles. This can be used to characterize the different phases of crack growth, in particular a rapid change in the back-face strain is seen when the fatigue failure enters the fast crack growth region. The S–N (or L–N ) curve is only directly applicable to constant amplitude fatigue, whereas in most practical applications for structural joints a variable amplitude fatigue spectrum is more likely. A simple method of using S–N data to predict variable amplitude fatigue is that proposed by Palmgren (1924) and further developed by Miner (1945). The Palmgren–Miner (P–M) rule can be represented by X ni ¼C N fi

(1)

where ni is the number of cycles in a constant amplitude block, Nfi is the number of cycles to failure at the stress amplitude for that particular block and can be obtained from the S–N curve, and C is the Miner’s sum and is ideally assumed to equal 1. Using Eq. (1), the fatigue life of a sample in variable amplitude fatigue can be predicted from an S–N curve obtained from constant amplitude fatigue testing of similar samples. However, there are a number of serious limitations to this method, primarily, the assumptions that damage accumulation is linear, that there is no damage below the fatigue threshold, and that there are no load history effects.

954

P–M Extended P–M Stress amplitude

Fig. 6 Comparison of P–M, extended P–M, and elementary P–M methods of predicting variable amplitude fatigue cycles to failure

I. A. Ashcroft

Elementary P–M

Log cycles to failure

Modifications to the P–M rule have been suggested to address some of the deficiencies; for instance, the extended and elementary P–M rules shown in Fig. 6 have been proposed to allow cycles below the fatigue threshold to contribute to the damage sum. Modifications to account for nonlinear damage accumulation and interaction effects have also been suggested; however, any improvements are at the expense of increased complexity and/or increased testing requirements, and the basic flaw in the method, i.e., that it bears no relation to the actual progression of damage in the sample, is still not addressed. Erpolat et al. (2004a) used the P–M law and the extended P–M law, in which cycles below the endurance limit also contribute to damage accumulation, to predict failure in an epoxy-CFRP double-lap joint subjected to a variable amplitude (VA) fatigue spectrum. The resulting Miner’s sum was significantly less than 1, varying between 0.04 and 0.3, and decreased with increasing load. This indicates that load sequencing is causing damage acceleration, i.e., that the P–M rule is nonconservative in this case. Some materials exhibit a stress level below which fatigue failure will not occur, known as the fatigue or fatigue limit. In this case, it may be possible to use the data from one sample to predict the fatigue limit for a different geometry or loading condition. A useful method is to experimentally determine the fatigue limit using a calibration sample and to then calculate the value of a suitable failure parameter at the fatigue limit. The fatigue limit can then be predicted for different geometries for failure in the same material by determining the load at which the fatigue limit value of the chosen failure parameter is reached in the new geometry. This is a potentially attractive method as the fatigue threshold value of the chosen failure criterion can be determined from an inexpensive test and then used to predict the fatigue limit for joints that would be expensive to test experimentally. However, in practice, there are the same difficulties as those in the prediction of the quasistatic failure load of adhesive joints, i.e., selection of an appropriate and robust failure criterion, dealing with the theoretical stress singularities, scaling issues and differences in the stress conditions in the simple test and application joints. Abdel Wahab et al. (2001a) predicted the fatigue threshold in CFRP-epoxy lap joints using a variety of stressand-strain-based failure criteria, taking the value of stress and strain at a distance of

33

Fatigue Load Conditions

955

0.04 mm from the singularity to avoid mesh sensitivity. They also used the plastic zone size as a failure criterion. Under high stress amplitudes, plastic deformation occurs and the fatigue life is considerably shortened. This is known as “low cycle fatigue” (LCF). Coffin (1954) and Manson (1954) proposed that Nf could be related to the plastic strain amplitude, Δep/2, in the LCF region. Δϵp ¼ BðN f Þβ 2

(2)

where B and β are material constants. The strain–life approach is more difficult to implement than the stress–life method as plastic strain is difficult to measure, particularly for nonhomogenous material systems such as bonded joints. Also, structural joints tend to be used in HCF applications, and hence the strain–life method has seen little application to adhesively bonded joints. Abdel Wahab et al. (2010a, b) proposed a low cycle fatigue damage law based on continuum damage mechanics and applied this to bulk adhesive samples and single-lap joints.

33.4.2 Phenomenological Methods In the phenomenological approach, fatigue damage is characterized as a function of a measurable parameter, most commonly the residual strength or stiffness after fatigue damage. The reduction in stiffness with fatigue damage, known as stiffness wearout, has the advantage that it can be measured nondestructively; however, it is not directly linked to a failure criterion and may not be very sensitive to the early stages of damage. The strength wearout method provides a useful characterization of the degradation of residual strength, but requires extensive destructive testing. In the strength wearout method, the joint’s strength is initially equal to the static strength, Su, but decreases to SR(n) as damage accumulates through the application of n fatigue cycles. This degradation can be represented by: SR ðnÞ ¼ Su
f ðSu , Smax , RÞnκ

(3)

where κ is a strength degradation parameter, Smax is the maximum stress, and R is the ratio of minimum to maximum stress (i.e., R = Smin/Smax). Failure occurs when the residual strength equals the maximum stress of the spectrum, i.e., when SR(Nf) = Smax. Shenoy et al. (2009b) proposed a modified version of this equation that they termed the normalized nonlinear strength wearout model (NNLSWM), which is given by: Ln ¼ 1


ðLu
Lmax Þ ðN n Þη Lu

(4)

956 1

Normalized residual load, Ln

Fig. 7 The normalized nonlinear strength wearout model applied to aluminumepoxy single-lap joints tested at different loads (Shenoy et al. 2009b)

I. A. Ashcroft

0.8

0.6 NNLSW model 0.4

Experimental

0.2

0 0

0.2

0.4

0.6

0.8

1

Normalized no. of cycles, Nn

where η is a constant and the normalized residual failure load, Ln, is: Ln ¼

LR ðnÞ Lu

and the normalized cycles to failure, Nn, is: Nn ¼

n Nf

where LR(n) is the quasistatic failure load after n fatigue cycles and Lu is the quasistatic failure load prior to fatigue loading. Figure 7 shows this model applied to aluminum alloy-epoxy single-lap joints tested at three different fatigue loads. It can be seen that the model fits all the data reasonable well, thus providing a simple method of predicting the residual strength in a joint under any combination of constant fatigue load and number of cycles. However, if the fatigue load varies, modifications to this approach may be required, as discussed in the next section. Schaff and Davidson (1997a, b) used the strength wearout method to predict the residual strength degradation of a composite material subjected to a variable amplitude loading spectrum. However, they noted a crack acceleration effect in the transition from one constant amplitude (CA) block to another, the cycle mix effect, and proposed a cycle mix factor, CM, to account for this. Erpolat et al. (2004a) proposed a modified form of Shaff and Davidson’s cycle mix equation to model the degradation of CFRP-epoxy double-lap joints subjected to a variable amplitude fatigue spectrum. They showed that this model represented the fatigue life of bonded joints under variable amplitude fatigue more accurately than Palmgren–Miner’s law. Shenoy et al. (2009c) proposed various further modifications to this approach, which they applied to aluminum alloy-epoxy single-lap joints subjected to various forms of VAF. The application of the cycle mix factor to predict strength wearout and cycles

33

Fatigue Load Conditions

Lu a

957

b

Residual load, LR(n)

c

d

e

CM or CMm

Lmax1 Lmax2

No. of cycles, n

Fig. 8 Illustration of the nonlinear strength wearout with cycle mix factor approach to predict variable amplitude fatigue

to failure for VAF is illustrated in Fig. 8. The figure shows two strength wearout curves, showing the reduction in the residual load for constant amplitude fatigue with maximum fatigue loads of Lmax1 and Lmax2. Both of these curves can be represented by a nonlinear strength wearout equation, such as Eq. (4). If a variable amplitude fatigue spectrum commences with a maximum fatigue load of Lmax2, then the decrease in the residual load with fatigue cycles initially follows path a–b. If at point b the maximum fatigue load increases to Lmax1, there is a horizontal jump to the strength wearout curve for the higher load and the residual load starts to decrease more quickly, following curve c–d. At point d, the maximum load is decreased to Lmax2, and we have another horizontal jump to the relevant strength wearout curve. However, at point e, the cycle mix factor is also applied, which has the effect of an immediate reduction in the residual load by an amount CM (or CMm). Shenoy et al. (2009c) proposed two cycle mix factors. The first was the cycle-independent cycle mix factor used by Erpolat et al. (2004a) given in Eq. (5). h γi CM ¼ α ðΔLmn ÞβLmax ðΔLmax, 1 =ΔLmn, 1 Þ

(5)

ΔLmn and ΔLmax are the mean and maximum load changes during the transition from one mean load to the other, α and β are experimentally determined parameters, and γ was assumed to be unity in this case. The second was termed the modified cycle mix factor (MCM) approach. This was developed after observing that a better fit to the experimental data could be made if the cycle mix factor was greater in the

958

I. A. Ashcroft

low cycle regime than in the high cycle regime. As it had already been noted that with these joints the fatigue life was propagation dominated in the low cycle regime and initiation dominated in the high cycle regime, the desired effect was achieved by making the cycle mix factor dependent on the extent of damage in the sample, as shown in Eq. (6).  CMm ¼

h i y ζ þ α ðΔLmn ÞβLmax ðΔLmax =ΔLmn Þ OL

(6)

where OL is the overlap length and ζ is a damage parameter. In Shenoy et al. (2009c), ζ was determined by fitting a power law curve to experimental plots of damage/crack growth against number of cycles. Hence, ζ was defined as: ζ ¼ m 1 ð nÞ m 2

(7)

where m1 and m2 are experimentally determined constants. It can be seen in Fig. 9 that the cycle mix parameter was applied when the fatigue loading became less severe, i.e., after the jump from d to e, but not when the fatigue loading became more severe, i.e., following the jump from b to c. This is consistent with Gomatam and Sancaktar (2006) who observed damage acceleration when “moderate” loading conditions followed “severe” loading conditions, but not when the order was reversed. It is also consistent with the mechanistic argument put forward by Ashcroft (2004) who attributed crack growth accelerations seen in adhesives subjected to intermittent overloads to the damage caused by the overloads to the adhesive ahead of the crack tip. It was proposed that this damage reduced the fatigue resistance of the adhesive and resulted in accelerated crack growth for the cycles of lower amplitude following an overload.

−2 Gc

−3 Log (da/dN), m/cycle

Fig. 9 Experimental fatigue crack growth curve for a CFRP-epoxy double cantilever beam (Data from Ashcroft and Shaw 2002)

−4 da = C(G )m max dN

−5 −6 −7

Gth

−8 −9

−10 −11 1.8

1.9

2

2.1

2.2

Log (Gmax) J/m2

2.3

2.4

2.5

33

Fatigue Load Conditions

959

As with strength degradation, stiffness degradation rate can be considered as a power function of the number of load cycles. Although stiffness degradation has the advantage that it can be measured nondestructively, it does not give a direct indication of the residual strength of a fatigued structure. However, if this link can be made, then stiffness degradation can be a useful method of in-service structural integrity monitoring.

33.4.3 Fracture Mechanics Methods The fracture mechanics approach deals predominantly with the crack propagation phase; hence, it is assumed that crack initiation occurs during the early stages of the fatigue cycling or that there is a preexisting crack. The rate of fatigue crack growth, da/dN, is then correlated with an appropriate fracture mechanics parameter, such as Griffith’s (1921) strain energy release rate, G, or Irwin’s (1958) stress intensity factor, K. Paris et al. (1961) proposed that da/dN was a power function of the stress intensity factor range, ΔK(=Kmax
Kmin), i.e.: da ¼ CΔK m dN

(8)

where C and m are empirical constants, dependent on factors such as the material, the fatigue frequency, the R-ratio, and environment. Although K is the most widely used fracture mechanics parameter in the fracture analysis of metals, it is more difficult to apply to bonded joints, where the constraint effects of the substrates on the adhesive layer complicates characterization of the stress field around the crack tip. Therefore, G is often used as the governing fracture parameter for adhesives if linear elastic fracture mechanics (LEFM) is applicable (i.e., localized plasticity). If an elastoplastic fracture mechanics (EPFM) parameter is required, owing to more widespread plasticity, then the J-integral (J) is generally used (Rice 1968). If creep is significant, then a time-dependent fracture mechanics parameter, such as C or Ct, should be considered, as discussed in Sect. 5. A plot of the experimentally measured crack growth rate against the calculated Gmax or ΔG often exhibits three regions, as illustrated in Fig. 9, which shows the fatigue crack growth curve for a CFRP-epoxy double cantilever beam (DCB). Region I is defined by the threshold strain energy release rate, Gth, in which crack growth is slow enough to be deemed negligible. Region II is described by a power law equation similar to Eq. (6) and is, hence, sometimes referred to as the Paris Region. In Region III, there is unstable fast crack growth as Gmax approaches the critical strain energy release rate, Gc. In a general form, the relationship between the fatigue crack propagation rate and a relevant fracture parameter, Γ, can be represented by: da ¼ f ðΓ Þ dN

(9)

960

I. A. Ashcroft

The number of cycles to failure can be determined from: 0a 1 ðf da A Nf ¼ @ f ðΓ Þ

(10)

a0

where a0 is the initial crack length and af is the final crack length. Equation (10) can be solved using numerical crack growth integration (NCGI). A simple method of predicting VA fatigue from a CA FCG curve is to perform numerical crack growth integration. In this method, the crack size, a, and corresponding fracture parameter, such as strain energy release rate range, ΔG, is assumed to be constant throughout a CA stage. The crack growth rate per cycle, da/dN, for this stage can be obtained using an appropriate correlation, such as the Paris law. Multiplication of this rate by the number of cycles in the stage, n, gives the overall crack growth during the stage, Δa. This is used to find the crack size for the subsequent stage (ai+1 = ai + Δai) and the procedure is repeated until Gmax exceeds the fracture energy, Gc, or until ΔG becomes equal to ΔGth. This algorithm is summarized below. Repeat aiþ1 ¼ ai þ ni :

da ðΔGi ðRi , ai , . . .ÞÞ dN

(11)

Until Gmax, i+1  Gc {G increasing with a} or Until ΔGmax, i+1  ΔGth {G decreasing with a} Abdel Wahab et al. (2004) proposed a general method of predicting fatigue crack growth and failure in bonded lap joints incorporating NCGI and finite element analysis (FEA). The crack growth law was determined from tests using a DCB sample and this was used to predict the fatigue crack growth, and hence the fatigue life, of single- and double-lap joints manufactured from the same materials. The numerical integration technique can easily be adapted to the prediction of fatigue crack growth in variable amplitude (VA) fatigue. Erpolat et al. (2004b) applied the NCGI technique for the prediction of crack growth in CFRP/epoxy DCB joints subjected to periodic overloads. This tended to underestimate the experimentally measured crack growth, indicating a crack growth acceleration mechanism, and an unstable, rapid crack growth period was also seen when high initial values of Gmax were applied. This behavior was attributed to the generation of increased damage in the process zone ahead of the crack tip when the overloads were applied, and Ashcroft (2004) proposed a fracture mechanics-based model that could predict this behavior. It was assumed that under constant amplitude conditions, FCG can be represented by ΔGth, GC, and the two Paris constants, C and m, as represented by CA in Fig. 10. It can be seen that if a CA fatigue load with a strain energy release rate range of ΔGA is applied, then the resulting crack growth rate will be (da/dN )CA. In most cases, ΔGA will be calculated assuming undamaged material ahead of the crack tip, and this is likely to be adequate for predictive purposes. However, the crack resistance and hence (da/dN )CA will be associated with the damage zone

Fatigue Load Conditions

Fig. 10 The damage shift method of predicting fatigue crack growth under variable amplitude (Ashcroft 2004)

961

Log da/dN

33

OL CA yE

(da/dN)OL (da/dN)CA

ΔGA ΔGAC

Log ΔG

created by the CA fatigue loading. If an overload is superimposed onto the CA spectrum, then the damage ahead of the crack zone will increase and the resistance to crack propagation will decrease. It is proposed that this increased damage can be represented by a lateral shift in the FCG curve. The FCG curve associated with this increased damage is represented by curve OL in Fig. 10. It can be seen that the value of da/dN at ΔGA has increased to (da/dN )OL, i.e., that there is a crack acceleration effect, which is represented by a “damage shift” parameter, ψ E, which can be easily determined for a given spectrum through a simple experimental test program. Theoretically, all that is needed to determine ψ E is the crack growth rates under CA and VA fatigue for a single value of ΔGA. If ΔGA is increased, a critical point will eventually be reached at which Gmax of the overloads is equal to the value of Gc for the shifted FCG curve, OL. This point is shown as ΔGAC in Fig. 10. Unstable or quasistatic fracture then occurs. If G increases with crack length, then this will lead to catastrophic failure of the joint. However, if G decreases with a, as when testing DCB samples in displacement control, then the crack will eventually stop if the G arrest (Garr) value is reached before the joint has completely fractured. The value of ΔGA associated with the crack arrest point (ΔGarr) will now be much smaller, and hence crack growth will be greatly reduced. This is consistent with the experimental crack growth behavior shown in Fig. 11, where there is a rapid increase in crack length after approximately 5000 fatigue cycles, after which there is a period of lower crack growth rate than that predicted by NCGI. It can be seen in Fig. 11 that the damage shift model is capable of predicting both the initial crack jump and the subsequent reduction in the crack growth rate. The fatigue crack growth approach will only predict the correct fatigue life if it is dominated by the fatigue propagation phase. However, if the initiation phase is significant, then this approach will underestimate the fatigue life and a method of predicting the number of cycles before the macrocrack forms is required. This can be done empirically in a similar fashion to the stress–life approach, with the number of

962

I. A. Ashcroft

Fig. 11 Fatigue crack growth for CFRP-epoxy DCB subjected to overloads

90

Crack length, mm

85 80 75 70

NCGI Experimental

65

Damage Shift Model

60 0

0.2

0.4 0.6 Number of cycles x 106

0.8

1

cycles to fatigue initiation, Ni, being plotted as a function of a suitable stress (or other) parameter rather than cycles to total failure, Nf (Shenoy et al. 2009a). An alternative approach, suggested by Levebvre and Dillard (1999), is to use a stress singularity parameter as the fatigue initiation criterion. They showed that under certain conditions, the singular stress, σ kl, in an adhesive lap joint could be represented by: σ kl ¼

Qkl xλ

(12)

where Qkl is a generalized stress intensity factor, dependent on load, and λ is an eigenvalue that can be related to the order of the singularity. It was proposed that Ni could be defined in terms of ΔQ (or Qmax) and λ in a 3D failure map. This approach has been combined with a fracture mechanics approach by Quaresimin and Ricotta (2006) to obtain a predictive method explicitly modeling both the initiation and propagation phases. An alternative to this is to use one of the damage mechanics methods described in Sect. 4.4. Adhesives and polymer composites tend to exhibit fast fracture, and hence it may be preferable to design for a service life with no crack growth, using Gth as the design criteria. Abdel Wahab et al. (2001a) investigated this approach for bonded lap-strap joints. Elastic (G) and elastoplastic (J) fracture parameters, crack placement, and initial crack size were investigated. This approach has also been extended to samples subjected to environmental ageing by incorporating the method with a semicoupled transient hygromechanical finite element analysis (Ashcroft et al. 2003). Ashcroft and Shaw (2002) used Gth from testing DCB joints to predict the 106 cycle endurance limit in lap-strap and double-lap joints at different temperatures. The predictions were reasonable, apart from the double-lap joint tested at 90  C, which failed at a far lower load than predicted. This was attributed to accumulative creep in the double-lap joint, which could be seen in plots of displacement against cycles at constant load amplitude.

33

Fatigue Load Conditions

963

a

da/dN, mm/cycle

1.E-02 Change in fracture mechanism 1.E-03

1.E-04 Fracture in adhesive Mixed mechanism fracture 1.E-05

0

10

20 Crack size, mm

30

40

2.55

2.6

b −1.0

Log [da/dN ], mm/cycle

−1.5

Fracture in adhesive Mixed mechanism fracture Fracture in composite

−2.0 −2.5 −3.0 −3.5 −4.0 −4.5 2.4

2.45

2.5 Log [G ], J/m2

Fig. 12 (a) Fatigue crack growth rate as a function of crack length for fracture in the adhesive and a mixed mechanism fracture. (b) Fatigue crack growth plots for fracture in the adhesive layer in the composite and a mixed mechanism fracture

It is quite common in adhesively bonded joints to have more than one failure mechanism, for example, cohesive failure in the adhesive, failure at the interface, interlaminar failure, or matrix failure in composite substrates, occurring in the fatigue life of a joint. This will affect the fatigue crack growth, as illustrated in Fig. 12a. If the various mechanisms occur sequentially, then failure criteria for the joint can also be applied sequentially in a fatigue lifetime predictive methodology. However, if more than one mechanism is occurring at one time, anomalous fatigue crack growth can occur. This is illustrated in Fig. 12b for the case of a CFRP-epoxy lap-strap joint in which fracture is initially in the adhesive before developing into a

964

I. A. Ashcroft

mixed mechanism failure with an increasing proportion of failure in the composite adherend. It can be seen that in the mechanism fracture, the fatigue crack growth exhibits anomalous behavior as the crack growth rate decreases as strain energy release rate increases, which is not represented by a Paris-type crack growth law. Ashcroft et al. (2010) proposed a mixed mechanism fracture model of the following form:        da da da ¼ f Aa , , Ac dn m dn a dn c

(13)

where A indicates area fraction of failure mode, as seen in the fracture surface, and the subscripts m, a, and c represent mixed, adhesive, and composite, respectively. The simplest form of this law is a simple additive one, i.e.: 

da dn

 m

    da da ¼ Aa þ Ac dn a dn c

(14)

Equation (14) assumes there is a proportional relationship between the area fraction of a particular failure mode and its effect on the mixed FCG rate; however, this is not necessarily the case. It may be the case that one of the fracture mechanisms has a disproportionate effect on the mixed FCG rate. A simple way to accommodate this is to substitute the actual area fraction, Ai, in Eq. (14) with an effective area fraction, A0i , as illustrated in Eq. (15) 

da dn

 ¼ m

A0a

    da 0 da þ Ac dn a dn c

(15)

A method of mapping the actual area fraction to the effective area fraction is required, and Ashcroft et al. (2010) illustrated how power laws and polynomial relationships could be used for this purpose. It was shown that the proposed model was capable of predicting the anomalous fatigue crack growth behavior shown in Fig. 12.

33.4.4 Damage Mechanics Methods The fracture mechanics methods have the limitation that the initiation phase is not accounted for and that damage is through a single planar crack propagating through undamaged material, which may not physically represent the real behavior very well. The damage mechanics approach addresses some of these problems by allowing progressive degradation and failure to be modeled, thus representing both initiation and propagation phases. Continuum damage mechanics (CDM) requires a damage variable, D, to be defined as a measure of the severity of the material damage, where D is equal to 0 for undamaged material and 1 for fully damaged material. Abdel Wahab et al. (2001b) developed the following fatigue damage equation:

33

Fatigue Load Conditions

965

1 h iβþmþ1 D ¼ 1
1
Aðβ þ m þ 1ÞΔσ eq βþm Rv β=2 N

(16)

where Δσ eq is the von Mises stress range, Rv is the triaxiality function (which is the square of the ratio of the damage equivalent stress to the von Mises equivalent stress), m is the power constant in the Ramberg–Osgood equation, and A and β are experimentally determined damage parameters. The number of cycles to failure (Nf) can be determined from Eq. (16) using the condition at the fully damaged state, D = 1 at N = Nf, giving: β

Nf ¼

Δσ eq
β
m Rv
2 Að β þ m þ 1 Þ

(17)

Abdel Wahab et al. (2001b) found that the approach described above compared favorably with the fracture mechanics approach when applied to the constant amplitude fatigue of CFRP-epoxy double-lap joints. Abdel Wahab et al. (2010a) later extended this approach to the low cycle fatigue of bulk adhesive where damage evolution curves were derived assuming isotropic damage and a stress triaxiality function equal to one that agreed well with experimental measurements. In further work, Abdel Wahab et al. (2010b) applied the approach to single-lap joints which required determination of the triaxiality function to account for the multiaxial stress state in the joint, and it was seen that this value varied along the adhesive layer. The dependency of the triaxiality function on the joint type was further investigated by Abdel Wahab et al. (2011a, b). Although the CDM approach described above enabled the progressive degradation of the adhesive layer to be characterized, it did not allow the initiation and propagation phases of fatigue to be explicitly modeled. Shenoy et al. (2010a) used a damage mechanics-based approach to progressively model the initiation and evolution of damage in an adhesive joint, leading to crack formation and growth. In this approach, the damage rate dD/dN was assumed to be a power law function of the localized equivalent plastic strain range, Δep, i.e.:  m dD ¼ CD Δϵp D dN

(18)

where CD and mD are experimentally derived constants. The fatigue damage law was implemented in a finite element model. In each element, the rate of damage was determined from the finite element analysis using Eq. (18), and the element properties degraded as: E ¼ E0 ð1
DÞ σ yp ¼ σ yp0 ð1
DÞ β ¼ β 0 ð1
D Þ

(19)

966

I. A. Ashcroft

where E0, σ yp0, and β0 are Young’s modulus, yield stress, and plastic surface modifier constant for the Parabolic Mohr–Coulomb model, respectively, and D = 1 represents a fully damaged element, which was used to define the macrocrack length. The fatigue life was broken into a number of steps with a specified number of cycles, ΔN, and at each step, the increased damage determined using Shenoy et al. (2010a): Diþ1 ¼ Di þ

dD dN dN

(20)

showed that this method could be used to predict total-life plots, the fatigue initiation life, fatigue crack growth curves, and strength and stiffness wearout plots and could also be extended to variable amplitude fatigue (Shenoy et al. 2010b). Walander et al. (2014) experimentally studied Mode-I fatigue crack growth in rubber and polyurethane-based adhesives using a double-cantilever beam specimen. A damage growth law of a similar form to Eq. (20) related damage evolution with three material parameters: α, β, and σ th, which were determined from the experimental data. Good correlation between the experimental data and the proposed damage law for fatigue was reported. A similar approach was proposed by Khoramishad et al. (2010a) in which the rate of damage was related to the maximum principal strain, ϵmax, rather than the equivalent strain and a threshold strain, ϵth, was used rather than a plastic strain in order to define a strain below which fatigue damage does not occur.  ΔD C  ðϵmax
ϵth Þb , ϵmax > ϵth ¼ (21) ΔN 0, ϵmax  ϵth where C and b are experimentally determined material damage constants. In this case, the damage initiation and fracture was modeled using a bilinear cohesive zone model (as described in ▶ Chap. 25, “Numerical Approach: Finite Element Analysis”) and the damage variable was used to degrade the parameters of the cohesive model as illustrated in Fig. 13. This method was later extended to incorporate the effect of varying load ratio (Khoramishad et al. 2010b), variable amplitude fatigue (Khoramishad et al. 2011), and the effect of moisture degradation (Katnam et al. 2011). Pirondi and Moroni (2010) also proposed a method of predicting fatigue crack growth using cohesive zone elements within a finite element analysis. In this case, the damage rate was related to the strain energy release rate. Oinen and Marquis (2011) presented a damage model for the shear decohesion of combined clamped and adhesively bonded (i.e., hybrid) joints that incorporated a linearexponential cohesive zone model and a frictional contact model to account for a combined slip and decohesion response to loading.

33.5

Creep-Fatigue

Evidence of creep in the fatigue testing of bonded lap joints has been observed by a number of authors. It was seen in Sect. 4.3 that application of a standard fracture mechanics method can significantly overpredict the fatigue life if there is

Fatigue Load Conditions

967

Traction

33

Pure mode 1 I

eI

DF = 0

re

Pu

d mo

G 01c 0 G1c

Normal separation

n

tio

ara

r

ea

Sh

p se

1−1 G1c

DF = DF1−1

1−1 G1c

G 11c DF = DF1

1 G1c

e ag am ) e d (D F tigu ter Fa ame r pa Fig. 13 Degradation of cohesive zone model parameters with increasing fatigue damage (Khoramishad et al. 2010a)

significant accumulated creep. This is not surprising as polymeric adhesives will exhibit some degree of viscoelastic or viscoplastic behavior over a part of their operating range. The importance of the creep contribution to fatigue failure will depend on a number of factors such as the nature of the adhesive, the ambient temperature and humidity, the joint geometry, the mean fatigue load, and the frequency. Fatigue crack growth in such circumstances may still be adequately represented by LEFM or EPFM parameters; however, different fatigue crack growth curves will be required for different frequencies and temperatures, as shown in Fig. 14. However, if creep is significant, then fatigue crack growth may be better represented by a time-dependent fracture mechanics (TDFM)

968

I. A. Ashcroft

−3

Fig. 14 The effect of test frequency on fatigue crack growth plots for steel-epoxy DCB samples (Al-Ghamdi et al. 2003)

10 Hz 1 Hz

Log (da/dN), m/cycle

−4

0.1 Hz

−5

−6

−7

−8

−9 1.5

2

2.5

3

3.5

Log (Gmax), J/m2

parameter. Landes and Begley (1976) proposed a TDFM parameter analogous to Rice’s J-integral that they called C and defined it as follows: 

ð





C ¼ W dy
T i @ Γ

u_ gds @x

(22)

where



ϵð _ ij

W ¼

σ ij d_ϵij

(23)

0

Γ is a line contour taken counter clockwise from the lower crack surface to the upper crack surface, W is the strain energy rate density associated with the stress σ ij and strain rate ϵ_ ij, Ti is the traction vector defined by the outward normal n along Γ, u_ i is the displacement rate vector, and s is the arc length along the contour. C is valid

33

Fatigue Load Conditions

969

only for extensive steady-state creep condition, and Saxena (1986) proposed an alternative parameter, Ct, that was suitable for characterizing creep crack growth from small scale to extensive creep conditions. This was defined as:  _ 1  a, t, V c Þ Ct ¼
@U t B @a

(24)

where B is sample width, a is crack length, t is time, Ut is an instantaneous stress–power parameter, and V_ c is the load line deflection rate. In cyclic loading, Ct will vary with the applied load, and an average value of Ct, Ct (ave), may be used to characterize the fatigue-creep crack growth. Al-Ghamdi et al. (2004) proposed three methods of predicting creep-fatigue crack growth in bonded joints. In the first method, an appropriate fracture mechanics parameter is selected and plotted against the fatigue crack growth rate (FCGR = da/dN ) or the creep crack growth rate (CCGR = da/dt). A suitable crack growth law is fitted to the experimental data and crack growth law constants are determined at different temperatures and frequencies. Empirical interpolation can then be used to determine crack growth law constants at unknown temperatures and frequencies. This method can also be used to predict crack growth under conditions of varying frequency, load, and temperature. The second method was called the dominant damage approach and assumes that fatigue and creep are independent mechanisms and that crack growth is determined by whichever is dominant. In this case, crack growth should be predominantly either cycle or time dependent. For example, if creep is the dominant damage mechanism, then plots of da/dt against a suitable TDFM should be independent of frequency and can be used to predict the fatigue life in terms of time under load rather than number of cycles. The third method was termed the crack growth partitioning approach in which it is assumed that crack growth can be partitioned into separate cycle-dependent (fatigue) and time-dependent (creep) components, both of which contribute the total crack, as represented by the following equation: da ¼ dN



da dN

 þ fatigue

  1 da f dt creep

(26)

Al-Ghamdi et al. (2004) proposed the following form of Eq. (26): q

mCtðaveÞ da ¼ DðGmax Þn þ dN f

(27)

where the fatigue crack growth constants D and n are determined from high-frequency tests, and the creep crack growth constants m and q are determined from constant load crack growth tests. Under some conditions, this method still underpredicts the crack growth, in which case a creep-fatigue interaction term may be added:

970

I. A. Ashcroft q

mCtðaveÞ da ¼ DðGmax Þn þ þ CFint: dN f

(28)

where CFint: ¼ Rpf Ryc Cfc and Rf and Rc are scaling factors for the cyclic and timedependent components, respectively. Rf ¼

ðda=dN Þ ðda=dN Þ þ ðda=dtÞ=f

Rc ¼

ðda=dtÞ=f ðda=dN Þ þ ðda=dtÞ=f

and p, y, and Cfc are empirical constants.

33.6

Impact Fatigue

There is increasing interest in the effects of repetitive low-velocity impacts produced in components and structures by vibrating loads. This type of loading is known as impact fatigue. Typical plots of normalized force as a function of time from the impact fatigue testing of a bonded single-lap joint are shown in Fig. 15. It can be seen that the duration of the main peak is approximately 2 ms and that subsequent peaks are of significantly smaller magnitude. It can also be seen that the maximum force, Fmax, and the load time for the first peak, TF, decrease with the number of cycles, which is indicative of progressive fatigue damage in the adhesive. As the force is not constant in an impact fatigue test, it is useful to define a mean maximum force Fmax as:

1 Initial impact

0.8

Penultime impact

F (t )/F 0max

0.6

Last impact

0.4 0.2 0

−0.2 −0.4

0

1

2

3

4

5

6

Time, ms

Fig. 15 Variation in force and strain with time in a typical impact in the impact fatigue testing of a bonded lap joint (Casas-Rodriguez et al. 2007)

33

Fatigue Load Conditions

971

1 Impact fatigue Standard fatigue 0.8

0.6

0.4

0.2

0 1

10

100

1,000 10,000 100,000 1,000,000 Cycles to failure

Fig. 16 F–N diagrams for aluminum-epoxy single-lap joints in impact and standard fatigue (Casas-Rodriguez et al. 2007)

Fmax ¼

Nf 1 X Fmax N f i¼1 i

(29)

In Fig. 16, the normalized mean maximum force is plotted against the logarithm of cycle to failure and it can be seen that an approximately linear relationship is observed, with cycles to failure increasing as the force decreases, as expected. A plot for similar single-lap joints, manufactured with the same materials and processes, tested in standard fatigue with a constant force amplitude, sinusoidal waveform can also be seen in Fig. 16. It can be seen from this that the impact fatigue appears to be significantly more damaging than the standard fatigue. Casas-Rodriguez et al. (2008) also compared damage and crack growth in CFRP-epoxy lap-strap joints in standard and impact fatigue and found that the accumulated energy associated with damage in impact fatigue was significantly lower than that associated with similar damage in standard fatigue. It was also seen that the fracture surfaces, and hence, the fracture mechanisms, were quite different for the two types of loading. Ashcroft et al. (2011) also investigated the impact fatigue of notched bulk adhesive samples, developing an impact fatigue growth law based on the dynamic strain energy release rate.

33.7

Conclusion

It can be seen that a number of techniques have been used to model fatigue in bonded joints. Although most methods have limitations, all are useful in understanding and/or characterizing the fatigue response of bonded joints, and it is expected that further

972

I. A. Ashcroft

developments will continue in all the methods described. For instance, the traditional load–life approach is useful in characterizing global fatigue behavior, but is of little use in fatigue life prediction and provides no useful information on damage progression in the joint. However, combined with monitoring techniques, such as back-face strain or embedded sensors, and FEA, the method could potentially form the basis of an extremely powerful in-service damage monitoring technique for industry. The fracture mechanics approach is potentially a more flexible tool than the stress–life approach as it allows the progression of cracking to failure to be modeled and can be transferred to different sample geometries. Problems with the traditional fracture mechanics approach include: selection of initial crack size and crack path, selection of appropriate failure criteria, load history, and creep effects. Also the fracture mechanics approach does not accurately represent the accumulation and progression of damage observed experimentally in many cases. However, recent modifications to the standard fracture mechanics method have seen many of these limitations tackled. In future developments, it is expected that fatigue studies of bonded joints will continue to increase our knowledge of how fatigue damage forms and progresses in bonded joints and this will feed into the models being developed. In addition to further developments and extension to the approaches described above, it is expected that an area for huge potential advances is in the incorporation of increasingly sophisticated damage growth laws into finite element analysis models to develop a more mechanistically accurate representation of fatigue in bonded joint than is currently available. It is also hoped that many of the techniques described above will reach sufficient maturity to form the basis of useful tools for industry, both in the initial design and in-service monitoring of bonded joints in structural applications subjected to cyclic loading.

References Abdel Wahab MM, Ashcroft IA, Crocombe AD, Hughes DJ, Shaw SJ (2001a) Compos Part A 32:59 Abdel Wahab MM, Ashcroft IA, Crocombe AD, Hughes DJ, Shaw SJ (2001b) J Adhes Sci Technol 15:763 Abdel Wahab MM, Ashcroft IA, Crocombe AD, Smith PA (2004) Compos Part A 35:213 Abdel Wahab MM, Hilmy I, Ashcroft IA, Crocombe AD (2010a) J Adhes Sci Technol 24:305 Abdel Wahab MM, Hilmy I, Ashcroft IA, Crocombe AD (2010b) J Adhes Sci Technol 24:325 Abdel Wahab MM, Hilmy I, Ashcroft IA, Crocombe AD (2011a) J Adhes Sci Technol 25:903 Abdel Wahab MM, Hilmy I, Ashcroft IA, Crocombe AD (2011b) J Adhes Sci Technol 25:925 Al-Ghamdi AH, Ashcroft IA, Crocombe AD, Abdel Wahab MM (2003) J Adhes 79:1161 Al-Ghamdi AH, Ashcroft IA, Crocombe AD, Abdel-Wahab MM (2004) Proceedings of the 7th international conference on structural adhesives in engineering. IOM Communications, London, pp 22–25 Ashcroft IA (2004) J Strain Anal 39:707 Ashcroft IA, Shaw SJ (2002) Int J Adhes Adhes 22:151 Ashcroft IA, Abdel Wahab MM, Crocombe AD, Hughes DJ, Shaw SJ (2001a) Compos Part A 32:45 Ashcroft IA, Abdel-Wahab MM, Crocombe AD, Hughes DJ, Shaw SJ (2001b) J Adhes 75:61 Ashcroft IA, Abdel-Wahab MM, Crocombe AD (2003) Mech Adv Mater Struct 10:227

33

Fatigue Load Conditions

973

Ashcroft IA, Casas-Rodriguez JP, Silberschmidt VV (2010) J Adhes 86:522 Ashcroft IA, Silberschmidt VV, Echard B, Casas-Rodrguez JP (2011) Shock Vib 18:157 Casas-Rodriguez JP, Ashcroft IA, Silberschmidt VV (2007) J Sound Vib 308:467 Casas-Rodriguez JP, Ashcroft IA, Silberschmidt VV (2008) Comp Sci Technol 68:2663 Coffin LF (1954) Trans Am Soc Mech Eng 76:931 Costa M, Viana G, da Silva LFM, Campilho RDSG (2017) J Adhes 93:127 Crocombe AD, Richardson G (1999) Int J Adhes Adhes 19:19 Datla NV, Ulicny J, Carlson B, Papini M, Spelt JK (2011a) Eng Fract Mech 78:1125 Datla NV, Ulicny J, Carlson B, Papini M, Spelt JK (2011b) Int J Adhes Adhes 31:88 Dowling NE (1999) Mechanical behaviour of materials. Prentice Hall, New Jersey, pp 390–392 Erpolat S, Ashcroft IA, Crocombe AD, Abdel Wahab MM (2004a) Int J Fatigue 26:1189 Erpolat S, Ashcroft IA, Crocombe AD, Abdel Wahab MM (2004b) Comp A 35:1175 Farrow IR (1989) Damage accumulation and degradation of composite laminates under aircraft service loading: Assessment and prediction, Vol. I & II. PhD thesis, Cranfield Institute of Technology Gomatam R, Sancaktar E (2006) J Adhes Sci Technol 20:69 Griffith AA (1921) Phil Trans R Soc A221:163 Irwin GR (1958) Fracture. In: Flugge S (ed) Handbuch der physic VI. Springer, Berlin, pp 551–590 Katnam KK, Crocombe AD, Sugiman H, Khoramishad H, Ashcroft IA (2011) Int J Dam Mech 20:1217 Khoramishad H, Crocombe AD, Katnam K, Ashcroft IA (2010a) Int J Fatigue 32:1146 Khoramishad H, Crocombe AD, Katnam K, Ashcroft IA (2010b) Int J Adhes Adhes 32:1278 Khoramishad H, Crocombe AD, Katnam K, Ashcroft IA (2011) Eng Fract Mech 78:3212 Landes JD, Begley JA (1976) Mechanics of crack growth, ASTM STP 590. American Society for Testing and Materials, Philadelphia, pp 128–148 Levebvre DR, Dillard DA (1999) J Adhes 70:119 Little MSG (1999) Durability of structural adhesive joints. Ph.D. thesis, London, Imperial College of Science, Technology and Medicine Manson SS (1954) Behaviour of materials under conditions of thermal stress. In: National Advisory Commission on aeronautics. Report 1170. Lewis Flight Propulsion Laboratory, Cleveland, pp 317–350 Miner MA (1945) J Appl Mech 12:64 Nisitani H, Nakamura K (1982) Trans Jpn Soc Mech Eng 48:990 Oinen A, Marquis G (2011) Eng Fract Mech 78:163 Palmgren A (1924) Z Ver Dtsch Ing 68:339 Paris PC, Gomez MP, Anderson WE (1961) Trend Eng 13:9 Pirondi A, Moroni F (2010) J Adhes 86:501 Quaresimin M, Ricotta M (2006) Comp Sci Technol 66:647 Rice JR (1968) J Appl Mech 35:379 Saxena A (1986) In: Underwood JH et al (eds) Fracture mechanics: seventeenth volume, ASTM STP 905. American Society for Testing and Materials, Philadelphia, pp 185–201 Schaff JR, Davidson BD (1997a) J Comp Mater 31:128 Schaff JR, Davidson BD (1997b) J Comp Mater 31:158 Shenoy V, Ashcroft IA, Critchlow GW, Crocombe AD, Abdel Wahab MM (2009a) Int J Adhes Adhes 29:361 Shenoy V, Ashcroft IA, Critchlow GW, Crocombe AD, Abdel Wahab MM (2009b) Int J Fatigue 31:820 Shenoy V, Ashcroft IA, Critchlow GW, Crocombe AD, Abdel Wahab MM (2009c) Int J Adhes Adhes 29:639 Shenoy V, Ashcroft IA, Critchlow GW, Crocombe AD (2010a) Int J Fatigue 32:1278 Shenoy V, Ashcroft IA, Critchlow GW, Crocombe AD (2010b) Eng Fract Mech 77:1073 Walander T, Eklind A, Carlberger T, Stigh U (2014) Proc Math Sci 3:829

Creep Load Conditions

34

Paul Ludwig Geiß and Melanie Schumann

Contents 34.1 34.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Viscoelastic Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.2.1 Creep Under Static Stress Conditions (Retardation) . . . . . . . . . . . . . . . . . . . . . . . . . . 34.2.2 Creep Under Static Strain Conditions (Relaxation) . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.2.3 Generalized Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.3 Superposition Principles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.3.1 Boltzmann Superposition Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.3.2 Correspondence Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.3.3 Time-Temperature Superposition Principle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.4 Experimental Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.4.1 Accessing Creep Properties of Adhesives by Rheometric Test Procedures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.4.2 Retardation Test Methods with Adhesively Bonded Joints . . . . . . . . . . . . . . . . . . 34.4.3 Relaxation Test Methods with Adhesively Bonded Joints . . . . . . . . . . . . . . . . . . . 34.4.4 Relation of Stress Relaxation and Static Load Creep . . . . . . . . . . . . . . . . . . . . . . . . . 34.5 Predictive Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

976 977 979 982 985 986 986 987 988 989 989 991 1000 1003 1005 1005 1006

Abstract

Polymer-based adhesives that have been especially designed to feature a high degree of toughness, flexibility, or even deliberately viscoelastic properties exhibit a pronounced time-dependent stress-strain behavior under static mechanical load conditions. Therefore, a suitable mechanical approach to describe and experimentally access phenomena related to creep in polymers and adhesives

P. L. Geiß (*) · M. Schumann Faculty of Mechanical and Process Engineering, Workgroup Materials and Surface Technologies (AWOK), University of Kaiserslautern, Kaiserslautern, Germany e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_34

975

976

P. L. Geiß and M. Schumann

needs to consider elastic as well as viscous material properties. Creep occurs at load levels below the yield strength, and particularly, creep of viscoelastic materials by definition is assigned to reversible deformation being able to recover after unloading. In technical practice, the term creep is often being used to describe arbitrary time-dependent strain including viscous and elastic-viscous behavior. The concept of generating mechanical substitute models dates back to the late eighteenth century. The Maxwell model, for example, represents a serial connection of a spring-(elastic) and a dashpot-(viscous) element. Other viscoelastic models are discussed in terms of their creep (retardation) strain and relaxation stress response to a corresponding stress or strain input in the generalized form of a step-unit function. The resulting constitutive laws for stressstrain-time relationship under static stress or strain conditions can be applied in the limits of linear elasticity where the Boltzmann superposition principle and the correspondence principle are valid. The options for experimentally accessing creep and relaxation phenomena range from standardized procedures to advanced test methods, which are discussed in the context of experimental data. Applying the theoretical background of the standard models for viscoelasticity to the experimental results can improve the reliability of predictive methods intending to extrapolate results from short-term experiments to a long-term timescale.

34.1

Introduction

In the seventeenth century, the British physicist Robert Hooke (Purrington 2009) defined the basic relation between stress and strain as “Ut tensio, sic vis.,” meaning extension is proportional to force. This so-called Hooke’s law is valid for many inorganic materials in the linear elastic region, where force-induced deformations are reversible and the material will completely return to its initial shape after removing the force without any delay or residual plastic deformations. In the linear elastic region, constitutive material models are able to correlate strain to stress without the need to consider the history of stress and strain events a specific object or assembly has been subjected to previously. If an elastoplastic material is subjected to mechanical stress, it will also respond with instantaneous elastic deformation but above the yield strength deform plastically at a deformation rate governed by the process of deformation. Again, the behavior is not considered to be time dependent. In most crystalline or amorphous inorganic materials, strong metallic bonds or ionic bonds are responsible for a dense structure and high level of rigidity. After exceeding the linear elastic stress-strain area, in metal alloys plastic deformation is related to the generation and movement of dislocations in the crystalline structure such as edge dislocations, screw dislocations, and sliding along grain boundaries. Amorphous inorganic glassy materials generally exhibit brittle fracture properties on a short-term timescale but are known to also deform plastically and time dependently under long-term static load conditions or at elevated temperatures. Metals and alloys also respond to elevated temperatures with a reduction of the linear elastic

34

Creep Load Conditions

977

deformation threshold generally referred to as the yield strength, especially once the temperature approaches or has exceeded the recrystallization temperature. The mechanisms of plastic deformation and creep under static load conditions in organic polymer materials such as adhesives are somehow different from that and need to be considered when using structural adhesive bonds in the presence of so-called dead loads which frequently may occur due to the mere influence of gravity of the mass of a certain structure. In 1921, Hermann Staudinger (Priesner 1983) postulated that polymers consist of chain-like macromolecules with typically carbon-carbon bonds as the repeating unit responsible for creating the backbone of a polymer chain. The average molecular weight, its statistical distribution, the presence of secondary bonds between adjacent chains, the degree of entanglement, and crosslinking by side chains are the dominant parameters influencing the mechanical properties of polymer materials. Depending on the degree of geometric order, polymer materials may also develop crystallinity by self-organized alignment of chains or segments into stacks and on a superordinated level into spherulites. Examples of semicrystalline polymer materials include isotactic or syndiotactic polypropylene. Most adhesive polymer materials, however, show little or no macroscopic level of crystallinity. Plastic deformation in polymers can either take place by the creation of voids (crazing), by the rotation between chain segments, or by sliding between adjacent polymer chains. While crazing can be considered as a specific type of cracking on a micro- and nanoscopic size level, segment rotation and chain sliding are reversible physical phenomenon. Micro-Brownian movement is related to a change in conformation without dislocating the center of gravity of the polymer chain, while strain by macroBrownian is characterized by a local movement of the center of gravity of a polymer chain under the effect of, e.g., mechanical stress. As with the mechanisms of creep described above for inorganic crystalline or amorphous materials, the inelastic deformation of polymers is strongly related to the temperature, since in addition to mechanical stress, temperature serves as a source of activation energy to stimulate rotation and related Brownian movement on a nanoscopic level. In the range of the glass transition area, the ability of chain segments to rotate is strongly promoted with increasing temperature. Above the so-called glassy state of a polymer, the linear elastic properties diminish, and the relation between stress and strain can no longer be approximated by Hooke’s law with adequate accuracy. This is the reason why several viscoelastic models have been subsequently developed, to provide means of describing, analyzing, and predicting the mechanical behavior of materials exhibiting an elastic and viscous time- and temperature-dependent stress-strain relationship.

34.2

Viscoelastic Models

The term viscoelasticity combines viscous and elastic stress-strain flow characteristics. If material behavior is dominated by viscous flow, it is generally referred to as a fluid, whereas if the elastic properties dominate the mechanical properties of a material, it is considered to be solid. Most adhesives are applied in a liquid or

978

P. L. Geiß and M. Schumann

Time

Instantaneous elastic strain

Strain

Fig. 1 Schematic strain response of viscoelastic materials to a period of stress

Stress

pasty condition to allow wetting and promote spreading and then are required to phase change into a solid. In the liquid state, rheology provides the methods to differentiate between elastic and viscous flow characteristics, while dynamic mechanical analysis of cured adhesive polymers uses similar principles to access elastic and viscous parameters of the stress-strain response. If no stress threshold is required to initiate viscous flow, the behavior is called viscoelastic; if a certain stress threshold (yield strength) needs to be overcome before the beginning of time-dependent viscous flow, the behavior is addressed as viscoplastic. When subjected to constant stress for a certain period of time, three basic parts of strain response may typically be observed. Immediately after applying a certain amount of stress, instantaneous elastic strain takes place followed by delayed elastic strain. Delayed elastic strain will completely recover after removing stress. It is often referred to as primary creep. Following delayed elastic strain, viscous flow can often be observed for polymer adhesive materials with low crosslinking density. Strain caused by viscous flow leads to permanent deformation and is generally referred to as non-recoverable strain or secondary creep. Figure 1 illustrates schematically the strain response of such a material to a limited period of stress. The level of instantaneous elastic strain upon loading corresponds to the instantaneous strain recovery after removing the load. The aim of viscoelastic models is to use a combination of mechanical elastic and viscous standard elements to generate a comprehensive mechanical substitute system showing more or less simplified viscoelastic properties and thus to provide parameters for building constitutive equations suitable for the prediction of viscoelastic behavior and the characterization of viscoelastic materials. James Clerk Maxwell (Niven 2003) proposed in 1867 a simplified mechanical model consisting of a viscous dashpot in a serial connection to an elastic spring to describe viscoelastic material properties.

Delayed elastic strain

Unrecoverable strain

Time

34

Creep Load Conditions

979

Table 1 Comparison of various viscoelastic models and related differential equations, s: stress parameter, E: strain parameter, E: elastic parameter, : viscous parameter, l: relaxation time HOOKE-spring element (linear elasticity) σ = E0  e NEWTON-fluid (linear viscosity) σ ¼ η0  e_ KELVIN-VOIGT-solid σ ¼ E0  e þ η0  e_ MAXWELL-fluid σ þ λ0  σ_ ¼ η0  e_

ðλ0 ¼ η0 =E0 Þ

BURGERS-fluid (four parameter

 model)
 2 € ¼ η1  e_ þ η1Eη2 2  e€ σ þ Eη11 þ Eη12 þ Eη22  σ_ þ Eη11 η E2  σ

η1

E1

η2

E2

The spring element is considered to behave according to Hooke’s law in a linear elastic manner, while the mechanical behavior of the dashpot equates linear viscous flow of a Newtonian fluid. The following models of, e.g., Lord Kelvin and Woldemar Voigt (Thomson 1875) suggested a parallel connection of a dashpot and a spring element, while the Burgers model (Burgers 1935) named after the Dutch physicist Johannes Martinus Burgers consists of a Maxwell element and a KelvinVoigt element in series. Further generalized Maxwell or Kelvin-Voigt models connected an arbitrary or even infinite number of single elements to generate a spectral distribution of relaxation and retardation properties. Table 1 gives an overview of the most commonly used mechanical models for viscoelasticity and their corresponding differential equations.

34.2.1 Creep Under Static Stress Conditions (Retardation) The time-dependent increase in strain of viscoelastic polymers under static stress conditions is generally referred to as creep or retardation. To describe creep behavior of adhesive joints under static stress conditions using viscoelastic models, it is assumed that a specimen be subjected at a certain time t = 0 to a defined constant level of tensile stress σ = σ 0 or shear stress τ = τ0. The stress-strain history then provides the boundary conditions for solving the differential equation in the following way: e ¼ 0 for t < 0 σ ¼ 0 for t < 0 σ ¼ σ 0 ¼ constant for t  0

(1)

980

P. L. Geiß and M. Schumann

Creep Response of the Maxwell Model The single spring element of the Maxwell fluid (Niven 2003) does not exhibit any time-related strain behavior, and the response to stress σ 0 at t = 0 can be expressed as eðtÞspring ¼

σ0 E

(2)

The parameter E represents the elastic constant of the spring element. The time-dependent strain of the dashpot element due to viscous flow under constant stress can be described as follows: eðtÞdashpot ¼

σ0 t η

(3)

with η being a parameter related to the Newtonian viscous flow behavior of the dashpot element and the stress response of this dashpot being proportional to the strain rate. The overall time-dependent strain of the Maxwell fluid thus can be calculated as eðt Þ ¼

σ0 σ0 tþ η E

(4)

Dividing the viscous parameter η with a dimension unit of Pas by E with a dimension unit of Pa provides the parameter λ generally referred to as the relaxation time. The time-dependent retardation strain response of the Maxwell model can then be expressed as eðt Þ ¼

σ0 σ0 σ0
t tþ ¼  1þ ¼ σ 0  FðtÞ with λ ¼ η=E λ η E E

(5)

Strain

Fig. 2 Stress history and corresponding creep response for viscoelastic materials according to the Maxwell spring and dashpot model

Stress

The term F(t) generally is referred to as the creep function. After a certain time of creep corresponding to λ has passed, the ratio t/λ equals 1, and the viscous strain matches the initial elastic strain. The right side of Fig. 2 illustrates the resulting strain response for the schematic stress history indicated on the left side of Fig. 2. The Maxwell model allows the approximation of elastic and creep strain under static stress as a function of time by assigning elastic compliance leading to instant

0

Time

0

Time

34

Creep Load Conditions

981

elastic strain to the spring part of the model. The flow due to creep is represented by the linear slope related to the single dashpot element with linear viscous flow properties. In practice, viscoelastic adhesive polymers do not show a discontinuous sharp bend but rather a smooth transition from the initial elastic strain to steady-state creep. A further refinement of the viscoelastic model is therefore needed to create a model converging with actually observed viscoelastic behavior.

Creep Response of the Kelvin-Voigt Model The Kelvin-Voigt model (Thomson 1875) addresses this issue by using a parallel connection of a spring and a dashpot. The boundary condition for solving the corresponding differential equation is that spring and dashpot at all times share the same level of strain: eðSpringÞ ¼ eðDashpotÞ

(6)

The applied stress is shared by both the spring and the dashpot element: σ 0 ¼ σ ðSpringÞ þ σ ðDashpotÞ

(7)

The solution of the differential equation for the time-dependent creep response under these boundary conditions can be expressed as eðt Þ ¼

 σ0
t  1  e  λ ¼ σ 0  Fð t Þ E

(8)

Strain

Fig. 3 Schematic creep response to a stress step function for viscoelastic materials according to the Kelvin-Voigt model

Stress

where again the ratio λ = η/E is the relaxation time and F(t) is the creep function of the Kelvin-Voigt model. When a period of time with a duration of λ has passed, the exponent t/λ equals 1 and the factor e1 attains a value of  0.37. Accordingly, creep strain at relaxation time has reached  63% of the infinite ultimate value for this specific viscoelastic model. Figure 3 illustrates the schematic timedependent creep response of the Kelvin-Voigt model for a similar load history like in Fig. 2. Compared to the previously discussed Maxwell model and actual experimental data, the advantage of the Kelvin-Voigt model is that the strain exponentially approaches an ultimate strain limit without any discontinuities. Steady-state longterm creep and initial elastic strain, however, are not represented in the overall strain response according to the Kelvin-Voigt model.

0

Time

0

Time

982

P. L. Geiß and M. Schumann

Creep Response of the Burgers Model The combination of both the Kelvin-Voigt and the Maxwell model in the Burgers model (Burgers 1935) leads to an even more realistic approach. The solution of the corresponding differential equation can be obtained by simply adding the KelvinVoigt and Maxwell shares of the time-dependent strain response. 

 tE2  1 1 1
þ tþ 1  e η21 E1 η 1 E2   t 1 1 1
λ2 ¼ σ0  þ tþ 1e ¼ σ 0  Fð t Þ E1 E1  λ 1 E2

eðt Þ ¼ σ 0 

(9)

with the relaxation times λ1 = η1/E1 and λ2 = η2/E2 and F(t) being the creep function of the Burgers model. The assignment of the individual spring and dashpot elements according to this model has been shown in Table 1. The schematic strain retardation response of the Burgers model to a step-function stress input is illustrated in Fig. 4. The creep response according to the Burgers model in Fig. 4 covers all elementary aspects of time-dependent viscoelastic behavior including instantaneous elastic strain, secondary steady-state creep in the long-term area, and a delayed elastic strain transition behavior that can be, e.g., fitted to experimental data according to the choice of the η1, E1, η2, and E2 parameters. Increasing the number of interconnected spring and dashpot elements in building viscoelastic models will increase the degrees of freedom in fitting the models to experimental data and for generalized models based on an infinite number of single elements will match the continuum mechanics approach of solid and fluid dynamics. The challenge in selecting an appropriate viscoelastic model for a specific area of practical application is to obtain the simplest relation but still to achieve the desired level of accuracy.

34.2.2 Creep Under Static Strain Conditions (Relaxation) If a bonded assembly is subjected to a defined strain condition like a joint in a confined construction, creep-related phenomena in the adhesive polymer would lead

Primary creep (delayed elastic strain)

Strain

Stress

Secondary creep

Instantaneous elastic strain 0

Time

0

Time

Fig. 4 Creep response for viscoelastic materials according to the Burgers model

34

Creep Load Conditions

983

to a time-dependent reduction of stress. This process is generally referred to as relaxation. In describing relaxation behavior of adhesive joints using viscoelastic models, it is assumed that a specimen is subjected at a certain time t = 0 to a defined constant level of strain ε = ε0 in tensile load conditions or tan γ = tan γ0 in the case of shearloaded joints. The strain-stress history again provides the boundary conditions for the differential equation as follows: e ¼ 0 for t < 0 σ ¼ 0 for t < 0 e ¼ e0 ¼ constant for t  0

(10)

Relaxation of the Maxwell Model Relaxation of the simple Maxwell fluid model will decrease the initial level of stress caused by the elasticity of the spring element by viscous elongation of the dashpot until all stress has been completely relieved. The solution of the first-order linear differential equation in the case of ε = ε0 = const describes the time-dependent relaxation of stress for the Maxwell model as follows:
t σ ðtÞ ¼ E  e0  eλ ¼ e0  RðtÞ

(11)

with E being a parameter representing the elastic constant of the spring element, λ = η/E being the relaxation time, and R(t) being the relaxation function of the Maxwell model. The characteristic time-dependent stress relaxation response of the Maxwell model to a step-function strain input is illustrated in Fig. 5.

Relaxation of the Kelvin-Voigt Model Since the Kelvin-Voigt model consists of a parallel connection of spring and dashpot, the initial level of stress depends on the viscosity η assigned to the dashpot and the strain rate when applying initial strain at t = 0.

Strain

Stress

E·e0

0

Time

0

Fig. 5 Relaxation of stress according to the simple Maxwell model

Time

984

P. L. Geiß and M. Schumann

The stress response to a step-function type of strain signal as indicated on the left side of Fig. 5 would jump to an infinite level. Therefore, this model is not able to give a reasonable approximation of relaxation phenomenon under these preconditions.

Relaxation of the Burgers Model The extension to the four-parameter Burgers model provides again a more detailed description of relaxation in viscoelastic polymer materials. The initial differential equation and the assignments of parameters are as follows:  σþ

η1 η1 η2 þ þ E1 E 2 E2



  σ_ þ

η1  η2 E1  E2



  σ€ ¼ η1  e_ þ

η1  η2 E2

  e€

(12)

The initial stress upon application of a constant strain is governed by the Maxwell spring element E1, and relaxation on a long-term timescale is dominated by the Maxwell dashpot η1, while the Kelvin-Voigt section with E2 and η2 governs the delayed elastic relaxation. Solving the differential equation by Laplace transformation is, e.g., comprehensively described in Betten (2008). Another possibility to work with the parameters if the viscoelastic Burgers models in evaluating and predicting relaxation phenomena is by using a recursive definition of the relaxation function σ(t). Therefore, the total period of time to be considered is first divided into nt incremental steps, and the calculation of strain for each consecutive step is based on the value of the preceding one. Starting at t = 0, again assuming a strain-stress sequence according to a strain step function for relaxation experiments with e ¼ 0 for t < 0 σ ¼ 0 for t < 0 e ¼ e0 ¼ constant for t  0

(13)

both dashpots η1 and η2 will give a rigid response to an instantaneous strain input e0. In consequence, the total initial strain of the Burgers model in the first incremental step is governed by the spring element E1 and can be calculated as σ ð nt ¼ 1Þ ¼ E 1  e 0

(14)

During the next incremental step, σ (nt = 1) will cause the dashpot η1 to expand to a strain of eη1 ðnt ¼ 2Þ ¼

σ ð nt ¼ 1Þ  Δt η1

(15)

During this step, no load has been transferred to the second spring element E2, and the increase of strain in dashpot η2 can be approximated in the same manner: eη2 ðnt ¼ 2Þ ¼

σ ð nt ¼ 1Þ  Δt η2

(16)

34

Creep Load Conditions

985

The total viscous strain eη1(nt = 2) + eη2(nt = 2) will reduce the elongation of the E1 spring element to a value of e0  (eη1(nt = 2) + eη2(nt = 2)) and reduce the overall stress in the next step to   σ ðnt ¼ 2Þ ¼ E1  e0  eη1 ðnt ¼ 2Þ  eη2 ðnt ¼ 2Þ

(17)

The last incremental effect to be considered according to this recursive model is that the strain in dashpot η2 has transferred a part of the load to the spring element E2. For the calculation of the next step, this spring element carries the load σ E2 ðnt ¼ 3Þ ¼ E2  eη2 ðnt ¼ 2ÞÞ

(18)

The stress causing dashpot η2 to further expand is reduced accordingly to σ η2 ðnt ¼ 3Þ ¼ σ ðnt ¼ 2Þ  E2  eη2 ðnt ¼ 2Þ

(19)

The resulting increase in strain of dashpot η2 for the next step will then be eη2 ðnt ¼ 3Þ ¼

σ ðnt ¼ 2Þ  E2  eη2 ðnt ¼ 2Þ  Δt η2

(20)

Following the third increment, the calculation is repetitive, and the accuracy depends on the choice of the number of increments Δt in relation to the total time span to be considered. The recursive definition can easily be implemented into computer-based spreadsheet calculation tools and be extended to other springdashpot combinations if desired.

34.2.3 Generalized Models The greater the number of connected spring and dashpot elements in a viscoelastic model, the better a fitting to the material behavior of a specific polymer can be accomplished. In the generalized Maxwell model, a spring and dashpot are connected to an arbitrary number of Maxwell elements in parallel. In the generalized Kelvin-Voigt model, a Maxwell element is connected to an arbitrary number of Kelvin-Voigt elements in series. The notional transition from a finite number of elements to an infinite number (i.e., a spectrum) of elements leads to models where Γ(λ) expresses the probability density of relaxation times in the generalized Maxwell model. The relaxation function of the generalized Maxwell spectrum then is 1 ð

RðtÞ ¼ 0

Γ ðλÞ  eðt=λÞ dλ

(21)

986

P. L. Geiß and M. Schumann

In a similar way, for an infinite number of elements in the generalized KelvinVoigt model, f (λ) may be used to express the probability density of retardation times, and the creep function F(t) for the spectrum in the case of the generalized KelvinVoigt model can be written as 1 ð

Fð t Þ ¼


 f ðλÞ  1  eðt=λÞ dλ

(22)

0

The functions Γ(λ) and f (λ) characterize the retardation and respectively the relaxation spectra and thus represent the distribution of retardation and relaxation time values in a specific material. Further details are given in Haddad (1995) and de With (2006).

34.3

Superposition Principles

Describing and predicting viscoelastic properties of polymer materials or adhesively bonded joints on the basis of analytical mathematical equations is only justified in the limits of linear viscoelasticity. Linear viscoelasticity is typically limited to strain levels below 0.5%. Furthermore, linear viscoelastic behavior is associated to the Boltzmann superposition principle, the correspondence principle, and the principle of time-temperature superposition.

34.3.1 Boltzmann Superposition Principle According to the Boltzmann superposition principle, the time-dependent strain caused by two individual stress incidents σ 1(t) and σ 2(t) can be calculated as ε1(t) + ε2(t) if σ 1(t) would cause ε1(t) and σ 2(t) would cause ε2(t). In consequence, according to the Boltzmann superposition principle in the linear viscoelastic range, strain responses from subsequent stress histories may simply be summarized if according to the criterion of proportionality at any instant of time, the strain response behaves proportional to the stress level: f ½c  σ ðtÞ ¼ c  f ½σ ðtÞðc ¼ constantÞ

(23)

The criterion of superposition of input histories can be expressed accordingly as f ½σ 1 ðt1 Þ, σ 2 ðt2 Þ, . . . , σ n ðtn Þ ¼ f ½σ 1 ðt1 Þ þ f ½σ 2 ðt2 Þ þ . . . þ f ½σ n ðtn Þ

(24)

Figure 6 illustrates how the Boltzmann superposition principle will affect the total creep strain in a step-loaded creep experiment. When at a time t1 the initial stress σ 1 is applied, then the time-dependent strain e1(t) is induced in

34

Creep Load Conditions

987

e3 Strain

Stress

s3 s2 s1

t1

t2

e2 e1

t1

t3

t2

t3

Time

Time

Fig. 6 Effect of the Boltzmann superposition principle on creep strain in a step-loading creep experiment

e1 ðtÞ ¼ σ 1  J ðtÞ

(25)

According to linear viscoelasticity, the compliance J(t) is independent of stress, i.e., J(t) is the same for different stress levels at a particular time. Once the following stress increment σ 2–σ 1 is applied at time t2, the additional strain increase due to stress increment σ 2–σ 1 is e2 ðtÞ ¼ ðσ 2  σ 1 Þ  J ðt  t2 Þ

(26)

In a similar way, the additional strain increase due to σ 3σ 2 can be calculated as e3 ðtÞ ¼ ðσ 3  σ 2 Þ  ðt  t3 Þ

(27)

34.3.2 Correspondence Principle The correspondence principle following the Boltzmann superposition principle allows the conversion of the common mechanical relationships of linear elasticity theory into linear viscoelasticity simply by replacing σ by time-dependent σ(t) and ε by time-dependent ε(t). Young’s modulus E is accordingly transformed to the creep modulus EC(t) = σ/e(t) or the creep compliance J(t) = e(t)/σ, respectively, and the relaxation modulus ER(t) = σ(t)/e. These time-dependent parameters can be determined from tensile creep and relaxation experiments. In compression or shear mode, the corresponding parameters of moduli are calculated in a similar manner. In linear viscoelasticity, the creep function and the relaxation function are interrelated, and each would permit the derivation of the viscoelastic constitutive relation under the condition that the principle of time invariance can be applied, meaning that the material is influenced only by σ(t) and ε(t) and no other influencing variables are being effective. In summary, in the idealized case of the isothermal linear viscoelasticity under the conditions of static stress or strain, the linear viscoelastic behavior of a material may adequately be described by the creep function F(t), the relaxation function R(t), the retardation spectrum, or the relaxation spectrum.

988

P. L. Geiß and M. Schumann

34.3.3 Time-Temperature Superposition Principle The time-temperature superposition principle is based on the assumption that in the linear viscoelastic region, the different mechanical parameters exhibit similar temperature dependence. Under this precondition, an increase in temperature will cause a similar change in mechanical properties as an increase in time for the timedependent mechanical viscoelastic parameters. In the case of a cyclic mechanical stress or strain excitation of a viscoelastic material, an increase of temperature will have a similar effect as a reduction of frequency according to the principle of timetemperature superposition. The principle of time-temperature superposition therefore is commonly applied on, e.g., data obtained by dynamic mechanical thermal analysis. The dynamic mechanical thermal analysis (DMTA) provides an effective method for measuring transitions due to morphological or chemical changes, or both, in a polymer as it is heated/cooled through a specified temperature range. For the dynamic mechanical thermal analysis (e.g., according to ASTM D 4065), a specimen of known geometry is subjected to a mechanical oscillation either at fixed or natural resonant frequencies. Elastic (storage) and viscous (loss) moduli of the specimen are then measured at varying temperatures and frequencies. Plots of the elastic modulus (E0 , G0 ) related to the spring properties in the viscoelastic behavior or the loss modulus (E00 , G00 ) related to the viscous dashpot properties are indicative of the viscoelastic characteristics of the specimen. The ratio E00 /E0 or G00 /G0 is defined as the loss factor tan δ which provides a measure of the mechanical damping due to inner friction of a viscoelastic polymer material. Rapid changes in viscoelastic properties in a particular range of temperatures, times, or frequencies are generally referred to as transition regions. Dynamic mechanical testing provides a method for determining elastic and loss moduli as a function of temperature, frequency or time, or both. A plot of the elastic modulus and loss modulus of material versus temperature provides a fingerprint of elastic and viscous properties as a function of temperature or frequency. Experimental data obtained during such a dynamic mechanical analysis at different temperatures and frequencies can be converted using a shift function a = f(T ) with T being the temperature, to merge temperature-dependent data into a master curve. This master curve then, e.g., represents the mechanical properties for a chosen temperature T0 at time or frequency levels beyond the range that had been experimentally applied in collecting the underlying data. The shift function was originally published by Williams, Landel, and Ferry (Ferry 1970) and proved to be applicable for most elastic and amorphous polymers and in its general form is fitted to give best conformity of the master curve by two parameters C1 and C2: log aT ¼

C1  ðT  T 0 Þ C2 þ T  T 0

(28)

If the glass transition temperature TG is chosen as T0, for many polymer material parameters, C1 = 17.44 and C2 = 51.6 K have provided a good approximation.

34

Creep Load Conditions

989

This method can also be applied to the extrapolation of time- and temperaturedependent creep behavior. Experimental creep curves first need to be obtained at a series of different temperatures over a specific time period, followed by the values of compliance plotted on a logarithmic timescale. After one creep curve at a chosen temperature is defined as reference, creep curves at other temperatures are then shifted one by one along the log timescale until they superimpose to a single curve in the ideal case. Curves above the reference temperature are shifted to the right, and those below are shifted to the left. This procedure can be applied to predict long-term creep compliance on the basis of short-term tests at different temperature levels in the range of linear viscoelasticity. In practice, polymers and adhesive joints are often being used under service conditions exceeding the limitations of linear viscoelasticity. Indications for nonlinear viscoelastic properties include the observation that creep compliance changes for different stress levels and that the mechanical behavior not only depends on the summarized history of previous mechanical incidents but also from the sequence in which they have been applied. In this case, the simplified constitutive equations of linear viscoelasticity will fail in describing time-dependent mechanical properties. Further reference to nonlinear viscoelasticity is given in Haddad (1995) and Brinson and Brinson (2008).

34.4

Experimental Testing

34.4.1 Accessing Creep Properties of Adhesives by Rheometric Test Procedures The term “rheology” was first introduced by Bingham and Crawford in 1948 (Scott Blair 1969). Rheology today provides a large range of experimental procedures to study the flow of fluids and the plastic deformation of soft solids. Most scientific rheometric instruments are designed to apply a constant or oscillating torque to a sample located in a precisely adjusted and thermally controlled gap between test plates which may carry a flat or conical surface. The parallel plate (or plate-plate) geometry has the advantage of being able to take preformed sample discs which can be especially useful when working with film or tape adhesives. The main disadvantage of parallel plates is due to the fact that the shear rate varies across the radius of the sample. In most cases, an average value for the shear rate is displayed in the test results. The cone to plate geometry is designed to maintain a uniform shear rate by narrowing the shear gap toward the center of the sample. Rheological tests are especially useful to experimentally determine the limits of linear viscoelasticity of an adhesive bulk sample, to investigate its zero-shear viscosity as a measure for creep under long-term loading, and to obtain a data basis for the calculation of relaxation time spectra.

990

P. L. Geiß and M. Schumann

Limits of Linear Viscoelasticity (LVE) In a strain sweep under dynamic oscillation, increasing levels of strain are applied at a certain frequency. According to ASTM D 7175, the departure of the linear viscoelastic region is defined as the point at which the elastic modulus G0 deviates by more than 10% from a constant (plateau) value. It should be considered that the choice of test frequency has an impact on the limit of LVE with higher frequencies causing a shift of the LVE limit to lower strain values. Figure 7 illustrates the oscillatory strain-sweep results for a hot melt adhesive in plate-plate geometry with a diameter of 25 mm at 70  C and a test frequency of 0.1 Hz obtained with a Bohlin Gemini 150 rheometer. Departure from linear behavior occurs in this case at about 30% of deformation. Reference on creep in relation to the rheology of pressure-sensitive adhesives is given in ▶ Chap. 15, “Pressure-Sensitive Adhesives (PSAs)” by C.W. Paul. Zero-Shear Viscosity At sufficiently low shear rates, creep-sensitive adhesives exhibit a so-called zeroshear viscosity (ZSV). Creep-sensitive adhesives include non-crosslinked hot melt adhesives or non-crosslinked pressure-sensitive adhesives. This parameter is indicative for the bulk adhesive’s resistance to permanent deformation under long-term loading. Zero-shear viscosity (ZSV) may be derived for such materials from rheological creep tests under constant shear stress. The zero-shear viscosity is determined as the “steady-state viscosity” at sufficiently low levels of shear stress in the linear 30'000 Δ 10 %

Elastic modulus [Pa]

20'000

Viscous modulus [Pa] 15'000

10'000 End fof LVE

Elastic / viscous modulus [Pa]

25'000

5'000

0 −1.0

−0.5

0.0

0.5

Log deformation [ - ] Fig. 7 Identification of the limit of linear viscoelasticity from an oscillatory strain-sweep test

1.0

34

Creep Load Conditions

991

1 Instantaneous elastic strain

Log deformation [-]

Secondary creep

Delayed elastic strain Primary creep 0.1

Creep curve Unrecoverble strain

Recovery curve Instantaneous elastic strain 0.01

0

400

200

600

Time [S ] Fig. 8 Identification of secondary creep in a rheometric creep test

viscoelastic region, where strain curves should be independent of the level of stress. The steady-state viscosity criterion relates to the change in strain as a function of time and may be included in the experimental parameters of most rheometric instruments to facilitate the selection of a sufficient timescale. As the adhesive acquires secondary creep, the slope of the strain vs time diagram represents the zero-shear viscosity, and the differential of the slope approaches 1.00. Most rheometer software gives readout of the slope as the test progresses and can be set to automatically accept the data when this steady-state criterion is attained. Using state-of-the-art software tools for nonlinear regression analysis, the viscoelastic parameters and relaxation time corresponding to a preselected model according to Table 1 may be fitted to the experimental data of rheometric creep/creep recovery tests. Figure 8 illustrates the time-dependent strain of a hot melt adhesive in a rotation-mode experiment under a constant stress of 1000 Pa, in plate-plate geometry with a diameter of 25 mm and a temperature of 70  C. In this specific case, the zero-shear viscosity in the region of secondary creep converges to a level of approximately 1 MPas.

34.4.2 Retardation Test Methods with Adhesively Bonded Joints Due to the polymer nature of most structural engineering adhesives, their viscoelastic deformation behavior is strongly time and temperature dependent and associated to creep (retardation) under static or so-called “dead-load” condition.

992

P. L. Geiß and M. Schumann

There are several standards dealing with creep and long-term mechanical properties. ISO 899 Part 1 “Plastics – Determination of creep behaviour” specifies a method for determining the tensile creep of plastics in the form of standard bulk test specimens like specified in ISO 527. The method is suitable for use with rigid and semirigid and filled plastic materials and can be applied to the testing of adhesive bulk specimen in the form of dumb-bell-shaped test specimens molded directly or machined from molded sheets. The method is intended to provide data for engineering design and research and development purposes. Acquiring data for engineeringdesign purposes requires the use of extensometers like video extensometers allowing a precise measuring of the gauge length of the specimen during the test. A multisample tensile creep tester with video extensometer is displayed in Fig. 9. Tensile creep properties strongly depend on the preparation procedure, the dimensions of the specimen, and the test environment. The thermal history of the test specimen has also shown to have a significant effect on creep behavior. Consequently, when exact comparative results are required, these factors must be carefully controlled.

Fig. 9 Multi-sample tensile creep tester with video extensometer

Creep Load Conditions

993

Creep diagram P2'

t1

s1

P2

Strain e(t)

Isochronous stress–strain diagram

s2

Stress s

34

P1' P1

P2'

P2

s2 s1

t2

P1 P1'

t1 Log time t2

Strain e(t)

Fig. 10 Conversion of creep data into isochronous stress-strain diagrams Creep diagram

Iso-strain diagram s2

P2'

e2

s1

P2 P1'

e1

s1

P1 t1

Log time

t2

P2'

P2

s2 Stress s

Strain e (t)

e3

e3 P1' e2

P1

t1

e1 Log time

t2

Fig. 11 Conversion of creep data into iso-strain diagrams

In determining creep properties, a series of specimens are subjected to static loads at different levels, and their increase in strain over time is measured. Data may be either presented directly as creep strain on a logarithmic timescale, as creep modulus EC(t) = σ/e(t), or as isochronous stress-strain diagrams where the relation between stress on the axis of ordinates over strain on the axis of abscissa is presented for different time levels. The schematic approach to converting creep data into an isochronous stress-strain diagram is illustrated in Fig. 10. Other ways to present creep data include compilation to stress-time diagrams with continuous lines indicating time-dependent strain limits. Therefore, this way of plotting creep data is frequently referred to as an iso-strain diagram. The schematic approach for converting creep data into an iso-strain diagram is illustrated in Fig. 11. A third alternative way of presenting the time-dependent stress-strain relationship due to creep under static load conditions is to plot the creep modulus EC (t) on a logarithmic timescale. Provided that the materials creep is linear viscoelastic, data at different stress levels should cause similar time-dependent values of creep modulus. The schematic approach for converting creep data into a creep modulus versus time diagram is illustrated in Fig. 12. Since the long-term behavior of bonded joints is strongly influenced by the adhesive-surface interaction at the interface, a prediction of creep and relaxation

994

P. L. Geiß and M. Schumann

Creep diagram

Creep modulus diagram s2 s1

P1, P2

P1'

s / e (t)

Strain e (t )

P2' P2

P1', P2'

P1 t1

Log time

t2

t1

Log time

t2

Fig. 12 Conversion of creep data into creep modulus versus time diagrams

for adhesives cannot only be founded on data obtained from adhesive bulk specimen in tensile mode but should consider shear load conditions and geometric bond line size effects close to practice adhesive joint assemblies. The standard ISO 15109 “Adhesives – Determination of the time to rupture of bonded joints under static load” is focused on time to rupture of bonded lap shear specimens under a specific static load. No time-dependent shear strain is recorded according to this standard. Initial stress levels are chosen as a percentage of the ultimate tensile lap shear strength in quasi-static experiments in accordance with ISO 4587. The results are expressed as a graph of stress versus time to failure including remarks on the type of failure pattern for each specimen. A more sophisticated procedure is described in ASTM D 1780 “Standard Practice for Conducting Creep Tests of Metal-to-Metal Adhesives.” This standard covers the determination of the amount of creep of metal-to-metal adhesive bonds due to the combined effects of temperature, tensile shear stress, and time. The primary intention of this test procedure using single lap shear specimens is to provide basic data for the choice of safe working stresses for applications in which the allowable deformation within the service life of the structure is the criterion of failure. The standard is intended to be used with standard testing machines in combination with extensometers allowing to measure shear strain close to the bond line of the employed lap shear specimen. For the purpose of conducting creep experiments with adhesively bonded joints under natural or accelerated aging conditions, several spring loaded creep test instruments have been suggested. Since constant-stress creep tests are desirable, the usual common practice would be using constant dead weight load tests with gravity causing the predefined level of load. Creep tests made by means of spring loading or fixtures, which involve deflection or strain measurements for the application of load, are generally considered to be unsatisfactory because the creep strain of the specimen will cause a reduction of stress due to, e.g., contraction of a tensile spring. However, if the total deformation or extension in the adhesive is large, corrections can be used to compensate for the decrease in stress because of the extension in the adhesive. Spring-loaded units have the advantage of reduced weight

34

Creep Load Conditions

995

compared to creep testers applying the required creep load directly by means of masses. Commercial creep test instruments overcome the problem of bulky load weights by using cantilevers to amplify the load to the desired stress level during the creep experiment. Special care needs to be taken to eliminate friction in the cantilever movement since this is likely to tamper the test results. Spring-loaded instruments for conducting creep tests on single lap shear specimens in tensile or compressive mode are described, e.g., in ASTM D 2294 “Standard Test Method for Creep Properties of Adhesives in Shear by Tension Loading (Metalto Metal)” or ASTM D 2293 “Standard Test Method for Creep Properties of Adhesives in Shear by Compression Loading (Metal-to-Metal).” In ASTM D 2294, the tension creep test apparatus consists of a hollow loading chamber, a solid extension rod with provisions for attachment of test specimens, and a spring. A testing machine is required to apply the static load, and a microscope is being suggested to measure the amount of creep shear strain. According to ASTM D 2293, the load conditions are adjusted in a less accurate manner by compressing a spring between two bushings to the desired load by tightening a nut. The applied load can be estimated by deflecting the spring of a given measured amount as predetermined from a calibration curve. The standard ASTM D 4680 “Standard Test Method for Creep and Time to Failure of Adhesives in Static Shear by Compression Loading (Wood-to-Wood)” also applies a compressive force to shear specimens to monitor the creep properties of wood adhesives in lap shear geometry. Since creep is considered as a key weakness of pressure-sensitive adhesives, various specific test methods and standards have been developed to evaluate the creep resistance of pressure-sensitive adhesives like in the European Standard EN 1943 (“Self-adhesive tapes. Measurement of static shear adhesion”), FINAT (Federation Internationale des Fabricants et Transformateurs d’Adhesives et Thermocollants sur Papier et autres Supports), test method FTM 8 (Resistance to shear from a standard surface), or Pressure-Sensitive Tape Councils test method PSTC 107 (Harmonized International Standard for Shear Adhesion of Pressure-Sensitive Tape) by either monitoring the time- and load-dependent displacement of an adhesive specimen under shear load or simply recording the time to failure. The result of the so-called SAFT-test (“Shear Adhesion Failure Temperature“) indicates the temperature at which a sample has failed that has been subjected to an environment with steadily rising temperature under static shear load. Figure 13 represents the typical course of a creep experiment using single lap shear specimen bonded with a viscoelastic acrylic adhesive under tensile load. The curve progression can be divided into three phases of creep. At the end of phase I, instantaneous elastic strain and delayed elastic strain are completed, and the creep progress in phase II is dominated by secondary creep. At the end of phase II, creep accelerates leading to failure of the specimen at the end of phase III. Failure can either occur due to cohesive fracture of the adhesive, due to interfacial sliding, or due to adhesive debonding at or close to the adhesive substrate interface. A set of creep experiments at different load levels can be used to build creepstrength-time curves indicating at which time- and load-related limits a failure under

996

P. L. Geiß and M. Schumann Phase I

Phase II

Phase III

10

Creep [mm]

8 6

Accelerated creep and failure

4 Elastic strain

2

Steady-state creep

0 0

24

48

72 Time [h]

96

120

144

Fig. 13 Typical course of a creep experiment for single lap shear specimen with a viscoelastic acrylic adhesive (Geiß 1998)

static load condition is to be expected for a certain adhesive-substrate combination. The construction of a so-called “creep rupture envelope” by connecting the points of the onset of the tertiary creep stage is shown in Fig. 14. The creep rupture envelope in Fig. 14 was obtained at a temperature of 60  C. Creep rupture envelopes for the same type of adhesive joint at different temperatures may show different characteristics if, e.g., the failure mode is changed by the influence of heat. The empirical prediction of long-term behavior from such technical test methods may often not lead to the desired level of certainty since the failure of specimens under static shear loads may involve random fracture mechanisms which cause a wide scattering of the results in terms of time to failure. In addition to this, the failure of viscoelastic adhesives in creep experiments often occurs when high strain values far beyond the limits of linear viscoelasticity have been reached which are not tolerable in most technical applications and therefore do often not contribute satisfactory to the application-related long-term predictability. Excessive shear strain should generally be avoided in adhesive joints because cohesive failure is likely to occur once shear deformation leads to a buildup of tensile stress at the end of the overlap bond area. In practice, it has been found that strain in assemblies with rubbery moisture curing one-component polyurethane adhesives should be limited to values of shear strain tan γ below 0.5. Specific viscoelastic pressure-sensitive tapes may even tolerate shear strains of tan γ up to 2. The maximum tolerable strain of structural adhesive joints using adhesives with a highly crosslinking density, however, is typically more than one order of magnitude lower than that. Due to the significance of limiting shear strain, creep data can be presented as iso-strain curves. This can be accomplished by taking stress-strain data from creep experiments at different stress levels and converting them to iso-strain lines (Fig. 15).

34

Creep Load Conditions

997

1.2 70 N 60 N

1

50 N 40 N

Creep [cm]

0.8

30 N

0.6

25 N 20 N

0.4

15 N 10 N

0.2

7.5 N 5N

0 0.1

1

10

100

1,000

Time [h]

Fig. 14 Creep-strength-time diagram based on a set of creep experiments at different levels of static load at a temperature of 60  C (Geiß 1998)

The presentation of creep data in the manner according to Fig. 15 allows to easily read stress limits if a certain allowable application-related strain level due to creep must not be exceeded within a given timeframe. Other ways to present creep data include isochronous stress-strain diagrams (Fig. 10) or creep modulus-time diagrams (Fig. 12). Attempts to project results from short-term creep tests to a longer timescale often fail because the increase of shear strain during a static load creep experiment will lead to a change of the state of stress in the adhesive joint and will eventually cause an excessive amount of superimposed tensile stress near the end of the overlap area. Therefore, the dotted lines in Fig. 15 merely represent a “guide to the eye” projection rather than a mathematically confirmed extrapolation. The problem of changing the state of stress in a static long-term experiment is less in relaxation tests compared to creep tests. Since the strain in relaxation tests is to be maintained at a constant level throughout the test time, a better accessibility to a mathematically based extrapolation may often be achieved. Section “4.3” will further address this matter. Beyond collecting empirical creep data for specific adhesives or adhesive joints, it is in some cases desirable to derive viscoelastic parameters in relation to the previously discussed mechanical substitute models of viscoelasticity from the test results. The challenge hereby is to distinctively separate instantaneous elastic and delayed elastic strain (primary creep) from non-recoverable strain (secondary creep) and to determine whether the principles of linear viscoelasticity may or may not be applied.

998

P. L. Geiß and M. Schumann 0.09

Shear stress [N/mm2 ]

tan g =1

0.07

tan g = 0.9 tan g = 0.8

0.05 tan g = 0.7

0.03

tan g = 0.6 tan g = 0.5

0.01 0.01

0.1

1 Time [h]

10

100

Fig. 15 Load-dependent time to reach different deformation limits indicated by iso-strain curves derived from static load creep experiments in lap shear specimens with a viscoelastic acrylic adhesive (Geiß and Vogt 2007)

A good method to reduce the amount of test specimens needed to collect creep data on different stress levels is to take advantage of the so-called steploading experiments. In an experiment with step loading, the stress is again and again increased after a certain period of time. The viscoelastic response in consequent load steps in Fig. 16 should be similar if the Boltzmann superposition principle can be applied and the material is behaving according to linear viscoelasticity. In the experiment illustrated in Fig. 16, the first step (caused by a load of 32 N) has led to an initial deformation of the sample of a bit less than 4 mm. During the following load steps with a repeated stress growth of 125% of the preceding load, the additional creep response decreases with higher total strain values, and also the slope at the end of each step decreases, although the total stress becomes larger. This indicates that the viscoelastic parameters of the specific material at these levels of creep change due to a strain hardening type of behavior, and the creep response does not represent linear viscoelastic properties. In some cases, it may be useful to dissent from the method of creep experiments under a constant static load and to subject the adhesive joint to a sequence of periodical static load events. The additional information obtained in this manner is the creep and the creep recovery behavior of an adhesive joint with viscoelastic properties (Fig. 17). The test in this case was carried out using single lap shear specimens and a time controlled mechanical lifting device to apply and remove a balanced weight causing a periodically repeating history of static load inputs.

34

Creep Load Conditions

999

Fig. 16 Creep response of an acrylic viscoelastic adhesive in a step-loading experiment (Geiß 1998)

12

Creep [mm]

10 8 6

62.5 N

50 N

4

40 N

2

32 N

0 0

48 Time [h]

72

96

10 Experiment

9 8

Burgers model

7 Creep [mm]

Fig. 17 Creep and creep recovery of a lap shear specimen bonded with an adhesive based on a SIS block copolymer formulation (Geiß 1998)

24

6 5 4 3 2 1 0 0

48

96

144 Time [h]

192

240

288

The line connected by data points in Fig. 17 represents the experimental data, while the continuous curve relates to the numerically simulated behavior based on the four parameter Burgers model for viscoelasticity. The load history consists of repeatedly loading for 24 h followed by 24 h of recovery. The steadily increasing residual strain after recovery gives a clear indication of the viscous dashpot properties η1 in the Maxwell section of the Burgers model. The E1 spring properties can easily be fitted according to the repeating elastic strain in the creep and creep recovery sequence. The parameters η2 and E2 in the Kelvin-Voigt part of the Burgers model could be adjusted to adequately fit the experimental data during the first two cycles. In this specific test, an early indication of damage accumulation could already be observed in the third cycle by a deviation of the experimental data from the predicted behavior based on the Burgers parameters. In the following cycles, this deviation increased steadily, leading to the failure of the specimen during the seventh load cycle.

1000

P. L. Geiß and M. Schumann

This example was intended to illustrate how a combination of experiment and simulation based on simple viscoelastic models can contribute to generate added value from creep experiments without expanding their complexity. A recurring challenge in designing creep experiments is the choice of the load levels and width of their distribution. Quasi-static lap shear tests at low strain rates indicate the short-term load capacity of the joint. As a rule of thumb, the higher the strain rate dependency of the modulus and fracture strength of an adhesive is, the lower the static load as a percentage of the ultimate fracture strength should be chosen if no other application-related considerations govern the level of static load to be used in a specific experiment. Structural adhesives may have a static load-bearing capacity in the range of 70% to even 90% of their short-term strength. Creep resistance of pressure-sensitive adhesives may be limited to 20% of their shortterm strength or even less. To overcome the risk of investing efforts in creep experiments resulting in strain levels beyond the area of relevancy for a certain application, relaxation test methods may provide an effective alternative test procedure.

34.4.3 Relaxation Test Methods with Adhesively Bonded Joints In the engineering process of bonded structures, the limits for maximum tolerable deformations in a structure to maintain its functionality and appearance are often important for the design of the adhesive joint. The shear strain of adhesive joints is generally expressed as tan γ = s/d, where s is the displacement of the adherends and d is the thickness of the adhesive layer. Once the admissible tan γ deflection is established, relaxation tests may provide an effective method to determine the related static load capacity of an adhesive or adhesive joint or the required size of the bond area to support a given load under such limitations. While creep experiments apply a constant load to the sample and follow the increase of strain over time, the relaxation method applies a fixed strain to the specimen at the beginning of the test and records the decrease in stress over time caused by relaxation of the adhesive polymer. The basic difference in creep (retardation) testing versus relaxation testing is illustrated in Fig. 18. To obtain accurate relaxation data, the elastic behavior of the test instrument needs to be taken into account. The elastic compliance of a fixed test frame would generally influence the designated level of constant strain if the stress and the associated load of the test frame decrease during the course of the experiment. The following experimental relaxation data in Fig. 19 were therefore obtained in single overlap shear relaxation tests using a testing machine with noncontact video extensometry and closed-loop control to apply the initial strain at a predefined constant strain rate and then keep the true shear strain in the bond area on a precisely constant level throughout the test. On the left side of Fig. 19, data have been plotted on a logarithmic timescale with linear ordinate units. Conversion to a double logarithmic scale on the right side of Fig. 19 leads to a linearization of the relaxation characteristic, which allows a better

Relaxation:

Strain 0

Time

0

Time

0

Time

0

Time

Stress

Creep: (retardation)

1001

Stress

Creep Load Conditions

Strain

34

Fig. 18 Comparison of input and response in creep (retardation) and relaxation tests

1.0 Relaxation (tan γ = 1)

Shear stress [N/mm2 ]

Shear stress [N/mm2]

0.3

0.2

0.1

0 0.001

0.01

0.1 1 Time [h]

10

100

Relaxation (tan γ = 1)

0.1

0.01 0.001

0.01

0.1 1 Time [h]

10

100

Fig. 19 Relaxation of shear-loaded adhesive joints with a viscoelastic acrylic adhesive, left: linear ordinate, right: logarithmic ordinate

projection and extrapolation of relaxation on a prolonged timescale. The relaxation data in Fig. 20 have been acquired in a similar manner starting from different shear strain levels tan γ. In the range of the applied strain values between tan γ = 0.5 and tan γ = 1, a good linear conformity at times of relaxation beyond 1 h can be observed. Linearity in this case means that the decrease in shear stress as a function of time can be approximated by the following equation: yðtÞ ¼ b  tð1=aÞ

(29)

On a double logarithmic scale, the equation transforms to yðtÞ ¼ ð1=aÞt þ b

(30)

1002

P. L. Geiß and M. Schumann 1

Shear stress [N/mm 2]

tan γ = 2 tan γ = 1.5

0.1 tan γ = 1 tan γ = 0.5

0.01 0.0001

0.04 N/mm2

0.01

100

1 Time [h]

Fig. 20 Relaxation of shear-loaded adhesive joints from different strain levels with a viscoelastic acrylic adhesive

1.6

0.0

Tensile stress [N/mm2]

Log tensile stress [N/mm2]

25% strain

1.4

10% strain

1.2

5% strain

1.0

2% strain

0.8 0.6 0.4 0.2

–0.2 –0.4 –0.6 –0.8 –1.0 –1.2 –1.4

0.0 0.0

0.5

1.0 Time [h]

1.5

0.0

0.5

1.0 Log time [h]

1.5

Fig. 21 Relaxation behavior of an elastoplastic adhesive on a double logarithmic scale (Geiß and Vogt 2007)

If the creep behavior of an adhesive or adhesive joint meets this criterion, the 1/a and b parameters can easily be retrieved by linear regression. The factor b is proportional to the initial load applied to the specimen (elastic properties of the adhesive), and the slope parameter a relates to the relaxation properties of the adhesive. The data in Fig. 21 have been obtained in relaxation experiments using bulk specimen of elastoplastic adhesives at initial strain levels ranging from 2% to 25%. Similar to the shear stress relaxation of the adhesive joints in Fig. 20, the relaxation behavior in the long-term timescale in Fig. 21 exhibits a sufficient degree

34

Creep Load Conditions

1003

Table 2 Parameters obtained by linear regression of the relaxation curves in Fig. 21 Initial strain [%] 2 5 10 25

Linear regression parameters 1/a 0.0736 0.0709 0.0684 0.0909

log b 0.9046 0.5653 0.3413 0.0482

of linearity in the double logarithmic representation of the experimental data. The parameters obtained by linear regression are listed in Table 2. It was found that in this case the parameter (1/a) relating to relaxation neither depended on the initial strain values up to 10% nor on the corresponding initial stresses from which the relaxation tests started. The initial stress levels calculated from the log b parameter were in this case in good agreement with the actual measured stress values at the beginning of the relaxation experiments.

34.4.4 Relation of Stress Relaxation and Static Load Creep Taking into account the transferability of creep and relaxation results, the following considerations may be useful in predicting static load limits of adhesives or adhesive joints. The experimental data of time-dependent stress from relaxation tests can be converted and applied to estimate maximum load values under creep conditions. This can be done considering the assumption that the declined value of shear stress after a specific time in the relaxation experiment will cause less deformation when this residual level of stress is constantly applied in a corresponding creep experiment. To approve this concept experimentally, single overlap shear specimens were subjected to stress relaxation at room temperature after imposing an initial strain of tan γ = 1. During relaxation, time-dependent stress values were determined for 0.5, 2, 8, and 16 h (Fig. 22). The results of four corresponding creep experiments applying the time-dependent load values from Fig. 22 indicate that the predicted strain limit of tan γ = 1 was not exceeded in either one of the creep tests and the congruence between relaxation and creep becomes better in the long-term scale (Fig. 23). All of the above conclusions have been drawn for isothermal conditions without considering a thermal fluctuation under creep conditions but which in practice represent the usual case during the service life of structural adhesive joints. Considering temperature effects in creep-dependent lifetime prediction can follow a worstcase scenario in which either the definition of the load-dependent fracture envelope or the test for compliance with predefined strain limits is carried out at the maximum

1004

P. L. Geiß and M. Schumann

0.1

Shear stress [N/mm2]

Relaxation (tan γ = 1)

s = 0.051 MPa s = 0.0398 MPa s = 0.0317 MPa

0.01 0.1

1

s = 0.0287 MPa

10

100

Time [h]

Fig. 22 Relaxation data for aluminum lap shear specimens bonded with a viscoelastic acrylic adhesive (Geiß and Vogt 2005) 10

Strain tan [ γ ]

Creep

tan γ0.5h = 0.68 s = 0.051 MPa

tan γ2h = 0.68 s = 0.0398 MPa

tan γ8h = 0.81 tan γ16h = 0.84 s = 0.0317 MPa s = 0.0287 MPa

1

0.1 0.1

1

10

100

Time [h]

Fig. 23 Results of creep experiments under static stress at ambient temperature corresponding to load values extracted from Fig. 22 (Geiß and Vogt 2005)

temperature to be expected during service life. This conservative approach is likely to lead to excessive contingency reserves. Alternatively, extended effort must be taken to collect creep data in the above described manner at different temperatures relevant for a specific application and to then estimate total levels of creep under in service conditions based on the probability of different temperature exposures during service life using predictive numerical methods.

34

Creep Load Conditions

34.5

1005

Predictive Methods

Long-term life prediction for adhesives and adhesive joints based on short-term laboratory tests remains a demanding challenge. Testing is generally carried out with the intention of simulating and often accelerating time-dependent effects that may happen to a joint during its lifetime under less severe conditions. Extrapolation methods generally intend to establish correlations involving temperature, time to rupture, and stress so that the least possible number of accelerated tests may be used to derive a sufficient lifetime prediction. The basic starting point for building a descriptive creep function in correlation to experimental data for the purpose of extrapolation is the following equation: eðtÞ ¼ Fðσ, t, T Þ ¼ f ðσ Þ  gðtÞ  hðT Þ

(31)

Besides the level of stress, both time and temperature need to be explicitly included in the representation since both parameters significantly affect the creep response. The functions f(σ), g(t), and h(T) are introduced on the basis of the assumption of superposition of the effects of stress, temperature, and time. While the viscoelastic models discussed in Sect. 2 may provide a starting point for establishing the function to describe the time- and stress-dependent part of the creep function, the time-temperature superposition equation of Williams, Landel, and Ferry (Sect. 2) should be considered for building time- and temperature-related descriptive empirical creep functions for the purpose of extrapolating and predicting creep behavior of a certain adhesive material or an adhesive joint. Due to the large bandwidth of adhesive mechanical polymer properties, no further generalized recommendation for a valid procedure can be given at this point. It must further be pointed out that such predictive methods based on numerical evaluation can only be applied under the premise that no mechanisms of aging, being either independent or related to the presence of mechanical load, are superimposing the long-term behavior and thus the durability. Common detrimental environmental influences on the static load-bearing capacity of adhesive joints include an increase of compliance and reduction of the glass transition temperature due to absorption of humidity or water or swelling under the influence of other media. Other aging effects may include corrosive attack starting at the bondline, debonding at or close to the interface between adhesive and substrate, or chemical aging of the adhesive polymer under the influence of temperature, oxygen, or ultraviolet or other ionizing radiation. Further reference on aging of adhesively bonded joints is given in ▶ Chaps. 31, “Effect of Water and Mechanical Stress on Durability,” and ▶ 32, “Adhesives in Space Environment”.

34.6

Conclusions

The enhanced plasticity and toughness of modern adhesives has become an important property in applications where dissipation of energy in the case of impact, the ability to compensate thermal movement, or the reduction of vibration leads to added

1006

P. L. Geiß and M. Schumann

value and improved service performance. Viscoelastic behavior can theoretically be predicted and analyzed by means of mechanical models including combinations of elastic and viscous elements to simulate the time-dependent viscoelastic stress-strain response to mechanical loads. Due to the increased use of elastomer adhesives, there has been a strong need to develop experimental techniques leading to reliable engineering data in less time which generate materials data that can be implemented into modern design tools such as finite element simulation and used to predict the durability of adhesive joints under specific service conditions. Creep experiments under constant load conditions are quite familiar and have been described in various international standards because they require rather little experimental equipment. The amount of material-specific data, which can be retrieved from such tests, is, however, limited, and the validity of long-term extrapolation remains often questionable. Relaxation tests require a higher level of instrumental accuracy but may provide a better pathway in acquiring material-specific data from smaller number of individual tests and specimens. Since adhesive technical data sheets generally lack information on creep and relaxation properties, the design of structural adhesive joints to be used under static creep load conditions will require a significant effort and time for the acquisition and evaluation of experimental creep or relaxation data. Predictive numerical methods to extrapolate experimental creep data or to provide a lifetime prediction in terms of creep rupture probability need to consider time- and stressrelated factors as well as the strong temperature dependency of creep-related phenomena. On the basis of the creep and relaxation functions derived from viscoelastic models in combination with shift factors based on the principle of time-temperature superposition, an adequate modeling of time- and temperature-dependent long-term stress-strain behavior becomes feasible. Extrapolation of short-term data can only be made if it is known that the material is not subject to superimposed time-dependent changes in properties which may lead to a reduced load-bearing capacity and damage after a stable period. Extrapolation generally should not exceed more than 1 logarithmic unit of time beyond experimental data without adequate justification and in no case beyond 1.5 units as, e.g., recommended in the EOTA (European Organization for Technical Approvals) Guidance Document 003 “Assessment of working life of products” which is relevant for products of the building and construction industry.

References Betten J (2008) Creep mechanics, 3rd edn. Springer, Heidelberg, pp 237–248 Brinson HF, Brinson LC (2008) Polymer engineering science and viscoelasticity, an introduction. Springer, New York Burgers JM (1935) First report on viscosity and plasticity. Committee for the study of viscosity. Academy of Sciences, Amsterdam de With G (2006) Structure, deformation, and integrity of materials, volume II: plasticity, viscoelasticity, and fracture. Wiley-VCH, Weinheim, pp 569–645

34

Creep Load Conditions

1007

Ferry JD (1970) Viscoelastic properties of polymers, 2nd edn. Wiley-Interscience, New York Geiß PL (1998) Verarbeitungskonzepte und Belastungskriterien für Haftklebstoffe. Dissertation, University of Kaiserslautern. Hinterwaldner Verlag, Munich Geiß PL, Vogt D (2005) Assessment and prediction of mechanical long-term properties of adhesives with high plasticity. J Adhes Sci Technol 19:1291–1303 Geiß PL, Vogt D (2007) Durability of pressure sensitive adhesive joints. In: Proceedings of the 30rd pressure sensitive tape council tech conference (PSTC) Haddad YM (1995) Viscoelasticity of engineering materials. Chapman & Hall, London Niven WD (ed) (2003) The scientific papers of James Clerk Maxwell, vol 2. Dover Publications, New York Priesner C (ed) (1983) H. Staudinger, H. Mark und K. H. Meyer: Thesen zur Größe und Struktur der Makromoleküle. Wiley-VCH, New York Purrington RD (2009) The first professional scientist: Robert Hooke and the Royal Society of London. Birkhäuser, Basel Scott Blair GW (1969) Rheology – a brief historical survey. J Texture Stud 1:14–18 Thomson W Baron 1st Kelvin (1875) Elasticity, 9th edn. Encyclopaedia Britannica, London

Durability of Nonstructural Adhesives

35

James D. Palmer

Contents 35.1 35.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Durability Considerations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.1 Joint Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.2 Choice of Substrates and Adhesive Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.3 Plasticizer Content of Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.4 Surface Preparation and Primer Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.2.5 Methods of Application and Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.3 Methods for Estimating and Evaluating Durability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.3.1 Elevated Temperature and UV Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.3.2 Antioxidants and Stabilizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.3.3 Water, Humidity, and Saline Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.3.4 Stress and Fatigue . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.3.5 Assessment by Combinations of Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.4 Durability and the Construction Products Regulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.4.1 Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1010 1011 1011 1012 1012 1013 1014 1015 1016 1018 1021 1022 1023 1025 1026 1027 1027

Abstract

In all bonded assemblies, the durability of the adhesive or sealant joint is important to the integrity of the article concerned. The considerations of joint design, adhesive, sealant and substrate selection, surface preparation, and primer selection along with methods of application including mixing procedures, application techniques, and joint assembly are all reviewed. Methods of evaluation of durability and test regimes including exposure to elevated and reduced temperatures, UV radiation, moisture, saline solutions, stress, and fatigue, both J. D. Palmer (*) Staffordshire, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_35

1009

1010

J. D. Palmer

individually and combined in cycles, are considered. The relevance and application of these test condition to specific situations are discussed, including the role of antioxidants and cross-linking additives to improve durability. Reference is made to the corresponding standards according to the major standard organizations, such as the International Standards Organization (ISO), the European Committee for Standardization (CEN), and the British Standards Institute (BSI), and also to European directives and regulations and national legislation.

35.1

Introduction

A review of the literature shows that the majority of the published research relates to the study of the durability of structural adhesives and that there is little published specifically on the durability of nonstructural adhesives; however, the mechanisms and factors that lead to the failure of nonstructural adhesive bonds are the same as those for structural adhesives and are presented in detail elsewhere in this book. This chapter, therefore, illustrates the principle factors affecting durability, the available methods for estimating and evaluating durability, and the options for improving the performance of bonded materials utilizing examples from the world of adhesive and sealant applications. In conclusion, the importance of durability in the realm of building and construction is discussed in relation to the forthcoming EC Construction Products Regulation. Structural adhesives are defined as adhesives that can resist substantial loads and that are responsible for the strength and stiffness of the structure (Adams et al. 1997) and are generally rated by their shear strength. Nonstructural adhesives then are lower strength materials and tend to be rated by their peel strength (Petrie 2007). However, this is a generalization that can be misleading as in both cases the adhesive is responsible for the structural integrity of the bond, if not necessarily the entire structure. Durability is the ability to maintain the required performance over a given or long time, under the influence of the potential degradation factors, including temperature, humidity, water, UV radiation, abrasion, chemical attack, biological attack, corrosion, weathering, frost, freeze-thaw, and fatigue. Durability is, therefore, dependent on the intended use of the product and its service conditions, and the assessment of durability can relate to the bonded assembly as a whole or to its performance characteristics. In either case, the underlying assumption is that the performance will be maintained at an acceptable level, in relation to its initial performance, throughout its working life (EC Guidance Paper F 2004). A large number of standard test methods and specifications have therefore been developed to define adhesive and sealant applications and performance requirements including durability. Many of these are application or market specific and designed around the intended use and service conditions, ranging from specifications for wall covering adhesives (BS 3046: 1981) to test methods for determination of adhesion and cohesion properties of sealants after exposure to heat water and artificial light through glass (BS EN ISO 11431: 2002).

35

Durability of Nonstructural Adhesives

35.2

1011

Durability Considerations

Whenever an adhesive joint is made as part of a product or device, there is always concern about the joint durability whether this is in a box of breakfast cereal, the sole of a shoe, or an external panel of a motor car. The consequences of joint failure can, therefore, range from nuisance to endangering lives. As already stated, the durability or permanence of a bonded assembly is dependent on the intended use and service conditions to which the bond will be exposed. However, the joint design, choice of substrates, adhesive selection, substrate preparation, and primer selection, where appropriate, plus the method of application and assembly all have significant impact on the service life of adhesively bonded materials. Most or all of these considerations are interdependent, for example, the joint design and substrates chosen will limit the range of suitable adhesives that can be employed. In a similar way, the durability of a sealed joint is only as good as the adhesion of the sealant (and primer) to the surfaces forming the joint. Primers and/or sealants will adhere to surfaces only if those surfaces are properly prepared. A very large proportion of all sealant joint failures result from poor or inadequate surface preparation.

35.2.1 Joint Design Many factors need to be considered when designing a bonded assembly, which will utilize adhesive or sealant. They include the purpose of the joint and the stresses likely to be encountered, the substrates to be joined and their physical properties, such as coefficient(s) of linear expansion, the type of materials forming the joint, maximum and minimum temperatures predicted for the components and their heat capacity and conductivity, and any other external influences such as exposure to chemical attack (e.g., water, acids, solvents, etc.), physical movement from vibration, etc. Joint geometry is a prime consideration, and bonded joints need to be designed so that loading stresses will be directed along the lines of the greatest strength of the adhesive or sealant. In general, joints may be subject to tensile, compressive, shear, or peel stresses, often in combination, and adhesives and sealants are strongest in shear, compression, and tension and perform less effectively under peel or cleavage loadings. Sufficient adhesive surface or bonded area must also be provided between the adherends to enable the transfer of forces via the bonded joint. In general adhesive joints are relatively low in cross-sectional area compared to sealant joints, and an additional consideration in sealant joint design is the designed movement accommodation required of the sealant. During their service life, sealants are subjected to mechanical stresses due to movement, and for a given amount of movement, wide joints will place less induced stress on the sealant than narrow joints, with implications for their durability. For example, consider a 5 mm movement in a joint: 10 mm wide bead – the movement is 5/10
100% = 50%. 20 mm wide bead – the movement is 5/20
100% = 25%.

1012

J. D. Palmer

35.2.2 Choice of Substrates and Adhesive Selection In contrast to other traditional mechanical joining methods such as riveting, welding, and screwing, adhesive bonding has little or no adverse effect on the material characteristics of the surfaces to be joined. However, in general, in the area of nonstructural adhesives, designers do not consider that materials have to be bonded, paying more attention to appearance and other aesthetic properties, and consequently the service life of bonded assemblies can be severely limited by the choice of substrate. Materials that contain plasticizers, or processing oils, for example, or have a mismatch in their coefficient of thermal expansion can result in bond strength deterioration with many adhesive types in relatively short timescales. Plasticizers are a common source of nonstructural bond failures and are discussed in more detail below. Formulators have developed many adhesive and sealants to suit most substrate types and properties, and adhesive selection is most commonly carried out by first considering the substrates to be bonded. Most manufacturers of adhesives and sealants publish guides of this type, and many can be found on the Internet to assist designers. The choice of an adhesive or sealant for a specific situation will be governed by consideration of several key parameters in relation to the manufacturing process to be employed: the setting time required, the initial performance requirements, the durability and ultimate performance requirements, the total product and process costs, and the associated health, safety, and environmental hazards associated with the product use. Additionally, sealants are most often selected on the basis of the intended joint movement accommodation requirement. The selection of adhesive is critical to fulfilling the design requirements in many applications, but in many industries, the adhesive component of the finished product represents a small percentage of the overall cost; however, its importance in terms of quality and durability of the finished products assumes much greater proportions. This is particularly the case in the manufacture of inflatable lifesaving equipment, such as life rafts and jackets. Solvent-borne polychloroprene adhesives are commonly used to bond rubberized fabrics in the manufacture of these items, and, for obvious reasons, all possible measures are utilized to ensure the durability of the bond. While the water resistance and general durability of correctly formulated polychloroprene adhesives are relatively good, the addition of isocyanate prepolymers, as part of the application process, to these systems is commonly employed to produce further cross-linking and enhanced longevity.

35.2.3 Plasticizer Content of Substrates Many plastic polymers contain plasticizers to make them softer and more flexible as determined by the finished product requirement for a specific application. Plasticizers are generally high boiling point organic liquids with molecular weights in the 400+ region, which have a degree of compatibility with the polymer concerned. In

35

Durability of Nonstructural Adhesives

1013

polyvinyl chloride, for example, the most commonly used plasticizers are the phthalate esters, such as dioctyl, dinonyl, and didecyl phthalate and their iso isomers. In the most flexible materials, commonly made from plastisols, where the polymer is dispersed in plasticizer and then heated to melt the polymer into the plasticizer, forming a homogeneous liquid, which forms a solid on cooling in a molding process, the plasticizer content can be in the order of 50% by weight. Since plasticizers are not bound chemically to the polymer and are held by relatively weak intermolecular forces, most are capable of migrating to the interface, where they can come into contact with the adhesive or sealant layers. Plasticizers also have an affinity for most adhesive polymers and can migrate across this interface resulting in softening of the adhesive or sealant and consequent reduction in physical properties and, in the worst case, cohesive adhesive failure. This migration of plasticizer into adhesive or sealant will also cause change in the physical properties of the plastic substrate itself such as dimensional changes and even surface stress cracking. One application area where this is of particular relevance is the adhesive bonding of floor and wall covering materials, and the relative resistance of adhesive systems to plasticizer migration can be determined utilizing the standard method described in EN 1903: (2008), “Adhesives – Test method for adhesives for plastic or rubber floor coverings or wall coverings – Determination of dimensional changes after accelerated aging.” This method gives a measure of the suitability of a plastic or rubber floor or wall covering/adhesive combination by monitoring dimensional changes during defined conditioning sequences when bonded to a specific substrate. The floor or wall covering test pieces, with dimensions of (250  5) mm
(250  5) mm or (300  5) mm
(300  5) mm, are first preconditioned for 6 h at 80  C  2  C before measuring their dimensions and are then bonded to uncoated fiber cement panels using the test adhesives and dimensions measured again. The test assemblies are then stored at 50  C for a 13 day period each followed by 24 h conditioning at room temperature prior to measuring the dimensions again. The interpretation being that the more cycles to which the test pieces are exposed, without significant dimensional change, the more durable the adhesive and floor or wall covering combination will be in service.

35.2.4 Surface Preparation and Primer Selection Correct surface preparation is critical to achieving optimum bond strength and joint reliability. Cleaners, such as surfactants and solvents, are used to ensure that surfaces to be bonded are clean and free from dust and other loose materials, grease, oils, and other impurities that may affect the adhesion and ultimate bond strength. Pretreatments are also used to enhance bonding performance including mechanical treatments such as grinding and blasting to produce a rough surface texture and effectively increase the bonded area; physical treatments such as corona discharge, flame, or plasma treatments to improve bonding performance by increasing surface energies; and primers or adhesion promoters.

1014

J. D. Palmer

Primers and adhesion promoters are normally low-viscosity solutions of either isocyanate prepolymers or organosilanes, which function by surface penetration into and chemical adhesion to the substrate(s) and chemical adhesion to the subsequently applied adhesive or sealant, increasing bonding performance and durability. Again, formulators have developed ranges of primers to suit most adhesive or sealant and substrate combinations, and selection guides are widely available.

35.2.5 Methods of Application and Assembly Automated methods for preparation, adhesive application, and bond assembly give the best reproducibility and best control of durability; however, such automation is frequently not possible. Manufacturers’ recommendations in data sheets and on containers should always be followed and particular attention paid to storage conditions, mixing procedures, adhesive or sealant application, and joint assembly.

Storage Conditions Storage temperature extremes can cause problems, which limit the effectiveness and shorten the service life of adhesives and sealants. Temperatures below zero can irreparably damage emulsion adhesives and sealants causing agglomeration or coagulation of the polymer emulsion. Many modern products are made freezethaw stable by the addition of chemicals such as glycols, but generally prolonged storage at low temperatures is inadvisable. Conversely, high storage temperatures can lead to premature polymerization of reactive adhesives severely impeding bonding performance. High humidity can also have adverse effects; for example, adhesives supplied in powder form require adequate protection from moisture vapor during storage to prevent premature curing or agglomeration making use and application difficult. Mixing Procedures Adequate mixing of multicomponent adhesives and sealants requires particular attention. Unless supplied in pre-weighed kit form, the correct proportioning of components is essential, and the duration and intensity of mixing with the correct type of mixer should be utilized to produce a homogeneous mix that is not overheated and with minimal unintentional air entrainment. However, excessive shear during mixing can damage some types of emulsion adhesives and sealants. Application In all adhesive and sealant applications, the aim is to apply a uniform layer of material of the appropriate thickness across the entire area to be joined to achieve the optimum performance. In many manual or semiautomatic applications, this is achieved by using spacers or jigs to control product thickness. Pressure-fed application equipment is also often utilized to ensure a consistent flow of product.

35

Durability of Nonstructural Adhesives

1015

Joint Assembly Once the adhesive or sealant has been applied to the substrate or substrates, the joint should be assembled within the allowable “open” or “assembly” time, that is, between the minimum and maximum recommended times. With solvent-borne contact adhesives, for example, sufficient time for solvent evaporation must be allowed to prevent solvent entrapment, which can lead to premature failure, especially when bonding impermeable substrates. The maximum open time with hot-melt adhesives, which set on cooling, can be very short and should be closely monitored. Ambient conditions can also affect the bond formation process. For example, high relative humidity can lead to condensation of moisture on the exposed drying surface of a solvent-based adhesive, as the heat of evaporation of the solvent causes the temperature of the adhesive film to drop below the dew point of the surrounding atmosphere. Such a film of moisture will impede the coalescence of the two films of a contact adhesive. Workmanship Quality of workmanship is obviously a key factor in the durability of bonded assemblies, and standard test methods are written in such a way to try to minimize any variability in laboratory evaluations. However, laboratory conditions cannot accurately reflect the actual application methods and conditions employed, and there have been some attempts to study the influence of workmanship on durability. One such study by Zhao and Zhang (1997) looked at the influence of workmanship on the bonding of ceramic tiles to external walls. In this study, they looked at three variables, tile setting pressure, adhesive inner cavity, and adhesive open time, and measured their effects on tile bonding or pull-off strength. Not surprisingly, their results indicated that the influences of these variables are significant and that when some or all of these factors appear simultaneously, the combined effect is cumulative in the reduction of pull-off strength.

35.3

Methods for Estimating and Evaluating Durability

Having taken into account the above considerations and completed the selection process, it is always desirable to know that the bonded assembly will meet the durability expectations of the product, that is, be “fit for purpose.” For this reason, adhesive selection is often made on the basis of past experience or “heritage data,” and a proven track record is recognized as a suitable method for demonstrating fitness for purpose. For example, the approved document in support of Regulation 7 of the UK Building Regulations – workmanship and materials – lists “past experience” as one of the ways of demonstrating fitness for the purpose of a material. However, such historical data is frequently not available, and other methods of assessment need to be employed. Durability tests that measure an adhesive or sealant’s performance retention during exposure to the main elements of its service

1016

J. D. Palmer

environment have, therefore, been developed, and this is usually measured in some accelerated manner because natural processes are too time-consuming to allow qualification in an appropriate timescale. At this point, it is worth noting that accelerated and natural aging tests do not always produce the same results. Ashcroft et al. (2001) undertook a comparison of aging joints in natural and laboratory conditions, with aluminum adherends, bonded with structural adhesives. A conclusion of their work was that there was no simple method of determining the long-term durability from accelerated tests and that excessive temperature and humidity will trigger degradation mechanisms, which are not representative. In general, accelerated aging tests tend to overestimate the reduction of joint strength. The published standard (ISO 9142 2003), Adhesives – Guide to the selection of standard laboratory aging conditions for testing bonded joints, which describes laboratory aging conditions under which adhesive joints may be exposed to various environmental influences, climatic or chemical, for the purpose of assessing the effects of such influences on certain properties, recognizes these limitations, warning in the introductory abstract that “The results obtained using the procedures described in this international standard are not necessarily applicable to the determination of the service life of a bonded assembly because there is no direct relation between the test results and the behavior of a bonded assembly over a period of time under service conditions. However, for certain specific applications, experience with the procedures may enable a correlation to be established.” Obviously, the purpose of an accelerated aging test is to cause a more rapid deterioration of the bonded assembly, and by paying due consideration to the actual service conditions, the accelerated aging conditions can be selected to recreate as closely as possible the failure mechanisms that would be seen in those actual service conditions. The accelerated aging conditions usually selected involve some environmental factors, which the bonded assembly might encounter, such as elevated temperature, UV radiation, high humidity, water immersion, exposure to aqueous salt solutions in spray form, and/or combinations of these, which may be in periodic cycles, as well as other predictable service factors such as stress and fatigue, where normal service loads can be applied in combination with the environmental accelerating factors above, or extreme loads, beyond normal service, to induce premature degradation.

35.3.1 Elevated Temperature and UV Radiation Perhaps the simplest durability evaluation is the exposure of hot-melt adhesives to elevated temperature for prolonged periods with their durability assessed by their color stability; the longer the period of exposure before significant discoloration occurs, the more durable the adhesive. While this test gives an indication of the ability of the adhesive to withstand long periods in the melt tank at or above the application temperature, it also gives an indication of the durability of the adhesive in

35

Durability of Nonstructural Adhesives

1017

service. Discoloration of the adhesive can be particularly significant with lightcolored substrates. All polymers containing hydrogen undergo oxidation by a free radical chain mechanism; however, the rate at which oxidation occurs varies with the chemical structure of the polymer. Polymers that are commonly used in adhesive formulations differ considerably in their sensitivity to oxidation. Acrylates, for example, are highly resistant to oxidation and do not generally require antioxidants at ambient temperatures, whereas unsaturated elastomers are highly susceptible to oxidative decomposition and require a relatively high concentration of antioxidants for protection. Both heat and light accelerate the oxidation of polymers, both increase the rate of reaction, but heat also increases the rate of inward diffusion of oxygen and the loss of antioxidants and stabilizers from the adhesive or sealant by volatilization. Mechanical deformation of the end product during service (“fatigue”) also accelerates the oxidation of adhesives and sealants. Oxidation of the polymeric components results in a deterioration of the physical properties, which can ultimately lead to bond failures, and the addition of antioxidants and stabilizers to improve durability is normally necessary. The deleterious effects of oxidation on polymer properties results from the breakdown of very unstable hydroperoxides in the polymer chain in the presence of light. This in turn leads to chain scission (molecular weight reduction) and the production of new free radicals which accelerate the oxidation process (Scott 1998). Elevated temperature-accelerated aging techniques rely on the dependence of the rate of reaction on temperature as expressed in the Arrhenius equation, where the rate constant k of a reaction and the absolute temperature T are related using the gas constant R: k ¼ AexpðEa=RT Þ

(1)

In its original form, the pre-exponential factor A and the activation energy Ea are considered to be temperature independent (IUPAC Compendium of Chemical Terminology 2nd Edition 1997). Elevated temperature is often used to estimate the in-can stability of adhesives and sealant formulations. This is usually applied by a simple “rule of thumb” that says the reaction rate of the degradation mechanisms involved doubles for every 10  C increase in temperature, and therefore 1 week storage at 50  C approximates to 8 weeks at room temperature (20  C). Application of this rule will likely be conservative in the prediction of shelf life and must be verified by real-time validation testing conducted at room temperature for the targeted shelf life. One example of in-can stability testing is given in ISO 15908 (2002), Adhesives for thermoplastic piping systems – Test method for the determination of thermal stability of adhesives, where samples of the adhesives in their relevant packaging, a minimum three test pieces for each packaging and adhesive variant, are stored at room temperature (23  C  2  C and 50% relative humidity) and at elevated temperature (for 1 month at 50  C  2  C followed by 24 h at 23  C  2  C and 50% relative humidity). After this period, the adhesive viscosities are measured and

1018

J. D. Palmer

their performances determined by measurement of the shear strength of bonded relevant plastic pipe systems. A stable viscosity and maintenance of bonding performance is assumed to indicate in-can stability; however, the standard method also recognizes that actual shelf life (in-can stability) can only be verified by tests conducted at actual storage conditions for the targeted shelf life period.

35.3.2 Antioxidants and Stabilizers Antioxidants and stabilizers are added to most polymer systems to improve durability. Antioxidants fall into two major categories, primary and secondary, (Petrie 2004) and within primary antioxidants, there are two main chemical types, hindered phenolics and aromatic amines. Primary antioxidants function by donating their reactive hydrogen to the peroxy free radicals so that the propagation of subsequent free radicals does not occur. The most widely used antioxidants for polymer protection are phenolics. These products generally resist staining or discoloration. However, they may form quinoid structures upon oxidation, which lead to yellowing. Phenolic antioxidants include simple phenolics, bisphenolics, polyphenolics, and thiobisphenolics. Hindered phenolics, such as butylated hydroxytoluene (BHT), high molecular weight phenolics, and thiobisphenolics, are the most popular of the primary antioxidants. They are suitable as long-term stabilizers in almost all cases. BHT has long been the workhorse of the antioxidant industry, and it possesses US Food and Drug Administration (FDA) approval for food contact. However, BHT is a relatively volatile material that is gradually being replaced by higher molecular weight antioxidants that resist migration. Thiobisphenolics can also act as peroxide decomposers (secondary antioxidants) at temperatures above 100  C. Therefore, they are often found in higher-temperature applications. Amines, normally arylamines, may be more effective than phenolics, but most are staining and discoloring and lack FDA approval for use in contact with food. Amines are commonly used in the rubber industry, but they also find uses in applications such as wire and cable formulations and in polyurethane polyols. Secondary antioxidants are used in conjunction with primary antioxidants to provide added stability to the polymer. They function by decomposing any hydroperoxides to form nonreactive products. Typical secondary antioxidant compounds contain sulfur or phosphorous. The more popular secondary antioxidants are thioesters (thiodipropionic acid derivatives and polythiodipropionates) and organophosphites. Organophosphites provide color stability but have a hygroscopic tendency. Hydrolysis of phosphites can ultimately lead to the formation of phosphoric acid, which can corrode processing equipment. Tris-nonylphenyl phosphite (TNPP) is the most commonly used organophosphite followed by tris(2,4,-di-tert-butylphenyl) phosphite. Thioesters have high heat stability but have an inherent odor, which can be transferred to the host polymer.

35

Durability of Nonstructural Adhesives

1019

The by-products of polymer degradation can also have a damaging effect on the durability of adhesive and sealant bonds necessitating the addition of specific stabilizers. For example, polychloroprene polymers can decompose by dehydrochlorination, liberating hydrochloric acid, which can attack the adhesive or sealant and the substrates. The addition of metal oxides, such as zinc oxide or magnesium oxide, which act as acid acceptors helps to overcome this problem.

Pressure-Sensitive Adhesives The use of pressure-sensitive adhesives (PSAs) in modern manufacturing has developed significantly in recent years. No longer are PSAs limited to temporary attachment applications, such as packaging tapes, but are increasingly being used in more sophisticated applications such as mobile telephones for the bonding of keypad switch assemblies. Pressure-sensitive applications have also been expanded to include a wide range of temperatures, ranging now from 40  C to +120  C. For example, airline baggage tag labels need to perform in low and high temperatures, often over a time period of a few hours. Some tapes intended for the construction or automotive markets need to perform at even higher-temperature extremes. So prediction of durability and service life has assumed greater significance to formulators of this type of adhesive. PSAs are usually characterized by their three common properties of tack, peel resistance, and shear resistance, which can be determined by standard laboratory test procedures. Each of these properties is affected by temperature change. Indeed, as already noted, most adhesives exhibit different properties when subjected to temperature, and in the case of PSAs, the bond is almost always thermoplastic, softening at elevated temperature and becoming more brittle at lower temperatures. As the temperature increases, the adhesive mass softens and “wets” the substrate more rapidly and completely, and the tack and peel properties increase. At some point, however, the adhesive has softened to the point that its cohesive strength is less than the substrate adhesion, and the adhesive mass splits, failing cohesively, that is, the locus of failure has changed from adhesion to cohesive within the adhesive. As the temperature continues to increase, the adhesive mass continues to soften, and the tack and peel values continue to go down. In testing shear resistance, one always looks for cohesive failure. As the temperature increases, shear resistance goes down as the adhesive mass softens. This reduction in shear resistance will be fairly slow and steady, to the point where the mass is so soft that it will not support a load for any appreciable time. It is important, therefore, that adhesives be evaluated for performance over a wide temperature range and bearing in mind the difference between application temperature and end-use temperature. The application temperature, at which the adhesive is applied to a surface and is expected to achieve a bond, is, generally, limited to the lower end and a narrower band of the temperature range, whereas the end-use temperature, at which the adhesive must perform adequately once the bond is made, is usually at the extreme ends, both high and low, of the temperature scale. In standard laboratory tests, the limits of application temperature are determined by placing the test samples and substrates at the selected temperature and allowing

1020

J. D. Palmer

them to equilibrate. Sample bonds are made at this temperature, and the laminated samples are allowed to reside for a specific period, after which performance properties are measured at that temperature. The limits of end-use temperature can be determined by making the sample bonds at a temperature within the application temperature range and allowing the bond to build for a specified time at this temperature. After this residence period, the samples are placed at the test temperature for a period of time to achieve equilibrium and then the performance properties determined at that temperature. These test conditions are achieved by the use of environmental chambers, freezers, or ovens. Alternative indications for durability of PSAs in specific applications can be obtained from modifications to these methods for determining performance properties, and the two most commonly used tests are the peel adhesion failure time (PAFT) and shear adhesion failure temperature (SAFT). The PAFT is determined using 25 mm strips of adhesive coated onto standard polyester film and bonded to standard stainless steel panels. The bonded assemblies are placed in a suitable enclosure to maintain the required test temperature and a weight (usually either 500 or 1,000 g) attached to the polyester film such that the assembly will peel in a 180 configuration. The time for the bond to fail at the test temperature is recorded, often using a suitable automatic recording device. Similarly, the SAFT is determined using a square adhesive contact area of 25 mm2 applied on to a standard polyester film and bonded to standard stainless steel panels and then placed in an oven starting at 25  C. A 500 g weight is applied, and the temperature is raised by 5  C every 10 min and the point at which the sample fails is recorded. Alternatively, if appropriate, the actual substrates intended for the application concerned can be utilized to make the tests more meaningful to the end use. Accelerated aging tests are also employed to estimate the durability of PSAs. While there is no standard set of aging conditions, 70  C, the standard aging condition for rubbers, on which most original PSAs were based, is commonly used; however, other test temperatures may be used depending on the temperature sensitivity of the components to be bonded. Unapplied samples, meaning that the samples are aged with the release liner in place, after which the liner is removed and the samples are applied to the test surface to determine performance properties, will give an indication of the “shelf life” to end use, whereas applied samples, meaning that the samples are applied to the test surface for the aging period before the determination of performance properties, will give an indication of the service life of the product after it is bonded to a substrate. While pressure-sensitive adhesive performance at various temperatures can be determined by practical tests, as described above, it is also possible to predict the temperature performance of pressure-sensitive adhesives. The polymers used in pressure-sensitive adhesives will have a variety of glass transition temperatures (Tg), and for a pressure-sensitive adhesive to be effective, the formulated mass must have a Tg of between 50 and 70  C below the temperature at which the adhesive is to be used. At the same time, however, the Tg cannot be so low that the adhesive mass is so soft that it fails cohesively even at room temperature. A variety of analytical techniques used to determine the Tg of the adhesive mass

35

Durability of Nonstructural Adhesives

1021

include differential scanning calorimetry (DSC), differential thermal analysis (DTA), and dynamic mechanical analysis (DMA). Of these, the most commonly used method is DMA. DMA will measure the bulk properties of the adhesive mass, which will include the base polymer and any additives in it. It does not take into consideration any interfacial relationships between the adhesive and the surfaces to which it is bonded. DMA can determine the storage modulus (G0 ) and loss modulus (G00 ) of the adhesive mass by sweeping through a range of frequencies and temperatures. Dividing G00 by G0 yields the property called tan delta, and plotting tan delta against temperature will give peaks that relate to the Tg of the adhesive mass. The type of information developed by DMA and other thermal analytical techniques can, therefore, be used to formulate adhesives that will perform well at ambient conditions and at temperature extremes. However, final practical evaluations of the adhesive formulation, the substrates to be used in the application and the temperature ranges to be seen in the application, must be undertaken.

35.3.3 Water, Humidity, and Saline Solutions Moisture ingress during service is thought to be responsible for many examples of premature joint failure. Accelerated aging tests are, therefore, commonly carried out by exposing joints or structures to either water immersion or water vapor. Moisture ingress in joints can occur either by bulk diffusion through permeable substrates or the adhesive or by wicking across an interface or through cracks or voids in the adhesive. Once present, water molecules can cause a modification in the adhesive mechanical properties, for example, by plasticization or by a reduction in the glass transition temperature. Alternatively, it is proposed that interfacial bonds can be disrupted or modification of the adherend surface can result, for example, by hydration of a metal. The rate of water ingress by diffusion through the adhesive is principally determined by its diffusion coefficient. An increase in temperature will rapidly increase the rate of water uptake in line with Fickian diffusion models as the diffusion coefficient will increase significantly with temperature. Such a hot-wet environment is generally regarded as a very aggressive form of accelerated aging (Critchlow 2005). The installation of exterior artificial sports surfaces is one application where water resistance is required. Adhesives are used in this process to bond seaming tape to artificial grass or other carpet materials at the joints between adjacent lengths of carpet, and there are performance specifications for these bonded assemblies set by the national or international sports governing bodies, such as rugby, football, or tennis. The durability of these installations is assessed in the laboratory by accelerated aging of test bonds of the substrates involved using elevated temperature and water immersion as described in EN 13744 (2004). Test joints are immersed in a water bath maintained at 70  C for 336 h (14 days) before the physical properties of tensile and peel strength are determined by the standard method described in EN

1022

J. D. Palmer

12228 (2002) and compared with the properties determined with un-aged joints. In this example, the durability requirements are deemed to be satisfied if the minimum performance requirements are exceeded before and after accelerated aging. In other applications, a maximum allowable percentage deterioration in physical properties will constitute the specification requirement. Adhesives for ceramic tiles are specified by the standard EN 12004 (2007) (Adhesives for tiles – Requirements, evaluation of conformity, classification, and designation) and include performance requirements for cementitious, dispersion, and reaction resin adhesive types for all internal and/or external walls and floor and ceiling applications. Annex ZA, of this standard, sets out the requirements, which address the provisions of the Construction Products Directive (see Sect. 4 below) and include assessments of durability. Compliance with these requirements confers a presumption of fitness for purpose for the applications and form part of a CE mark. In these cases, durability is assessed by elevated temperature and/or water immersion accelerated aging or freeze-thaw cycling of test specimens and then determining tensile or shear adhesion strength against minimum requirements. Saline solutions are also commonly employed by immersion or spray to mimic sea, coastal, or road conditions. Initially developed to evaluate the corrosion resistance of coatings for metals, the salt spray test can be used to evaluate the durability of adhesive- and sealant-bonded assemblies. The apparatus for testing consists of a closed testing chamber, where a salted solution, most commonly a solution of sodium chloride, is sprayed by means of a nozzle. This produces a corroding environment in the chamber, and thus, pieces in it are attacked under this severe corroding atmosphere. Testing periods can range from a few hours (e.g., 8 or 24 h) to more than a month before the determination of physical properties for comparison with their initial values. Chamber construction, testing procedure, and testing parameters are standardized under national and international standards, which describe the necessary information to carry out the test and give the testing parameters such as temperature, air pressure of the sprayed solution, preparation of the spraying solution, concentration, pH, etc. Salt spray testing is popular because it is cheap, quick, well standardized, and reasonably repeatable. There is, however, only a weak correlation between the duration in salt spray test and the expected life in service, since the corrosive processes involved are very complicated and can be influenced by many external factors. Nevertheless, salt spray test is widely used in the automotive, construction, and aerospace industries.

35.3.4 Stress and Fatigue Stress- or fatigue-induced degradation of adhesive- and sealant-bonded assemblies in nonstructural applications is often characterized by cyclic loading of a test assembly. Polymeric materials, including adhesives and sealants, are viscoelastic in nature, and their responses in such tests tend to be time and frequency dependent. Higher frequencies often induce localized heat, resulting in temperature changes that

35

Durability of Nonstructural Adhesives

1023

can dramatically affect the results, and slow frequencies can also result in poorer performance than expected due to creep rupture. Therefore, care is needed when specifying the temperature and frequency of testing regimes (Dillard 2005). The lower-frequency “creep” or “creep-rupture” tests are time dependent. In these tests, joints are subjected to a nominally constant load. The tests may proceed for a chosen time period or may be continued until complete rupture occurs. For example, the durability of thermoplastic adhesives for nonstructural wood applications is classified in European standards EN 204 (2001) and EN 205 (2003) according to their ability to withstand various water treatments and relatively rapidly applied loads. However, an additional characteristic that can be specified is resistance to static load, which can be determined using the method described in EN 14256 (2007). This method is used to determine the ability of a test piece bonded with a thermoplastic adhesive, to support a given load for up to 21 days without fracture or excessive distortion, and specifies a mean survival time of 14 days.

35.3.5 Assessment by Combinations of Factors Some durability evaluations use a combination of these elements of accelerated aging. For example, an indication of the durability of adhesives used for bonding thermoplastic pressure pipe systems is given in EN 14814 (2007), Adhesives for thermoplastic piping systems for fluids under pressure – Specifications, which references the test method given in ISO 9311-3, where test pipe assemblies incorporating adhesively bonded joints are subjected to internal hydrostatic pressures up to 3.2 times the nominal maximum operating pressure for 1,000 h without leaking. Passing this test regime is a requirement of the initial type tests to demonstrate compliance with the specification but is also expected to give an indication of the “in-service” performance of the system.

Insulating Glass Units Lowe et al. (1994) studied the water durability of insulating glass units manufactured with polysulfide sealants by accelerated aging of glass to sealant adhesive joints at 95% relative humidity in an environmental chamber set at 60  C for up to 14 months. Among their conclusions were that the glass to sealant joints were weakened by this test regime; that the sealants absorbed large amounts of water, which attacks at the interface between the glass and sealant; and that the amount of interfacial failure increases with exposure. Studies such as this, in conjunction with those undertaken by sealant and raw material manufacturers, have been instrumental in the development of the standard for insulating glass units (double glazed units) EN 1279 (a standard in six parts published between 2002 and 2008) which describes how to assess durability by accelerated aging of test double glazed units. The function of the polymer sealant utilized to seal the edges of the insulating glass units is to hold the unit together and prevent the penetration of moisture into the unit and escape of insulating gases placed between the panes. Durability is evaluated by exposure to temperature and humidity cycles, which attempt to replicate the actual in-service

1024

J. D. Palmer

exposure conditions of glazing units in buildings, that is, elevated temperature combined with high humidity, along with low-temperature cycles, and interspersed with periods of longer duration with exposure to high humidity. The efficiency of the seal is then determined by measuring the amount of moisture that has penetrated into the unit, by gravimetric analysis of the desiccant, which is contained within the hollow aluminum spacer bar that forms the gap between the panes in the construction of the unit. The rate of escape of insulating gas (usually argon or sulfur hexafluoride) through the sealant is also determined using gas chromatography by containing the test unit in a rig and sweeping the annular space around the unit with air into the chromatograph. The standard specifies maximum penetration levels and rates for both test regimes. Nonstructural sealants for use in construction works are classified by their performance characteristics as determined according to ISO 11600. Sealant manufacturers, who adopt this classification scheme, use designations, which indicate the sealants end use (either general construction, type F, or glazing, type G), its movement accommodation capability (in the range 7.5–25%), its modulus (defined as high or low), and its elasticity or plasticity, totaling 11 sealant classes, with all these performance characteristics determined using specified test methods. This classification system is a helpful aid to specifiers who can match their design requirements to the sealant performance characteristics. Within Europe, this classification system is to be augmented with the affixation of a CE mark to indicate suitability for purpose and the performance requirements, including durability, for building applications defined in new standards currently being developed (FprEN 15651 parts 1 to 5: 2009). The durability clause in these new standards indicate that the sealant performance determined by classification to the appropriate ISO 11600 specification will indicate suitable durability of the sealant after aging. It also notes that nonstructural sealants are easy to use, ready to use, and easy to remove when any unavoidable or unexpected degradation arises.

Electronics There are numerous applications for adhesives in electronics manufacturing industries, some bringing additional property requirements for adhesive formulators, particularly those of conductivity and heat and stress dissipation. Licari and Swanson (2005) give a complete exposition on the subject. To exhibit electrical conductivity, the adhesives typically employed, such as epoxy resins, need to be filled with electrically conductive particles, for example, metal powder like silver, because they are inherently insulators. The selection of an electrically conductive (or insulating) adhesive, therefore, is most often based on the adhesives’ conductivity or resistivity and its ability to retain those properties for the expected service life of the electronic equipment. Accelerated aging tests have been developed by industry that can give indications of durability and also be used for qualification of adhesives. Typical regimes involve heat aging for 1,000 h at 150  C, combined with any or all of the following: moisture exposure, thermal cycling, vibration, and thermal or mechanical shock, with the

35

Durability of Nonstructural Adhesives

1025

performance expectation that bond strengths and electrical properties will be maintained throughout. One major cause of failure in this field is large mismatch in the coefficient of thermal expansion of the substrates concerned, especially when bonded with high modulus adhesives, and so adhesives have been formulated from more elastomeric resins, which can effectively dissipate the stresses induced in these situations and can be demonstrated by the accelerated test cycles described above.

35.4

Durability and the Construction Products Regulation

The largest commercial market for adhesives and sealants is their use in building and construction. The Construction Products Directive (89/106/EEC) (CPD) is currently undergoing review and will be remade as the Construction Products Regulation (CPR) probably coming into force in 2012 or 2013. European regulations, unlike directives, are translated directly into law in the member states without the need for national statutes and instruments (and the potential for differing national interpretations). Thus, the basic work requirements (formerly essential requirements in the CPD) of the CPR will become compulsory in all member states. The CPD (and CPR) recognize that member states have provisions, including requirements, not only to building safety but also to health, durability, energy economy, protection of the environment, and other aspects important in the public interest. The articles of the Regulation require that these basic work requirements are satisfied during an economically reasonable working life, where the working life is the period of time during which the performance will be maintained at a level compatible with the fulfillment of the basic work requirements and an economically reasonable working life presumes that all relevant aspects are taken into account, such as costs of design, construction, and use; costs arising from hindrance of use; risks and consequences of failure of the works during its working life and costs of insurance covering these risks; planned partial renewal; costs of inspections, maintenance, care, and repair; costs of operation and administration; disposal; and environmental aspects. For adhesive and sealant materials, their durability in service, in relation to these basic requirements, is connected with the characteristics of the products, and the mandates issued by the European Commission for the development of harmonized technical specifications (harmonized European Norms – hEN), that is, standards and, for certain products, European Technical Approvals (ETAs), also include durability aspects. As a passport for a product to be marketed and used in the single market without the need for any other additional requirements, a CE mark may be affixed when it can be demonstrated that all the provisions of the CPR including these basic work requirements have been satisfied. Thus, standards and technical specifications must include provisions for the assessment of durability. A main route to this assessment involves the performance testing of a product to determine the variation in its characteristics under a given action or cycle of actions. However, the standard and

1026

J. D. Palmer

specification writers are required to strike a balance between the cost of testing and the additional information that can result from such tests. For products that are covered by hENs, the specific durability assessment requirements are stated in Annex ZA of the standard, and the results of that durability assessment form part of the aforementioned CE mark.

35.4.1 Sustainability The increasing concerns about the impact of the activities of man on the environment have brought new challenges to adhesive and sealant manufacturers as customers have begun to specify requirements relating to the sustainable consumption and production of the earth’s resources. Such considerations relate not only to the raw materials used to manufacture the adhesive and sealant products but also to the end of life options, including recycling possibilities, of the products and articles made with them and their ultimate destination and impact on the environment. In order to minimize all these impacts, full life cycle analysis techniques are employed to identify where improvements can be made. As adhesives and sealants are most commonly employed to join dissimilar materials, so the ability to de-bond assemblies to facilitate recycling activities has become an additional consideration for the formulator and a new area for research workers and adds another new dimension to considerations of durability in-service, with the now required coexistence of durable, yet dismantle-able properties. The dismantling and recycling of electrical appliances have received much publicity in the western world, particularly in relation to the exposure of workers in the developing world to the toxic chemicals, which are contained within them and are of relatively high value for recycling purposes. However, the promotion of waste recycling in the construction industry particularly in the west is becoming increasingly important, because of the dwindling availability of landfill sites. The building materials currently used in the market utilize a wide range of components from metals to inorganic and organic substances often to improve their strength, function, or aesthetic appearance, but their separation is often difficult to achieve, and combined materials cannot be recycled or reused. Obviously, in any manufacturing sector, if adhesively bonded joints can be made to separate easily by some method, the recycling of such combined materials becomes much easier, and new technologies have been developed in an attempt to satisfy this requirement. One of these technologies recognizes that hydrogen bonding and intermolecular forces are the main forces that join adherends and a method that extends intermolecular distance will facilitate bond dismantling. This is achieved by incorporating thermally expansive particles into the adhesive or sealant, such that when they are expanded by heating, typically above 100  C, the intermolecular distance is extended by the expansive forces, bond strength is reduced, and the materials are separated. Ishikawa et al. (2005) conducted a study on these types of adhesive for use in the construction industry. They examined the durability of the bonded assemblies by accelerated aging tests at 60  C and found that there is no deterioration in bonding performance

35

Durability of Nonstructural Adhesives

1027

and that joints could be easily separated by heating between 150 and 180  C. From these studies, they were able to conclude that durability and dismantle ability were able to coexist in the systems tested.

35.5

Conclusions

This chapter describes and discusses the main factors affecting the durability of nonstructural adhesives and sealants, and the methods used for their estimation, and their applicability to actual service life prediction. The major conclusions that can be drawn are the following: 1. Appropriate adhesive and sealant selection, which takes into account the properties of the materials to be joined and the expected service conditions and service life, is essential to ensure the durability of bonded assemblies, and much advice and assistance is available from manufacturers. 2. Standard test procedures have been developed to assist this selection process and predict durability in the absence of “heritage” data to establish suitability for purpose. However, these standard test methods rely on acceleration techniques to produce deterioration of properties in reduced timescales, which may induce mechanisms and pathways that are not present under “normal” service conditions. So these tests need to be supported by determinations conducted under “normal” service conditions. 3. Durability and its estimation are becoming increasingly important in all industrial manufacturing sectors for the calculation of economic service life of products including their human health and environmental impacts and cost of maintenance. This is particularly relevant in the construction industry.

References Adams RD, Comyn J, Wake WC (1997) Structural adhesive joints in engineering, 2nd edn. Chapman & Hall, London Approved Document to support Regulation 7 – Material and workmanship (1999) (H.M.S.O) Ashcroft IA, Digby RP, Shaw SJ (2001) J Adhes 75:175 BASA (1999) Guide to the ISO 11600 classification of sealants for building construction BASA (2008) Industry guide to the professional application of construction sealants on site BS 3046 (1981) Specification for adhesives for hanging flexible wallcoverings BS EN ISO 11431 (2002) Building construction. Jointing products. Determination of adhesion/ cohesion properties of sealants after exposure to heat, water and artificial light through glass Critchlow GW (2005) In: Packham DE (ed) Handbook of adhesion, 2nd edn. Chichester, Wiley Dillard DA (2005) In: Packham DE (ed) Handbook of adhesion, 2nd edn. Chichester, Wiley EN 12004 (2007) Adhesives for tiles. Requirements, evaluation of conformity, classification and designation EN 12228 (2002) Surfaces for sports areas – determination of joint strength of synthetic surfaces

1028

J. D. Palmer

EN 1279 parts 1 to 6 (2002 to 2008) Glass in building. Insulating glass units EN 13744 (2004) Surfaces for sports areas – procedure for accelerated aging by immersion in hot water EN 14256 (2007) Adhesives for non-structural wood applications – test method and requirements for resistance to static load EN 14814 (2007) Adhesives for thermoplastic piping systems for fluids under pressure. Specifications EN 1903 (2008) Adhesives – test method for adhesives for plastic or rubber floor coverings or wall coverings – determination of dimensional changes after accelerated ageing EN 204 (2001) Classification of thermoplastic wood adhesives for non-structural applications EN 205 (2003) Adhesives. Wood adhesives for non-structural applications. Determination of tensile shear strength of lap joints European Commission, Guidance Paper F (2004) Durability and the construction products directive FprEN 15651 parts 1 to 5 (2009) Sealants for non-structural use in joints in buildings and pedestrian walkways Ishikawa H, Seto K, Shimotuma S, Kishi N, Sato C (2005) Int J Adhes Adhes 25:193 ISO 11600 (2003) Building construction. Jointing products. Classification and requirements for sealants ISO 15908 (2002) Adhesives for thermoplastic piping systems. Test method for the determination of thermal stability of adhesives ISO 9142 (2003) Adhesives – guide to the selection of standard laboratory ageing conditions for testing bonded joints IUPAC (1997) Compendium of chemical terminology, 2nd edn Licari JJ, Swanson DW (2005) Adhesive technology for electronic applications. William Andrew, Norwich Lowe GB, Lee TCP, Comyn J, Huddersman K (1994) Int J Adhes Adhes 14:85 Petrie EM (2004) Antioxidants for adhesives – web publication by SpecialChem Petrie EM (ed) (2007) Handbook of adhesives and sealants, 2nd edn. New York, McGraw-Hill Scott G (1998) Stabilisation of adhesives against environmental peroxidation – Society for Adhesion and Adhesives. One day meeting on ‘Stability of Adhesives’ Zhao ZY, Zhang WL (1997) Int J Adhes Adhes 17:47

Part VII Manufacture

Storage of Adhesives

36

Hans K. Engeldinger and Cai R. Lim

Contents 36.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.1.1 Basics of Adhesive Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.1.2 Classification of Adhesives Based on Storage Criterion . . . . . . . . . . . . . . . . . . . . . 36.2 Storage at Room Temperature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.2.1 Solvent-Based Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.2.2 Water-Based Dispersion Adhesive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.2.3 Hot-Melt Adhesive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.2.4 Reactive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.3 Storage at Low Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.3.1 Basics of Adhesive Storage at Low Temperatures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.3.2 Temperature and Reactivity: The Arrhenius Equation . . . . . . . . . . . . . . . . . . . . . . . . 36.3.3 Storage Between Freezing Point and Room Temperature . . . . . . . . . . . . . . . . . . . . 36.3.4 Storage Below Freezing Point . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1032 1032 1033 1034 1035 1037 1038 1040 1043 1043 1046 1048 1048 1049 1049

Abstract

A chapter on adhesive storage should be closely aligned with the current market practice. As such, a major part of this chapter focuses on the shelf life and safety aspect of adhesives. The adhesive shelf life is dependent on the adhesive system and the storage conditions, in particular, the temperature. For a stronger emphasis on the effect of temperature on adhesive shelf life, this chapter is subdivided into Storage at Room Temperature and Storage at Low Temperature. H. K. Engeldinger (*) Product Development HAF Tapes, tesa SE, Norderstedt, Germany e-mail: [email protected] C. R. Lim Research & Development, tesa SE Hamburg, Norderstedt, Germany e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_36

1031

1032

H. K. Engeldinger and C. R. Lim

Due to the flammability and possible health hazards of solvent-based adhesives and majority of the chemically reactive adhesives, their storage usually requires special safety measures. Adhesive manufacturers are required to provide information on the shelf life and storage conditions in the product technical data sheet, as well as a description of the possible hazards in the safety data sheet (Material Safety Data Sheet, MSDS), e.g., in the event of a fire. Due to the wide range of adhesive systems available, it is not possible to devise a standard storage guideline applicable to all adhesives. In this chapter, adhesives are divided into four categories: solvent-based, water-based, hot-melt, and reactive adhesives. For some adhesives, their physical characteristics should also be considered (e.g., liquid or solid in the form of powder, granules, or film). Considering the variety of adhesive systems, it is necessary to provide a short introduction on the adhesive to indicate the type of packaging and storage conditions needed and its potential hazards. Similarly, in cases where the adhesive shelf life may be extended through storage at low temperatures, it should also be indicated on the packaging, although this applies to limited adhesive systems and applications. In order to explain the quantitative effect of low temperatures on adhesive storage, a brief subtopic on activation energy and temperature-dependent chemical reaction (Arrhenius equation) is provided.

36.1

Introduction

36.1.1 Basics of Adhesive Storage The proper storage of an adhesive is a prerequisite to maintain the adhesive properties for the adhesion process and, beyond that, for the final bond performance (bonding strength, heat resistance, and other specific properties). The significance of the appropriate storage conditions is usually presented in the quality assurance documentation of the adhesive. Essentially, the adhesive is guaranteed by the manufacturer throughout the shelf life under the storage conditions specified in the product technical data sheet. The optimization of the adhesive shelf life normally commences with the adhesive development, with storage stability tests integral to the design specifications. As a rule, manufacturers are required to inform customers of the adhesive shelf life under the recommended conditions in the product information. Within this period, the specified adhesive properties should be guaranteed for delivery as well as for the final bond performance after appropriate processing steps. The manufacturing date, shelf life, and suitable storage temperature of the adhesive should also be indicated on the packaging. While many products can be stored at room temperature, some may require low temperatures for a prolonged shelf life. The aspect of storage at low temperatures will be addressed separately.

36

Storage of Adhesives

1033

Besides complying with international and national laws and regulations, institutional measures to ensure environmental protection and occupational health and safety must also be adhered to during adhesive storage. They are applicable throughout the supply chain, from production to delivery by the adhesive manufacturer, to distribution by the agent if necessary, and, finally, to the end user after processing. In the European standard, EN 12701 (European Standard 2001) titled “Determination of words and phrases relating to the product life of structural adhesives and related materials,” definitions and requirements applicable to structural adhesives with storage life limited by possible change of properties are clearly specified. The compliance with safety procedures constitutes another notable aspect of adhesive storage. A series of adhesives are classified as hazardous materials due to the presence of highly flammable solvents or components presenting health hazards such as cyanoacrylates or acrylates containing high monomer contents. In addition, hardeners of reactive adhesives such as isocyanates, amines, or peroxides should be considered. An important reference is the REACH (Registration, Evaluation, Authorization and Restriction of Chemicals -EC 1907/2006) regulation, implemented by the European Parliament in 2007 to ensure that manufacturers take the responsibility to manage risks from chemicals and to provide safety information on the chemicals. It is also an international practice to include exposure controls and other measures in the safety data sheet, so as to ensure the safety of personnel handling of the adhesives. The manufacturers are responsible for compiling the necessary documents and are required to provide them to customers at all times. Special provisions and measures regarding the storage facilities of adhesives are established according to building and environmental regulations. Among them, fire control, which includes firefighting equipment, evacuation plan, and fire engine access route, is an important point for consideration. At the same time, attention should be accorded to the prevention of water pollution which may be caused by the leaching of hazardous materials into groundwater. This chapter mainly covers the aspect of adhesive storage, according to the industrial standard. The storage of adhesives in households or offices, usually in small quantities (e.g., in tubes), is mostly subjected to other conditions and regulations.

36.1.2 Classification of Adhesives Based on Storage Criterion Due to the existence of different adhesive systems and the potential hazards associated with each system, there are different types of packaging as well as storage conditions and shelf life. Adhesives can be classified according to specific criterion and properties. In the literature, one can find various classifications based on the assembly process, delivery form, adhesion mechanism, or application. Within the same adhesive group, the adhesives can be further categorized based on their physical states or characteristics: liquid, paste, or solid (through viscosity measurements and rheological characterization) or solvent based, water based, or those

1034

H. K. Engeldinger and C. R. Lim

without volatile content (through solid content measurements). Another classification can be made based on the adhesion mechanism: physically dried adhesives from solution, solidified hot-melt adhesive, or chemically cross-linked single- and two-component reactive adhesives. This chapter attempts to classify the different adhesives based on a few key points concerning storage. Based on storage criterion, adhesives can be classified into four main groups: solvent-based, water-based, hot-melt, and reactive adhesives. These groups mainly differ in their shelf life and potential hazards. Nonetheless, within such classifications, overlaps are possible. For instance, there are reactive hot-melt adhesives and reactive solvent-based adhesives. Other considerations include the delivery form and packaging, which differentiate solid adhesives from liquid adhesives. Liquid adhesives are usually stored in sealed packaging, containers, barrels, or tanks at room temperature between 18
C and 25
C. Homogeneous products can be stored for a year or more due to the limited physical or chemical changes within the system. For nonhomogeneous products, due to the possible sedimentation of fillers or segregation of different phases during storage, stirring is usually necessary to return the system to its original state before use. Any premature chemical reactions caused by inappropriate storage conditions (e.g., higher temperature or longer duration) are irreversible, and they usually render the adhesives ineffective or unusable. Solid adhesives such as hot melts are relatively easier to handle during packaging and storage, whether they are in the form of blocks, in crushed form as granules, in powdered form in cartons or sacks, or in the form of rolled sheets. During the storage of reactive adhesives, the influence of humidity is another important consideration besides temperature, especially for moisture-sensitive bonding agents such as cyanoacrylates or single-component polyurethanes, which harden upon contact with moisture, like hardeners of reactive adhesives. Among special reactive adhesives are those which react under anaerobic conditions or in UV light and require storage under oxygen-rich or dark conditions, respectively. The packaging of such adhesives should also be adequate to prevent any undesired chemical reactions before use. Reactive adhesives which are particularly sensitive require storage in a conditioned environment, with regulated temperature and humidity. To achieve fast adhesion during mass production, highly reactive adhesives containing catalysts are often required. However, they are not suitable for room temperature storage. Instead, low temperatures are required. Special measures of low-temperature storage and the affected adhesive groups will be addressed in Sect. 3.

36.2

Storage at Room Temperature

The term “room temperature” is sometimes misused. It was meant to define the temperature in a conditioned room, usually enclosed, such as an office or a work area, e.g., laboratory. The temperature range may vary according to the location, e.g., as narrow as 20–23
C in a measurement laboratory or as wide as 18–25
C in rooms

36

Storage of Adhesives

1035

where temperature sensitivity is not as crucial, such as storage hall for general goods. Unless specified, room temperature mentioned in this chapter refers to a temperature range of 18–25
C. Additionally, it is a basic condition in adhesive storage to keep the adhesive in its original sealed packaging before use, as far as possible, even if humidity does not always appear to influence the adhesive shelf life.

36.2.1 Solvent-Based Adhesives Solvent-Based Adhesive Group This class of adhesives consists of solutions of polymers, resins, and, where applicable, other additives in organic solvents. They have low to moderate viscosities and solid contents ranging from 15% to 40% in weight, which correspond to solvent contents ranging from 60% to 85% in weight. They are applied as PSA or contact adhesives. The base polymer should be soluble in common solvents. The rate of dissolution is dependent on the polarities of both the base polymer and the solvent and the molecular weight of the polymer. Predominantly, acrylate, polyvinyl acetate, natural or synthetic rubber, polychloroprene, polyurethane, polyvinylpyrrolidone, or polyvinylchloride (PVC) constitute the base polymer for such adhesives. Nonetheless, solutions combining phenolic resins and rubber are also used. Common solvents used are those which are easily available and cost effective. They should have a good solubility but low toxicity. Therefore, the use of chlorinated solvents such as dichloromethane or trichloroethylene has been effectively banned in the EU since December 2010 despite their good solubility, as they may be carcinogenic or pose health risks (e.g., optic neuropathy). In general, the following solvents are used: petroleum naphtha 60/95, petroleum naphtha 80/110, ethyl acetate, acetone (dimethyl ketone), butanone (methyl ethyl ketone), toluene (methylbenzene), ethanol, isopropanol, or sometimes tetrahydrofuran, which is a suitable solvent for PVC. For adhesive storage, key data of the corresponding solvents such as boiling point, flash point, and the rate of evaporation are necessary. Table 1 shows a summary of these data for the solvents mentioned above. (Flash point refers to the lowest temperature at which a volatile liquid can vaporize to form an ignitable mixture in air.) The rate of evaporation is a ratio of the time required to evaporate a measured amount of liquid to that of a reference liquid (ethyl ether). Storage and Shelf Life Storage vessels of solvent-based adhesives which include jerry cans, drums, and containers with capacities of up to 30, 200, and 1,000 L, respectively, are made of metals such as aluminum or steel. Drums either have a sealable lid or are sealed but equipped with a faucet outlet. Sealed drums should only be stored upright. Otherwise, the drum lid may be easily damaged on contact with adhesives, especially those containing aggressive solvents, causing the contents to leak. Plastic vessels are

1036

H. K. Engeldinger and C. R. Lim

Table 1 Properties of solvents used in solvent-based adhesives Solvent Acetone Butanone Ethanol Ethyl acetate Isopropanol Tetrahydrofuran Toluene

Boiling point (
C) at 1,010 mBar 56 80 78 77 82 66 111

Flash point (
C) according to DIN EN 22719 19 1 +12 4 +12 17 6

Rate of evaporation ethyl ether = 1 2.1 6 62 2.9 11 2.3 6

not suitable as storage containers for solvent-based adhesives, as solvents may diffuse through the container walls – unless they are protected by a barrier coating. Solvent-based adhesives are generally stable during storage, as they hardly undergo alterations as homogeneous solutions in sealed vessels. Hence, they usually have a recommended shelf life of up to 2 years. However, for some, the viscosity may change gradually with time. Though uncommon, some solvent-based adhesives may contain fillers or reactive components, which affect the storage stability through phase separation or cross-linking.

Hazard Potential Solvent-based adhesives are classified as hazardous goods, as they contain a significant portion of highly flammable solvents. Thus, it is essential to comply with laws and regulations concerning fire prevention and firefighting. Adhesive containers should be labeled with the appropriate hazard symbols. In storage facilities, no ignition sources should be present or allowed to come into contact with solvents or their fumes, and the generation of sparks should be avoided. Additionally, it is necessary for all electrical equipment to be explosion proof and all containers should be earthed. It is also mandatory for personnel to be equipped with safety shoes having conductive soles to minimize the accumulation of static electricity. Similarly, production equipment or vehicles should not generate any sparks. The maximum quantity of solvent-based adhesives allowed in a storage room is limited (to 10,000 L in Germany). In an “active storage,” where the containers with adhesives would be opened or siphoned, an exhaust system must be installed to minimize the accumulation of solvent fumes. Secondary containment is often necessary to prevent the adhesives from entering the earth or groundwater in the event of a spill as shown in Fig. 1. Further information on this aspect may be obtained from national or international regulations such as REACH (the new European law concerning chemicals) or specialized contributions from corporate entities such as Fundamentals of Hazardous Materials by Denios, Germany Fundamentals of Hazardous Materials (2009/ 2010).

36

Storage of Adhesives

1037

Fig. 1 Storage of flammable solvents and solvent-based adhesives in a protective storage chamber (Photo: Denios)

36.2.2 Water-Based Dispersion Adhesive Water-Based Adhesive Group Water-based dispersion adhesives consist of oligomers dispersed in water with surfactants as dispersing agents and stabilizers. Before dispersions, natural rubber latex and casein were used to manufacture such adhesives. Currently, polyvinyl acetate (PVAc) and polyacrylate serve as synthetic base polymers for water-based PSA, while polychloroprene and thermoplastic polyurethane serve the same purpose for water-based contact adhesives. Compared to solvent-based adhesives, water-based dispersion adhesives have several advantages: environmentally friendly and nonflammable. In addition, a larger quantity of the adhesive can be dispersed in water without a significant increase in the viscosity. Thus, a solid content of 40–70% in weight is common, and this is almost twice of that compared to solvent-based adhesives. Storage and Shelf Life Water-based dispersion adhesives are mostly stored in plastic containers. Smaller containers with capacities of up to 30 L are made of polyethylene (PE), while larger containers which come in 100-L vessels or up to 800-L containers are made of polypropylene (PP). To ensure stability, the container thickness should be

1038

H. K. Engeldinger and C. R. Lim

proportional to its size. In fact, such containers are often made of glass fiberreinforced polypropylene. Larger containers can also be stabilized using metal casings or cages. However, one drawback of water-based dispersion adhesives is the limited storage stability due to various influencing factors. In particular, they must be protected from frost, as freezing usually causes irreversible damage to the dispersions. Furthermore, the stability of such dispersions is pH dependent. In fact, many of them (except PVAc dispersions) coagulate in neutral or acidic conditions, leading to the failure of the adhesive system. This can be avoided by the addition of neutralizing agents such as ammonium hydroxide or alkali hydroxides with pH values in the range of 9–10. Water-based dispersion adhesives also face the problem of possible contamination through microbial attacks, especially when industrial water or mains water is used during their production. During microbial attacks, dispersions undergo partial coagulation, resulting in brownish discoloration and a strong odor. With that, the adhesive becomes unusable. Thus, a microbiological analysis is necessary to determine the bacteria count before storage. With that in mind, most manufacturers of water-based dispersion adhesives recommend storage durations of 6–12 months at 5–25
C.

Hazard Potential As water-based dispersion adhesives do not pose any health hazards and are nonflammable, there is no critical concern associated with their storage. Despite so, precautions relating to the concerned chemicals and information provided in the safety data sheets should be taken note of, in the event of a spill. Contact with the skin and splashes into the eye should also be avoided, like most chemicals. Upon contact with chemicals such as ammonia, some dispersion adhesives emit an odor. For environmental concerns, contaminants such as alkaline or acidic materials should not be discharged into the groundwater.

36.2.3 Hot-Melt Adhesive Hot-Melt Adhesive Group Hot-melt adhesives range from viscous to solid forms in storage at room temperature. To allow proper application during the adhesion process, they have to be melted at elevated temperatures. Following that, they solidify on cooling to form strong bonds between the substrates. Polymers such as ethylene-vinyl acetate, polyamide, and polyester constitute the base polymers of harder hot-melt adhesives with higher melting points, while the softer ones consist of polyacrylates, styrene-butadiene copolymers, or thermoplastic polyurethane. However, as such polymers exhibit lower adhesion and tack properties, they are usually used in combination with tackifiers. While the soft and viscous hot-melt adhesives exhibit better tack than solid ones, the solid hot-melt adhesives provide a stronger bonding. There are different delivery forms available: soft hot-melt adhesives are preferably packed into compact blocks or cartons, while highly viscous hot-melt granules are usually powdered with fine mineral powder such as talcum or silica to inhibit the

36

Storage of Adhesives

1039

clumping of the granules. This allows them to remain fit for storage. In contrast, as solid hot-melt adhesives do not form clumps easily, they are available in the form of powder, granules, or to some extent, sticks. Hot-melt adhesives also exist as films on release paper, winded into rolls. Such films can be precut into sheets, which are used immediately for mass productions, instead of being stored for later use.

Storage and Shelf Life During manufacture, hot-melt adhesives are mostly filled in their molten state at high temperatures, after which they solidify upon cooling. They are available in compact forms in 200-L vessels or in cartons, which are coated with release agents, and are finally delivered as blocks. Pressure vessels are often used as special storage containers as they are suitable for adhesive storage under pressure, after the heating and melting of the adhesives. Many of these adhesives undergo crystallization during cooling after manufacture. However, this does not impact the adhesion properties, as crystallization is reversed during the subsequent heating process before bonding. Such reversible behavior is characteristic for thermoplastic materials. Powdered and granulated hot-melt adhesives are usually prepared by crushing the solid form and then preferably packed and stored in 25-kg paper sacks. As hot-melt adhesives are sensitive to moisture, the paper sacks should be coated with polyethylene (PE) to prevent deterioration by the diffusion of moisture through the paper sacks. Polyamide is not used as a moisture barrier because it can absorb up to 3% weight of water during extended storage. Polyester is also not an option due to similar reason, though it has a lower moisture absorption capacity. The larger the surface area of the particle or smaller the particle size, the faster is the rate of absorption. During subsequent heating between 150
C and 180
C, the presence of moisture in the adhesive leads to the formation of bubbles or foam in the adhesive joint, which reduces the adhesive stability and leads to quality problems. During production of hot-melt films, hot-melt adhesives undergo heat extrusion and cooling on release-coated paper or film. When soluble polymers such as thermoplastic polyurethane are used, the solution can be coated onto the release paper, dried, and converted into a larger roll, before it is slit into smaller rolls. A large roll may measure 250 μm in thickness (with release paper), 2000 m in length, 2 m in width, and weigh up to 1 t. Hot-melt adhesives tend to flow under high pressure loads even at room temperature. Thus, large rolls should only be stored for a short period of time on a flat support or upright manner, otherwise deformation would occur. For longer storage durations, the problem may be avoided by fixing the roll on a rotating rack, which allows it to slowly rotate continuously, as shown in Fig. 2. For adhesive tapes with liner, regardless of the size of the rolls, humidity should be taken into consideration. At a high humidity, the paper within the liner absorbs moisture from the environment despite the silicon interface, especially after drying. When this occurs, the paper tends to swell. This affects the structure of the paper, which may then be transferred to the adhesive film. However, this concerns only the outermost layers of the roll. This effect can be avoided either by air-tight packing of individual rolls or the conditioning of the entire storage room to a temperature of 20–23
C and a humidity which is lower than 50%.

1040

H. K. Engeldinger and C. R. Lim

Fig. 2 Rotating racks for adhesive jumbo rolls in a temperature-regulated storage hall (Photo: tesa Werk Offenburg)

In general, by adhering to the moisture-related measures, hot-melt adhesives can be stored unchanged for a long period of time. The recommended shelf life of hot-melt adhesives ranges from 18 to 24 months, provided they are stored at 10–25
C, under dry conditions.

Hazard Potential Since most hot-melt adhesives do not contain volatile components, they are not highly flammable and do not pose any health hazard during normal handling. As such, they are also used in packaging for grocery. Unlike the rest, no special measures are required for an incident-free storage.

36.2.4 Reactive Adhesives Reactive Adhesive Group Reactive adhesives comprise all adhesives which undergo a chemical reaction during the adhesion process and in doing so, solidify. Single-component systems can react chemically in the presence of external influences such as heat, UV light, moisture, or anaerobic conditions. On the other hand, two-component systems usually react through the mixing of reactive components, e.g., resins and catalysts, at room

36

Storage of Adhesives

1041

temperature without external influences. During storage and before processing, reactive adhesives often exist in forms ranging from liquid to paste. The most important types of two-component adhesives are moisture-sensitive polyurethane, comprising polyol resin and isocyanate catalyst, and heat-sensitive epoxy, comprising bisphenol-A or bisphenol-F resins and polyamide catalyst. Both single- and two-component adhesive systems are available in the market. Singlecomponent reactive adhesives, which include cyanoacrylate and silicone, are also widely known, with the latter often being used as a sealant. A particular adhesive group used for heat-resistant bonding is the singlecomponent reactive adhesive, formulated as a combination of phenolic resin and nitrile rubber, in the form of heat-activated films (HAF). These contain latent hardeners for the phenolic resins, which are activated at temperatures above 120
C. The advantages include clean and precise handling. UV-sensitive and anaerobic-sensitive adhesives also belong to the reactive adhesive systems, which should be stored under chilled conditions.

Storage and Shelf Life of Two-Component Reactive Adhesives Resins and hardeners of two-component adhesives should be stored in separate metal drums, jerry cans, vessels, or larger containers. The quantity of each component to be stored is determined according to their mixing ratio. Polyurethanes and epoxy resins differ in the type of chemical reactions they undergo during the hardening process. On top of that, both systems are used in combination with different hardeners and have different storage requirements. Polyurethanes are traditionally produced from the reaction of isocyanates and polyols. Isocyanates (e.g., methylene diphenyl diisocyanate [MDI]), the most commonly used hardeners for polyurethanes have low viscosities and undergo crystallization in chilled conditions below 15
C. At temperatures above 40
C, isocyanates undergo a chemical reaction (i.e., dimerization), which increases their viscosities. In addition, upon exposure to moisture, they react with water molecules to produce polyurea and carbon dioxide gas, which results in outgassing and the formation of a layer of film. To prevent such occurrence, isocyanates must be stored at temperatures in the range of 20–35
C and in the absence of moisture. Under these conditions, the shelf life is approximately 6 months. Polyols, unlike isocyanates, are generally chemically stable. However, in many instances, they contain additives such as catalysts, fillers, and pigments for special applications. During storage, mineral fillers may sediment at different rates depending on their densities and the viscosity of the bulk solution. In such situations, the material should be stirred immediately before use to ensure homogeneity. However, this should be done without generating entrapped air in the mixture. Hence, it is essential to notify the user with a “Stir before use” label on the container. Epoxy resins comprise bisphenol diglycidyl ethers. Having moderate to high viscosities, they are mostly delivered in jerry cans or vessels. While pure epoxy resins tend to crystallize upon storage for longer durations, this phenomenon can be reduced by

1042

H. K. Engeldinger and C. R. Lim

mixing bisphenol-A with bisphenol-F. Often, such adhesives are also loaded with fillers. Hence, they should be monitored for any sedimentation during storage. Hardeners are mostly polyether amines or similar amine composites. They have moderate to high viscosities and a typical amine smell. The storage of resins and hardeners in sealed containers is less critical compared to that of polyurethanes, as every component is chemically consistent. Thus, they can be stored for 12 months at temperatures between 20
C and 25
C.

Storage and Shelf Life of Single-Component Reactive Adhesives Single-component polyurethanes consist of prepolymers, a fraction of which is pre-reacted polyurethanes with terminal isocyanate groups. Due to their moisture sensitivity, they have to be stored away from moisture or possible contact with water, preferably in metal containers. They require similar storage conditions as MDI hardeners, although in the reduced forms, they are less reactive. In the absence of moisture, the viscosity increases during storage. As such, manufacturers usually recommend a shelf life of 3–6 months at 20–25
C. Single-component epoxy resins contain dicyandiamide as latent hardeners, which are activated at high temperatures – usually above 150
C. As such, they belong to the class of heat-activated adhesives. They are available in viscous or paste forms, and those of even higher viscosities are available as films on release paper. Apart from hardeners, some adhesives in this category also contain an additional activator, with which a faster rate of reaction can be achieved. This effectively shortens the duration needed for the adhesion process or lowers the activation temperature. However, this also leads to reduced storage stability of the adhesive. Heat-activated films (HAF) can be stored for 12–18 months under dry conditions at room temperature, depending on the type of phenolic resins present in the system. For special applications, such as in the electronic industry, clean room environment and subsequent dust-proof packaging and storage are necessary. Cyanoacrylate belongs to the single-component reactive adhesives. Due to its fast-acting characteristic, it is commonly known as “superglue.” In the presence of water, it undergoes rapid polymerization, causing the adhesive to set. Thus, it must be stored in leak-proof containers, completely protected from moisture or water during its storage. Otherwise, it may react prematurely and become unusable. This is guaranteed by the manufacturers through the use of suitable packaging, usually in small quantities from 5 g to 5 kg. When stored in their original packaging in dry conditions at 20–23
C, the shelf life ranges from 6 to 12 months. Anaerobic adhesives are derived from dimethacrylate esters and are often low in viscosities. They react only in the absence of oxygen and when in contact with a metal surface at the same time. In order to prevent premature polymerization, these adhesives should be packed in a way which allows them to be in constant contact with oxygen. Having a diffusion-resistant packaging can also ensure that oxygen does not diffuse out of the packaging during storage. They are mainly used as bolt adhesives or for the sealing and securing of mechanical devices. These adhesives should be stored in a cool environment. For instance, they can be stored at 5
C in their original sealed packaging for 1 year.

36

Storage of Adhesives

1043

UV-sensitive adhesives are based on acrylate-modified epoxy resins, polyurethanes, and, among others, polymers. They are single-component reactive adhesives, which undergo rapid polymerization upon irradiation with UV light with wavelengths ranging from 400 to 500 nm. These adhesives should be stored in small opaque containers of up to 50 mL in volume. They can be stored in their original containers for 6 months at 10–25
C.

Hazard Potential Due to their reactivity, reactive adhesives and hardeners have to be stored according to special regulatory requirements. The containers of such materials should be labeled with the appropriate warning symbols and possible hazards associated with the content, as indicated in the safety data sheets. Nonetheless, under normal storage in their original packaging, they are not expected to pose any hazards. Nonetheless, mandatory precautions to contain a spill due to leaking or damaged packaging are necessary. Such precautions include secondary containment (e.g., trays) and spill kits or absorbents at accessible locations to soak up hazardous liquids for subsequent disposal. Personnel handling spilled adhesives should also be equipped with personal protective equipment (PPE) such as coats or gowns, gloves, safety goggles, and respiratory apparatus. In general, any contact with reactive adhesives should be avoided. While cyanoacrylates bond to the skin in seconds, epoxy resins and polyurethanes can trigger allergies. Other hazardous adhesive raw materials include hardeners used in the manufacture of polyurethanes. One example is methylene diphenyl diisocyanate (MDI), which poses health risks due to occupational exposure such as respiratory tract irritation and sensitization. A more critical situation would be that of a fire. Suitable extinguishing agents or fire extinguishers should be available to combat fires which are small enough. Beyond that, larger fires may lead to uncontrollable chemical reactions with the accumulation of toxic by-products. For instance, single-component epoxy resins may trigger strong exothermic reactions with promoters at temperatures above 120
C. In the presence of more than 1 kg of the materials, such exothermic reactions can hardly be extinguished. As a result, the surrounding may be heated up to temperatures of 500
C or higher. Thus, in the event of a fire, the fire response team must be immediately summoned. The team should also be informed of the materials involved and provided with their safety data sheets.

36.3

Storage at Low Temperatures

36.3.1 Basics of Adhesive Storage at Low Temperatures The main reason for adhesive storage under cool or cold conditions is to extend the adhesive shelf life. For some adhesives, storage at low temperature is absolutely necessary, while for some, it is recommended where longer storage durations are necessary. However, there are also some adhesives which can be damaged when stored under cold conditions.

1044

H. K. Engeldinger and C. R. Lim

For adhesives requiring storage at low temperatures, the same conditions should also apply during transportation from manufacturers to the customers. In the event that the adhesives are exported out of cold or temperate climates to warmer ones, insulated containers are essential. To achieve the desired cooling, filled adhesive containers are stored in cooling chambers or freezers. In doing so, the cooling capacity of the adhesive should be balanced with the cooling efficiency and the temperature of the cooling system. As liquid adhesives are cooled, physical effects such as an increase in viscosity may occur through a reduction in the rate of filler sedimentation. Further cooling may even halt the sedimentation of the fillers. At the same time, solute components may crystallize out of the adhesive. Most importantly, cooling retards the chemical crosslinking reactions, thereby enhancing the adhesive shelf life. The cooling of adhesives for storage followed by reconditioning to room temperature before use usually incurs additional costs, and for larger adhesive quantities, this may be a very time-consuming process. The tempering of full containers or vessels may take up to a few days depending on the temperature difference within the contents, especially when the bulk material is not stirred. Large rolls of reactive adhesive films, requiring storage at low temperatures, have similar behaviors. In practice, the conditioning of cooled adhesives before handling may take longer than expected, as the application of heat has to be conducted with caution. Overheating may lead to unwanted outcomes such as premature cross-linking in reactive adhesives or elevated vapor pressures, which present a flash fire hazard in the presence of solvent-based adhesives. Another effect is the condensation of water vapor on the cooled packaging, as the adhesives are brought into warm conditions. If the adhesives are applied too fast before conditioning, the condensed water droplets may lead to defective bonding between the adhesives and the substrates. These points should be taken into consideration during a good logistic practice. The need for storage at low temperatures only concerns a small group of the commonly available adhesives in the market. In Sect. 1.2, adhesives are categorized into four groups: solvent-based, water-based, hot-melt, and reactive adhesives. In this section, key points of storage at low temperatures will be addressed based on these categories. Most adhesives requiring cool or cold storage conditions are reactive adhesives, many of which are single-component adhesives. For some, cross-linking reactions begin even at room temperature, thus compromising the adhesive shelf life. Apart from fluid adhesives, this problem also concerns reactive adhesive films and prepregs (pre-impregnated sheets). Unlike reactive adhesives, solvent-based adhesives seldom require such storage conditions as the reaction rates of reactive components, such as polyurethanes, epoxies, and phenolic resins, can be distinctly lowered by dilution with solvents. Thus, storage at low temperatures is only occasionally recommended for solventbased adhesives. For normal storage in storage halls with higher ambient conditions, the temperature should be conditioned to a level not beyond the upper limit of room temperature, i.e., 25
C (with the exception of up to 30
C, depending on the solvent) through cooling.

36

Storage of Adhesives

1045

Hot-melt adhesives do not require storage at low temperatures, because being pure thermoplastic materials, they are neither chemically reactive nor do they undergo physical changes at room temperature. Furthermore, they do not emit solvent vapors or other gases during their storage. Thus, at room temperature, these adhesives remain unchanged and can be stored for extended periods of time. In contrast to reactive adhesives, water-based dispersion adhesives generally cannot withstand such storage conditions and must be protected from low temperatures. In such dispersions, water acts as the carrier medium, in which the adhesive components are dispersed as microspheres. Thus, when water freezes at 0
C, phase separation occurs. As the water crystallizes and the adhesive components separate from the liquid medium and condense, the adhesive coagulates, as shown in Fig. 3a, b. When this

a

Water-based acrylic dispersion adhesive

Room temperature (3 h)

–20 °C (3 h) immediate

–20 °C (3 h) after conditioning at RT (1 h)

b

Room temperature (3 h)

–20 °C (3 h)

Fig. 3 (a) Effect of cold storage on a water-based acrylic dispersion adhesive, (b) coagulation of water-based acrylic dispersion adhesive due to cold storage

1046

H. K. Engeldinger and C. R. Lim

happens, the adhesive becomes unusable. Hence, water-based dispersion adhesives are mostly stored at room temperature. Nonetheless, some manufacturers recommend storage temperatures between 5 and 25
C to inhibit microbial attacks.

36.3.2 Temperature and Reactivity: The Arrhenius Equation The shelf life of reactive adhesives can be extended by reducing their reactivity (i.e., rate of collision of reacting molecules per unit time). This can be achieved by increasing the free path of the molecules, through the dilution of the adhesive. However, for the dilution to be effective for this purpose, a reduction of the solid content by approximately 30–40% is necessary. Hence, in practice, dilution is only used as a means to reduce the adhesive viscosity. A more effective way to reduce the reactivity would be to decrease the mobility of the reacting molecules through cooling. The Arrhenius equation describes the dependence of the chemical rate constant on the absolute temperature as follows: EA

k ¼ A  e RT ,

(1)

where k is the rate constant of the reaction. A is the pre-exponent to indicate the number of molecular collisions per unit time. EA is the activation energy of the reaction (J mol1). R is the gas constant, with a value of 8.314 (J mol1 K1). T is the absolute temperature (K1). Most chemical reactions have similar dependence on temperature: the reaction rate doubles with every 10
C increase in temperature and halves with every 10
C drop in temperature. In practice, this could translate, for instance, to an adhesive having a shelf life twice as long when stored at a temperature of 10
C instead of 20
C. Figure 4 shows the relationship between the temperature and the shelf life of an adhesive, with different reaction rate constants. Figure 5 shows the influence of storage time and temperature of a one-component reactive adhesive on the peel strength of bonded electronic composites, from which the corresponding storage shelf life was deduced. In this example, a heat-activated adhesive (epoxy resin/dicyandiamide) was used for bonding polyimide film to stiff printed circuit material. Before its use, the adhesive was stored over a long period of time at 23
C, 15
C, and 5
C. The specified peel strength for the application was 15 N cm1 (L-peel test according IPC-TM-650 2.4.9). The peel strength was measured at an interval of 2 months. Based on the specified peel strength of 15 N cm1, the corresponding shelf life at each storage temperature was recorded, as shown in Table 2. In general, the trend observed fulfills the law of Arrhenius.

36

Storage of Adhesives

1047

Shelf life (months)

100

10

1 –40

–20

Very fast

0 Temperature (°C) Fast

Moderate

20

Slow

Very slow

Fig. 4 Reactive adhesives: shelf life as a function of temperature, with different reaction rate constants

Peel strength (N/cm)

30 25 20 15 Shelf life limit 10 5 0

0

2

4

Storing at 23°C

6

8 10 12 Storage time (months) Storing at 15°C

14

16

18

Storing at 5°C

Fig. 5 Reactive adhesives: influence of storage time and temperature on peel strength

Table 2 Effect of storage Storage temperature temperature of reactive 23
C adhesive on shelf life 15
C 5
C

Shelf life 5.5 months 9 months 16 months

1048

H. K. Engeldinger and C. R. Lim

36.3.3 Storage Between Freezing Point and Room Temperature A temperature range between freezing point and room temperature is adequate to accommodate the majority of the adhesives which require storage at low temperatures. This can be described as “storage under cool conditions” in contrast to storage under 0
C, which is described as “storage under cold conditions.” Some manufacturers or suppliers provide only general storage instructions such as “cool and dry” or “cool and frost-free” instead of the exact storage temperature and duration, while some provide only a temperature range. The wide variety of adhesive types adds yet another dimension to the storage temperature. The following list (Table 3) shows a selection of some of the typical adhesive groups with the appropriate storage conditions.

36.3.4 Storage Below Freezing Point Adhesives which require storage below freezing point are usually stored at temperatures between 20
C and 25
C. Such conditions are generally necessary for reactive adhesives, due to two main reasons as follows. First, for normal reactive single-component adhesives, they are usually stored in large quantities after manufacture for a significant period of time, before they are filled into smaller quantities and delivered to the market. Hence, to eliminate any undesired chemical reactions during storage, low storage temperatures are essential. This group of adhesives includes cyanoacrylates and UV-sensitive epoxy resins. When stored at 20
C, in dark conditions, they have a minimum shelf life of approximately 2 years. Second, highly reactive single-component adhesives, such as catalyzed epoxy resins in the form of films, are viable at room temperature for only a few days. However, at 20
C, they can be stored for approximately 6 months. This category of adhesives also includes prepregs, which are used in the aviation industry for the construction of aircraft, e.g., to bond the molded parts together. Prepregs are made of woven glass fibers, or more recently carbon fibers, which have been soaked in and pre-reacted with a mixture of epoxy resins and hardeners. They usually undergo temporary storage at 20
C and are conditioned to room temperature before use, where they are assembled with the substrates and cured. Apart from epoxy resins, phenolic resins can also be used as raw materials for prepregs. Table 3 Storage of different types of adhesives Water-based dispersion adhesives Cyanoacrylate adhesives Single-component adhesives polyurethane Methyl methacrylate adhesives Phenolbutyral adhesive films Phenolic resin/rubber adhesive films UV-sensitive epoxy adhesives

5–25
C (Cooling chamber/room temperature) 5–15
C (Cooling chamber/fridge) 15–25
C (Cooling chamber/room temperature) 0–5
C (Cooling chamber/fridge) 5–10
C (Cooling chamber) 5–15
C (Cooling chamber) 5–15
C (Fridge, dark)

36

Storage of Adhesives

36.4

1049

Conclusion

In summary, this chapter has shown that, while the compliance to recommended measures during adhesive storage is absolutely necessary, it is usually manageable. The different types of storage and the corresponding shelf lives for different groups of adhesives and composites have also been addressed. In addition, the need to consider the potential hazards of the adhesives or their components during storage and transport is highlighted. Generally, external influences on the adhesive characteristics during storage, such as temperature, humidity, and other physical effects, are controllable. Where necessary, measures such as suitable packaging and containers, as well as the conditioning of the storage halls, should be implemented. On top of that, adhesive storage should also be taken into consideration in the development of new adhesives. Therefore, new adhesives should not only be tested against their applications but also their storability and shelf lives, before they are launched in the market. While such tests usually constitute the most time-consuming step during the design process, shortening the time taken for this step during the development phase could potentially result in customer claims later on. It is a common practice to subject new products or bound parts to accelerated aging in order to simulate storage over long periods of time. This can be achieved through oven storage or environmental cycling tests for a defined period such as 6 weeks. Such tests are often helpful, except when there are other influencing factors apart from temperature and humidity. In such circumstances, it is difficult to establish the exact storage stability. During storage, any unintended chemical reactions within the adhesives can also lead to problems during or after processing. Hence, new raw materials should be evaluated for potentially undesired chemical reactions to avoid disappointment when such problems arise during the storage of the end products. Such measures also apply when changes to the adhesive contents are required, for instance, when replacements of hazardous raw materials with environmentalfriendly substitutes are called for. In particular, there is an increasing drive toward the use of halogen-free adhesives in the electronic and automotive industries. In general, any change in adhesive formulations would require a series of testing for the storage stability, besides the standard property test profiles. Currently, another challenge is the desire of many industries to have an adhesive with a shorter wetting-out time for the adhesion process and, yet, better adhesion properties. This would require a highly reactive adhesive, which normally requires storage at low temperatures to guarantee a reasonably long shelf life.

References European Standard (2001) DIN EN 12701 Fundamentals of Hazardous Materials (2009/2010) Denios, Germany

Preparation for Bonding

37

Andreas Lutz

Contents 37.1 37.2 37.3

37.4

37.5

37.6

37.7

37.8 37.9

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Training and Education . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Health and Safety Aspects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.3.1 Preventive Measures While Storing the Adhesive . . . . . . . . . . . . . . . . . . . . . . . . . . 37.3.2 Preventive Measures While Working with the Adhesive . . . . . . . . . . . . . . . . . . . 37.3.3 Preventive Measures While Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mixing of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.4.1 Vacuum Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.4.2 Continuous and Batch Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.4.3 Vertical and Horizontal Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.4.4 Kneader . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.4.5 Agitator Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.4.6 Centrifugal Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesive Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.5.1 Storage Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.5.2 Packaging Size . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.5.3 Packaging Material . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Dispensing of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.6.1 Single-Component Dispensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.6.2 Two-Component Dispensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.6.3 Hot Melt Dispensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.6.4 Higher-Volume Dispensing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mixing and Metering of the Adhesives Before the Application . . . . . . . . . . . . . . . . . . . . . . 37.7.1 Low- to Medium-Volume Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.7.2 Medium- to Large-Volume Application . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Transferring the Product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Substrate Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.9.1 Preparation of Steel Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.9.2 Preparation of Aluminum Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1053 1055 1055 1056 1056 1057 1058 1060 1060 1061 1062 1062 1064 1065 1066 1068 1068 1069 1069 1069 1072 1072 1074 1074 1077 1077 1081 1082 1083

A. Lutz (*) Automotive, Dow Europe GmbH, R&D Automotive Systems, Horgen, Switzerland e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_37

1051

1052

A. Lutz

37.9.3 Preparation of Magnesium Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.9.4 Preparation of Low-Energy Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.10 Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37.11 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1084 1086 1087 1088 1089

Abstract

Adhesives are used today with increasing volumes in many applications in different industries and are supplied to the end customers in various forms. Very common are liquid adhesives, pastes, mastics, as well as adhesive films or tapes. Less common are powder systems for coating purposes. Adhesives are used where more conventional mechanical or thermal joining technologies can’t be used because the joint area is not accessible (e.g., for a spot welding or riveting gun), entire areas need to be bonded (tile or carpet bonding), or if improved mechanical joint performance is required. Improved mechanical performance could be increased joint stiffness, longtime durability, or corrosion resistance under environmental exposure. Bonding with adhesives is not a simple operation, and a lot of preparation work is necessary before the adhesive can be applied to the designated bonding area (Shields J, Adhesives handbook. Butterworths, London, 1985). A chain of preparation processes is required. It starts with the mixing of the adhesive and its storage before the transfer to the customer. At the customer, the adhesives need to be transferred to the operation area. The substrates to which the adhesive will be applied have to be well prepared. The final process step includes the dispensing of the adhesive to the substrate and the joining of the parts which are intended to be bonded. Some adhesive formulations require an accurate metering of different components. Some, sensitive adhesive operations require a monitoring of the application, e.g., by vision systems, to guarantee the correctness of the application. Each of the steps is important and contributes equally to the quality of the final adhesive application. See below the summary of the different process preparation steps: 1. 2. 3. 4. 5. 6. 7.

Adhesive mixing at the supplier Packaging of the adhesive at the supplier Adhesive storage at the supplier or the customer Transferring the adhesive from the storage to the application area Substrate preparation Metering and dispensing the adhesive Quality control

Depending for which application, in which industry, and by which end customers adhesives are used, different application techniques and dispensing and metering devices are chosen.

37

Preparation for Bonding

1053

For smaller bonding operations like sealing work at construction sites, repair work at car garages, or to fix tiles to a floor, adhesives are typically manually applied out of smaller containers like cartridges. If adhesives are used in a larger scale, like at an automotive line, for example, for metal bonding reasons or to bond windows to the frame of a car, a robot application is used, which pumps the adhesive out of drums and transfers it to a dozer and a gun for the final application. The adhesive itself can be applied in different ways like as a simple bead, sprayed, extruded, or by air assistance, called swirling. Which technique is used depends on factors like the application, the rheological property of the adhesive, and the application temperature. Therefore, the rheological properties of the adhesive formulation need to be tailored to the application. Lower viscous adhesives can be applied by spraying or extrusion beside the simple bead application. Higher viscous adhesives are often limited to bead application, or the application temperature needs to be significantly increased to be applicable by other methods. This chapter will discuss in further detail each item of the process chain: the storage of adhesives, the transfer to the application area, the different metering and dispensing technologies, the used mixing equipments, the substrate preparation, and the quality control. Proposals are made regarding people education and how to create a safe working environment.

37.1

Introduction

Adhesives are used for many applications in a lot of different industries like in automotive (Symietz and Lutz 2007), marine, packaging, or building and construction (see Part I, Applications). Typical applications examples are: • • • • • • • •

Bonding of car bodies Bonding of structural parts in airplanes (National Research Council 1976) Applying foils on new cars for protecting during transport Lamination of composites like GFC (glass fiber composites) Coating of cans, tubes, or pipes Chemical-resistant flooring Labeling in the food or packaging industry Packaging of food or pharmaceuticals

Adhesives are used to bond various different substrates such as metal, glass, wood, or plastic among many others (Houwink 1967). The appropriate adhesive and its chemistry have to be chosen in dependence of the application and the substrate which is bonded. A wide range of different chemistries are used to formulate adhesives. The chemical nature of the adhesive defines its field of application.

1054

A. Lutz

Most adhesives which are used in the industry today are using epoxy, polyurethane, acrylic, and silicone or rubber chemistry. One- or two-part adhesives are dominantly applied, although there are multiplecomponent adhesive systems in use as well. One-part adhesives are commonly supplied with an inhibited curing system which can be activated, e.g., by heat or by UV radiation. Two-part adhesives are supplied with separate components, one is the resin component and the other the hardener component. The latter contains reactive chemicals, which enable cross-linking and polymer building of the resin part through a chemical reaction of different functional groups. The hardening of the adhesives happens either through a chemical reaction as described above or through simply physically binding to the surface. The chemical reaction in case of a one-part adhesive depends on the chemical nature of the adhesive (Habenicht 1990). So do polyurethane- or silicone-based adhesives cure in the presence of moisture, whereas cyan-acrylic-based adhesives are inhibited by the presence of oxygen and cure under nonaerobic conditions. Other adhesives like epoxy one-part adhesives cure with heat. Two-part adhesives contain, beside the resin part, a hardener part which consists of reactive molecules or polymers which react together with the resin part. Typical examples are epoxy- or acrylic-based adhesive formulations. In the following a summary of properties for some representative adhesive systems is given: • Epoxies adhere very well to a lot of different substrates like steel (Symietz and Lutz 5/2008), aluminum, stone, or wood. They offer a high bonding strength, are resistant to chemicals, and show a good heat resistance. Epoxies are used to fill gaps and larger volumes and are applied in thinner layers. They can be formulated in different viscosities, low or high, and with different filler levels. Epoxies typically cure with heat or can be formulated as two-component products which cure in the presence of a separate hardener component. • Polyurethanes are polymers which are very flexible and tough and which show a very good performance under peel loading and good adhesion to nonpolar substrates like plastic or rubber, as well as to paints and ceramics. They mostly cure with humidity which can cause curing issues in case of extreme humidity conditions. • Acrylics can be cured in different ways like with peroxides or with UV light. The curing is much faster than the curing of epoxy or of polyurethane adhesives. Acrylics are used to fill gaps and to bond plastic or metal parts. They offer a good environmental and heat stability and show a good performance under peel and shear loading. It needs to be considered that acrylics can be incompatible with some metals which are used in dispensing equipments. Acrylics can be formulated to pressure-sensitive adhesives which adhere to surfaces simply by applying some pressure during the application. Adhesives are used in applications where other joining methods cannot be used or because they offer unique performance features like a high strength combined with

37

Preparation for Bonding

1055

high flexibility and superior durability (Lees (1997)). It is therefore very important to prepare the adhesive bonding as accurate as possible to take the best advantage of this bonding technology.

37.2

Training and Education

Many people may think that bonding by using adhesives just requires applying the adhesive to the bonding area and it will stick. But the use of adhesives for bonding purposes requires a well-defined process including steps like: • • • • • •

The correct choice of the adhesive Correct handling of the adhesive Correct mixing and metering Preparation of the surface where the adhesive is applied to Quality checks to guarantee the constant quality of the bond Use of curing sources for one-component adhesives like ovens or radiation sources

All this requires a certain level of know-how, not only of the chemist or engineer but also of the worker who is daily working with the adhesive. In addition, the suppliers of the substrates like the metal or plastic suppliers or the suppliers of the lubricants need a fundamental understanding about the adhesives, the necessary preparation prior to the application, and about the application itself. Training is required for all the people who are involved in the application of the adhesive. The workers who are daily dealing with the adhesive have to be instructed before using the adhesives and regularly trained as well. There are institutes at universities who are offering a certified education in adhesive technology and their application. A good educational background prevents from making mistakes in the preparation of the adhesive and the substrate, helps to choose the right adhesive solution, and helps to apply it correctly. Education is further discussed in ▶ Chap. 41, “Processing Quality Control.”

37.3

Health and Safety Aspects

Adhesives are a blend of different chemicals and therefore potentially hazardous to the environment and the user. The information about the risk can be found on the labels and the material safety data sheets. According to their risk and health, ranking precautions have to be taken to create a safe working environment. Health and safety aspects are further developed in ▶ Chap. 39, “Environment and Safety.” Figure 1 below shows what typically can be found on labels and in material safety data sheets and what will change in the near future.

1056

A. Lutz

Fig. 1 Change from old CHIP (Chemicals Hazard Information and Packaging) to new GHS (Globally Harmonized System of Classification and Labeling of Chemicals) regulation

37.3.1 Preventive Measures While Storing the Adhesive Adhesives are filled in different packaging forms and sizes. Used commonly for smaller volumes are cartridges, bottles, or tubes and for bigger volumes pails or drums. Generally all containers should be stored in a dry and cool place, preferably below 25  C. Some reactive or solvent-containing adhesives would require cooler storage, preferably temperatures below 10  C or even below 0  C. If not, the adhesive viscosity could increase during storage, which could negatively influence its application characteristics, physical properties like adhesion, or even mechanical properties like joint durability. For some adhesive systems, like water-containing products, too low storage temperatures on the other side could lead to crystallization. The adhesive containers ideally should be stored in dependence of their used chemistry. The same chemistry should be stored together; for example, epoxy adhesives should be stored separately from acrylic adhesives. The adhesive containers should be stored above collecting trays to prevent pollution in case of a leakage or spill. Depending on their hazardous potential, adhesives need to be stored in special places: • Flammable material should be stored in rooms which are especially fire resistant and explosion protected. • Adhesives which have a strong odor should be stored in a room with special ventilation.

37.3.2 Preventive Measures While Working with the Adhesive An adequate personal protection should be used if working with adhesives. The safety data sheet has to be consulted to obtain detailed information about the

37

Preparation for Bonding

1057

precautions which have to be taken if working with the specific adhesive. The given shelf lives should be respected or the adhesive supplier consulted if any adhesive is applied beyond the shelf life. The personal protection should at least consist of: • Protecting glasses or goggles • Long-sleeved lab coats • Gloves Gloves should be changed regularly and the used ones should be disposed. The material the gloves are made of should prevent the penetration of the adhesive through the glove to the skin. Latex gloves, for example, are not recommended to be used if working with epoxy adhesives. In this case nitrile gloves should be chosen. The lab coat should be regularly and professionally cleaned. The thickness of the coat should be thick enough to prevent the adhesive from migrating through the woven to the skin. Respiration masks might be used for solvent-containing adhesives, adhesives with strong odor, or if special substrate preparations are needed, which require the use of wet chemistry. Adhesives are chemical blends using different organic polymers. Some examples of special preventive measures, which are recommended to be taken in dependence of the chemistry of the adhesive, are listed below: • Some adhesives, like epoxy-, rubber- or acrylic resin-based formulations, are skin sensitizing and require the use of special gloves. • Acrylic-based adhesives and hardener compounds of two-part formulations can have a strong odor and would require working under proper ventilation. • Solvent-containing adhesives contain flammable solvents and must be kept away from any heating source, would require working under proper ventilation, or for some instances would require the use of respiration masks.

37.3.3 Preventive Measures While Processing Adhesives can be applied automatically by pumping the material out of a pail or drum and transfer it through hoses and possibly a dozer to the nozzle. The adhesive is applied from the nozzle to the substrate. Certain pressures and temperatures are applied to guarantee a good flow of the adhesive form the pail or drum over the dozer to the nozzle. The following precautions are recommended to be taken during processing of the adhesive: • Avoid an overheating of the adhesive which could lead to an increase of the viscosity and eventually can block the hoses or the dozer. • Shield the equipment properly because it is under pressure.

1058

A. Lutz

• Switch of the heat and pressure after longer breaks like overnight or over the weekend. • Use heated follower plates in case the adhesive viscosity is that high that it cannot be applied at ambient temperatures. Do not use heated pail or drum jackets. • Work on specially sealed floors or use collection trays to be prepared in case of a spill. On the other hand, larger amounts of adhesives are commonly manually applied like in case of applying chemical-resistant floor protections or for tile as well as carpet bonding. In such cases, special masks should be worn in addition to the other recommended equipments to protect against the odor of the adhesive or against evaporation of a solvent. After the application of the adhesive, other operations can be additionally applied like spot welding in case of joining doors for cars or grinding of parts after the repair of cars. In such cases, special precautions need to be taken in addition like: • Installing a proper ventilation if thermal joining methods are applied through the adhesive • Wearing dust masks during the grinding operation • Wearing any other protective equipments like long-sleeved lab coats or any eye protection like goggles

37.4

Mixing of Adhesives

Adhesives consist of different raw materials. Organic monomers and polymers usually are the main part of a formulation and mostly contribute to the adhesives’ mechanical and physical performance. In addition to the organic part of a formulation, other raw materials like inorganic mineral fillers or solvents are used. For reactive adhesives, curing agents and accelerators are added to the formulation to cross-link the organic part of the formulation, for example, at ambient or higher temperatures or in the presence of humidity, heat, or UV light. In single-component adhesives, all the raw materials are added in one component. In two-part adhesives, the formulation is separated into two components. One part, usually called part A, contains the organic resins, and the other part, called part B, contains the hardener and accelerator. Adhesives are mixed at the supplier and filled in either smaller container sizes like cartridges or in bigger sizes like pails and drums. For some applications the adhesive bulk and other components like the hardener part or solvents are mixed at the end user just before the application. The correct mixing of the adhesive is a precondition to achieve a good homogeneity of the mixture and to obtain the best mechanical performance of the bond. There are many different kinds of mixers and stirring elements available on the market. Which kind of mixer is adequate depends on chemical nature of the adhesive, the viscosity of the adhesive, and its sensitivity to external factors like

37

Preparation for Bonding

1059

humidity or oxygen. It is proposed to investigate together with the equipment supplier which equipment suits best for the intended application. Mixers which are used in the laboratory or the ones which are used for manufacturing could be very similar in their design. There is not much principal difference with regard to the different elements a mixer consists of. A mixer typically consists of a vessel in which the mixing happens. For laboratory purposes it can be a disposable small container like a simple metal can. The vessel can be surrounded by a heated jacket to allow heating or cooling during the mixing operation. A gear drives the stirring elements. There are many different stirring elements on the market. Which one is chosen depends on many different factors like what kind of adhesive is formulated, how high the viscosity is, or how shear sensitive the formulation is. Ideally, the mixer is connected to a vacuum pump and can be flushed with inert media like nitrogen. In addition, the entire equipment can contain a discharge valve, a vapor filter, and a control panel. A window to watch the mixing is recommended and commonly used. For easy access for maintenance and inspection reasons, the mixer usually is equipped with a lifting column. Figure 2 shows a lab planetary mixer like it is used to formulate smaller quantities of higher viscous adhesives. Fig. 2 Lab mixer with lift and planetary agitators, used to formulate adhesive samples of about 400 g

1060

A. Lutz

Different mixing methods are described in this section. ▶ Chapter 38, “Equipment for Adhesive Bonding,” also describes the mixing of adhesives but focuses on the equipment itself.

37.4.1 Vacuum Mixing Vacuum mixing is defined as such: mixing in a closed system under an internal pressure which is significantly lower than atmospheric pressure. Mixing under vacuum has an important influence on the adhesive’s end performance. It helps to reduce the porosity of a cured adhesive and therefore helps to maximize properties like strength and durability (Angos 2007). But, for some formulations like those which are based on cyan-acrylics, the presence of oxygen can be necessary for the stability of the product. Mixing in many cases happens under vacuum for the following additional reasons: • Deaeration to obtain a void-free product • Obtain an oxygen-free product, to prevent a decomposition in the presence of oxygen in case of sensitive formulation ingredients • To prevent a possible chemical reactions or a microbiological growth • Prevent contact of the mixture with water, in case of ingredients which can react in the presence of water • Dewatering or evaporating solvents of heat-sensitive formulations • To more easily change the physical state from, for example, a paste to a dry powder or from a slurry to a paste • For dosing of light powders or pastes into the formulation • Acceleration of mixing through, e.g., better and faster wetting of powders • Better dispersion of smaller-volume ingredients • A lower-density material is obtained, and therefore smaller volumes need to be handled Because of all the listed benefits it offers, vacuum mixing is mostly used for the mixing of adhesives at the adhesive supplier.

37.4.2 Continuous and Batch Process The mixing equipment can be divided into continuous types and batch types. In a laboratory, batch-type equipments are dominantly used. Although in an adhesive production both types are in operation, the batch-type equipment is more commonly used. A lot of adhesives like epoxy-based adhesives are mixed in a batch-type procedure. Batch systems can be either stationary or transportable, whereas continuous systems are stationary only. The dosing of the raw materials happens sequentially and

37

Preparation for Bonding

1061

contributes significantly to the overall batch time. The raw material can be added either manually or automatically. Automatic dosing is used if the adhesives are produced on a regular base, the same raw materials are used, and the volumes are high. A continuous system is a fully automatic operation. The raw materials are added once and not sequentially like if using the batch process. The most important unit of this system is the metering unit. It must be carefully calibrated and regularly inspected. A batch system is the most precise operation, because each ingredient is added separately and is accurately balanced.

37.4.3 Vertical and Horizontal Mixing In theory the mixing of the adhesive can be done vertically, horizontally or centrifugally. The mixing speed can be varied and adjusted to the product properties. The size of the vessel depends on the volume which will be mixed. It can vary from a few hundred grams for laboratory purposes to about 50–500 kg for upscaling purposes or to about 800 l to about 30,000 l for manufacturing. In vertical mixers the rotor is installed in a 90 angle to the bottom of the vessel, and for horizontal mixers, the rotor is parallel to the ground. The cylindrical vessel form is used for horizontal mixers. In case of vertical mixers, it can have a conical or a cylindrical form. Vertical mixers are very frequently used for laboratory purposes, because they are somehow easier to operate and require less space. For the manufacturing of bigger volumes, the choice of the mixer type depends on items like how the raw materials like to be charged and the final product likes to be discharged, and it depends on the available floor space. Figure 3a shows a planetary mixer head and butterfly agitators,

Fig. 3 Head and agitators of a planetary mixer which are used in production (a) and in a laboratory (b) to mix adhesive formulations. Butterfly agitators are used for the manufacturing mixer and finger blade agitators for the laboratory mixer. The production mixer has a volume capacity of about 1 t and the laboratory mixer of about 10 kg

1062

A. Lutz

and Fig. 3b shows a pilot mixer. Both are used to mix adhesive formulations of higher viscosities.

37.4.4 Kneader For very high viscous materials, kneaders are preferably used. They usually contain a kind of sigma blades, which are able to cut materials like solid rubbers apart and can knead them effectively. An extrusion screw is discharging the material out of the vessel. Figure 4 shows a kneader which is used in a laboratory for mixing of very viscous adhesives like rubber- or polyurethane-based adhesive formulations.

37.4.5 Agitator Design The design of the agitator should be in line with the required mixing function and the expected result of the mixing. For making simple blends, dispersions, or for making adhesives, a different agitator design will be chosen. Adhesive formulations can differ a lot with regard to their viscosity and flow behavior. Depending on the used type of polymers, their molecular weight, and depending on the used fillers, as well as on the ratio of organic polymer to mineral filler, the viscosity can vary from very low to very high. The flow behavior of low viscous materials is totally different from the flow characteristics of higher viscous products (Alhofen 2005). Important factors for the decision of which agitator would suit best are: • • • •

The viscosity of the mixture The shear rate which needs to be applied The nature of the ingredients (e.g., shear sensitive or not) The heat stability of the adhesive formulation

Fig. 4 Laboratory kneader

37

Preparation for Bonding

1063

Fig. 5 Two anchor agitators with different shapes

For very low viscous formulations, a simple anchor stirrer can be chosen. Figure 5 shows two anchor agitators, which only differ in their shape. However, most adhesive formulations require, because of their higher viscosity, agitators which offer a higher shear and flow. The following needs to be considered: • Higher viscous materials do not flow easily and need a special stirring equipment which possibly rotates and travels through each point of the vessel. • The agitator should offer a good heat transfer in the mixture, especially for higher viscous materials, to avoid local overheating. • The agitator gear needs to cope with the high load of a high viscous material. There exist many different agitator designs which are suitable to mix higher viscosities. Some of them are described in more detail below. Stationary high-shear agitators are used for materials with medium to high viscosities and for materials which flow well: • High-speed dispersers are used if medium shear and a higher flow is required. • Rotor-stator mixers can be used alternatively but offer a higher shear and a lower flow. Figure 6 shows a photo of a high-speed disperser. These systems are able to mix even higher viscosities if combined with an anchor blade in addition. Such systems are available as dual- or triple-shaft mixers. The slower-moving anchor removes the material from the vessel wall and the bottom and transports it to the high-shear agitator. Obviously, triple-shaft systems are able to

1064

A. Lutz

Fig. 6 High-speed disperser

handle even higher viscosities than dual-shaft systems. Because of the slow-moving anchor, the system is limited to well-flowing mixtures. In case of planetary agitators, the blades are moving through the material and each agitator can move on its own axis. Such planetary agitators are suitable for higher viscous and not easily flowing materials. Double planetary mixers are typically used for mixing of materials like adhesives, sealants, dental pastes, composites, foams, plastisols, or metal slurries. A scraping blade is commonly used in addition to remove the material from the vessel wall during mixing. Planetary agitators are most efficient at high viscosities because the shear rate in the paste increases proportionally with the increasing material viscosity. Different blades can be used. Typical agitator designs are finger blades, older rectangular-shaped blades, and newer HV blades. The latter are able to even mix formulations of highest viscosities, which usually are mixed by using a kneader or an extruder. Figure 7 shows a photo of finger blade agitators used in a planetary laboratory mixer.

37.4.6 Centrifugal Mixing Centrifugal mixing is used for high-speed mixing and was originally developed for dental materials. Disposal cups are used for the mixing of the adhesives. The cup and

37

Preparation for Bonding

1065

Fig. 7 Finger blade agitators which are used in a planetary laboratory mixer, including a scraper blade

a mixing arm are rotating in opposite directions at high-speed rates of up to 3,000 rounds per minute. Figure 8 shows the mixing chamber and the mixing cup of a speed mixer which is used in the laboratory for the development of adhesives. The biggest advantage over vertical or horizontal mixing is the required mixing time. It varies from a few seconds to a few minutes only. On the other hand, the volume which can be handled is limited which is a clear disadvantage over the other mixing technologies. It varies, depending on the speed mixer model, from a few hundred grams to about 3 kg. Speed mixing is mainly used for lab screening purposes in combination with a design of experiments. Many different cup dimensions are available. Direct mixing in convenient cartridges or syringes is possible.

37.5

Adhesive Storage

The manufactured adhesive can be stored in different packaging forms, like for smaller quantities in cartridges, tubes, or bottles and for bigger scales in pails or drums. Adhesive films are systems which consist of an adhesive dry film applied to a paper, a glass fiber, or an aluminum foil and are usually stored in boxes.

1066

A. Lutz

Fig. 8 Mixing chamber of a speed mixer including a mixing cup

A correct storage is very important to conserve the adhesive properties until the material is supplied to the end user and to avoid any loss of performance, change in rheological properties, or to avoid a possible decomposition of the material through an inappropriate storage. Adhesive storage is treated in more detail in ▶ Chap. 36, “Storage of Adhesives.”

37.5.1 Storage Condition The applied storage conditions depend on the adhesive formulation, specifically the chemical nature of the resins, and in case of reactive adhesives, it additionally depends on which curing and accelerator system is used. Generally, the filled adhesive should be stored in a safe environment, not exposed to heat, to humidity, or, for light-sensitive adhesives, not exposed to light. Good ventilation should be given. The specific storage conditions are commonly written on the labels and can be read in the material data sheets. Two- or multiple-part formulations have a significantly longer shelf life than one-component adhesives with integrated curing agent and accelerator (Lutz and Brändli 3/2010). The shelf life of one-component adhesives strongly depends on the chemical nature of the formulation and on the acceleration system. Multiple-part adhesives have a shelf life of about 6–24 months and single part of about 3–12 months only. The shelf life can be prolonged by a time factor of about 2–4 for one-part adhesives, through storage in a refrigerator below 10  C or by freezing below 0  C. Excluded from cooled storage are products which are known to show phenomena during storage, like phase separation, crystallization, or sedimentation of ingredients of the formulation. Some examples of adhesive formulations which require special storage conditions are given below:

37

Preparation for Bonding

1067

300 Epoxy adhesive A

Viscosity [Pas]

250

Epoxy adhesive B

200 150 100 50

s th

11

on

10

9

8

7

6

5

4

3

2

1

12

m

In

iti

al

0

Fig. 9 Shelf life of two epoxy adhesive formulations. Formulation B is significantly more stable than formulation A

• Reactive one-part adhesives, which can increase in viscosity over time depending on temperature, should be stored under air conditioning or in a cooled warehouse. Examples are accelerated one-part epoxy or rubber-based adhesives. • Some adhesive formulations may tend to crystallize during storage and would require a storage above a certain temperature. • Adhesives which cure with humidity like polyurethane or silicone adhesives should be stored in a dry environment. • UV light-curable epoxy or acrylic adhesives should be stored out of light. • Because of their strong odor, acrylic adhesives should be stored in a room with good ventilation. Figure 9 shows the viscosity stability over time at 23  C storage for two one-part epoxy adhesive formulations. Adhesive B increases only very little in viscosity over time and therefore is significantly more stable than adhesive A. Adhesive A would need to be stored in a cooled area, like in a fridge or an air-conditioned warehouse. Figure 10 shows the decrease of the lap-shear strength of shear specimens which are bonded with one-part epoxy adhesives and stored under 80% relative humidity at 23  C before being cured. Adhesive B shows a much better resistance to water absorption than adhesive A. If using adhesive A for bonding, the joint should be cured quickly, at least before about 1 week after the application of the adhesive. Adhesive B could stay uncured for about 6–8 weeks.

1068

A. Lutz

25 Epoxy adhesive A Lap shear strength [MPa]

Epoxy adhesive B 20

15

10

5

ks

7

w ee 8

6

5

4

3

2

1

In

iti

al

0

Fig. 10 Decrease of lap-shear strength of humidity-exposed lap-shear specimen

37.5.2 Packaging Size The size of the container should be chosen based on the adhesive consumption. If only small volumes are used, small packaging sizes like cartridges, bottles, or tubes should be chosen. Adhesives have a specific shelf life, and the use of bigger packaging sizes for small-volume applications would inevitably lead to disposing significant amounts of adhesive. If large amounts of adhesives are consumed, the application typically is done out of pails or drums using pumps which transfer the adhesive through hoses from the drum to the nozzle where the adhesive is applied to the bonding area.

37.5.3 Packaging Material Depending on the chemical nature of the adhesive or of the hardener compound in case of two-part systems, the material used for the container, for example, cartridges, can be different. Corrosive materials, for example, the hardener compound of a two-part epoxy system, should not be stored in metal containers. Cartridges or bottles made out of plastic are mostly used. Humidity-sensitive adhesives like polyurethane adhesives should be stored in containers which show a low diffusion rate of water. Aluminum cartridges or especially, e.g., metal-coated plastic containers are preferred in this case over simple plastic ones. Special plastic bags which contain, for example, an aluminum liner are recommended to be used for pails or

37

Preparation for Bonding

1069

drums. The plastic inliner prevents on one hand the contamination of the drum or pail with adhesive, which saves cleaning costs, and on the other hand limits the water diffusion into the material. The adhesives should be shipped in a package which is sealed in a way that no material could leak in case of a damage of the container.

37.6

Dispensing of Adhesives

Adhesives can be applied in a lot of different ways to the bonding area. The used methods for dispensing depend on which kind of adhesives is applied, whether oneor two-part adhesives are used, and how much adhesive needs to be applied. In this section, dispensing methods are described.

37.6.1 Single-Component Dispensing For most of the manual operations, the use of smaller packaging sizes like cartridges, bottles, tubes, or even syringes are fully sufficient. For cartridges, there is a broad range of different sizes, and they are made out of different materials like plastic or aluminum. The adhesive is statically applied through a nozzle by using a handgun. Different sizes and shapes of nozzles exist for simple bead application or for the extrusion of the adhesive. Different handguns are available: manual operated ones as well as pneumatically or electrically operated handguns. Which handgun suits depends on how much pressure can be applied for the application. Applying too high pressures could damage the cartridge which could result in a leakage. The viscosity of the material determines the amount of pressure which needs to be applied. High viscous materials can be applied by using a heated gun, or the cartridge can be stored in a cartridge oven before the application. For bottles, there are special bottle dispensers in use like simple hand pumps or volumetric dispenser. Adhesive dispenser controllers are used to offer a high accuracy to meet customer requirements for multivalve dosing or coating or spray operations. Figure 11 below shows typical dispensing equipment for one- and two-part adhesive formulations. It contains a cartridge gun, cartridges, and a nozzle.

37.6.2 Two-Component Dispensing Two-component adhesives are commonly filled in dual cartridges. Depending on the mixing ratio, the sizes for each of the two cartridges are different. Typical mixing ratios are 1:1, 2:1, 4:1, or even 10:1. Special handguns have to be used. There is a new system for two-part applications which is a kind of mono-cartridge including two separate chambers. The benefit is that normal handguns can be used. The nozzle used for applying two-part adhesives is a cylindrical tube, which contains the mixing

1070

A. Lutz

Fig. 11 Cartridge gun, plastic nozzle, aluminum cartridge for dispensing of one-component adhesives and mono-cartridge for dosing two-part adhesives

Fig. 12 Typical cartridge guns and statically mixers for dispensing of two-part adhesives. Bigger (a) and smaller (b) packaging size

elements. The elements are stationary parts that force the materials to combine as they travel through the length of the mixer. The length of the nozzle and amount of mixing elements vary, and the adequate type is typically recommended by the adhesive supplier. Special guns which can dynamically mix the two components in a special mixing chamber are also used, before applying it through the nozzle. Such dynamic mixing is preferably used for high viscous two-part adhesive formulations like some polyurethane adhesives used for car windscreen bonding or for structural bonding in trim shop. Figure 12 shows two typical handguns, a bigger- and a smaller-size one, which are commonly used for applying two-part adhesives. Figure 13 shows two different cartridge systems, which are used to store two-part adhesives. One is a monocartridge with two separate chambers in one cartridge; the other is a double-cartridge

37

Preparation for Bonding

1071

Fig. 13 Mono-cartridge TAH™ system as supplied by Nordson with two chambers (a) and double Mixpac cartridge systems as supplied by Sulzer for applying two-part adhesive formulations: 2:1 mixing ratio, 215 ml and 15 ml volume (b)

Fig. 14 Different kinds of static mixers

system with two separate cartridges for components A and B. Figures 14 and 15 show different kind of static mixers as well as different mixing elements. Figure 16 shows a dynamic mixing element used for mixing of two-part formulations.

1072

A. Lutz

Fig. 15 Different mixing elements, as supplied by Sulzer Mixpac and used for static mixing of two-part adhesive products

Fig. 16 Dynamic mixing element for mixing two-part formulations

37.6.3 Hot Melt Dispensing Hot melt adhesives are very high viscous formulations and can only be applied at high temperatures. Special heated handguns (glue guns) can be used, or the cartridges can be heated in a special cartridge oven to lower the material’s viscosity before the application (Satriana 1974). Figure 17 shows a cartridge oven and a heated cartridge gun.

37.6.4 Higher-Volume Dispensing For higher-volume applications or continuous operations like applications at car manufacturers or in the building and construction industry, the adhesive is applied out of pails or drums. Typical pails contain about 25–50 l of material and drums about 200 l. It is recommended to use plastic inliners to avoid a pollution of the drum or pail. In this case the drum or pail can be reused and just the plastic bag needs to be disposed. Figure 18 shows a typically used pail with a plastic inliner. Some of the inliner may contain an aluminum foil which prevents water penetration through the inliner to the bulk adhesive. Such systems are applied in case of water-sensitive materials like polyurethane-based adhesives.

37

Preparation for Bonding

Fig. 17 Cartridge oven (a) and heated gun (b)

Fig. 18 Pail and inliner for adhesive storage

1073

1074

37.7

A. Lutz

Mixing and Metering of the Adhesives Before the Application

One-part adhesive formulations do not require a special mixing process prior to the application, but two- or multiple-part adhesives need to be mixed before the application of the adhesive. Mistakes in mixing definitely have a negative effect on the quality of the bond. For example, if not applying the correct ratio of A to B component, this can give low initial strength values and bad results with regard to the longtime durability of the bond. There are several mixing and dispensing options, and each offers advantages and disadvantages.

37.7.1 Low- to Medium-Volume Application Most suppliers supply their multiple-component systems in prepacked forms like in cartridges. No metering is necessary in this case. For adequate automatic mixing, there are mixing devices for application out of dual cartridges like handguns or automatic guns and static or dynamic mixers. Static mixers like described above are by far mostly used. The following points are crucial for a good mixing of two-component adhesive systems: • The rheological properties of both components A and B should be adjusted in a way that both components mix well. • The length of the static mixer and the number of mixing elements should be defined depending on the formulation. • Before applying the static mixer, both components A and B need to be equally balanced. • The first few grams of material after mixing should be disposed. • Regularly change the static mixer and dispose the old one. Possible application issues are: • The adhesive can cure at the cartridge nozzle, and this could result in a leak at the cartridge plunger. • The application itself is limited to low viscosities; high viscous materials are difficult to dispense by using handguns. • If using automatic guns in combination with a high viscous material, the high pressure during the application can damage the cartridge plunger, and this can result in a leakage of the adhesive. Figure 19 shows the content of a typical two-part adhesive set which is used for the repair of parts of a car. It shows which different kinds of guns can be used and

37

Preparation for Bonding

1075

Fig. 19 Typical two-part adhesive set for manual applications

Fig. 20 Balancing components A and B before applying the static mixer (a). Dispose the first few grams of material (b)

contain, beside the adhesive cartridge, a bottle of activator and cloths to apply it and to wipe it off the substrate. Figure 20 below shows two important preparation steps before the application of a two-part material. Figure 20a shows the balancing of both components before

1076

A. Lutz

Fig. 21 Application of a two-part epoxy adhesive used for the body repair of cars

Fig. 22 Adhesive bond line after salt spray testing

putting the static mixer on, and Fig. 20b shows the disposal of the first grams of material before the adhesive is applied to the substrate. Figure 21 shows a typical bonding operation performed in garages for the repair of car bodies (Lutz 6/2009). The adhesive, in this case, is manually applied by using a cartridge gun and a static mixer, which are described above. The adhesive used for this application is a structural two-part epoxy adhesive, applied with a mixing ratio of 2:1. Figure 22 shows an adhesively bonded joint after 480 h of salt spray testing. The bonded surface shows no trace of corrosion compared to the surrounding area and

37

Preparation for Bonding

1077

Fig. 23 Spot-weld sealed by using a two-part adhesive

demonstrates the corrosion resistance and sealing potential of the applied adhesive. Figure 23 shows the use of mixed joining technologies, spot welding, and adhesive bonding. In this case, adhesives not only offer strength but also seal the spot-weld against possible galvanic corrosion.

37.7.2 Medium- to Large-Volume Application For medium- to large-adhesive applications, metering and mixing dispense systems are used for applying adhesives like epoxy, polyurethane, polyester, or silicone formulations. Such systems offer an accurate measure, mixing and dispensing for two-part adhesives, and an accurate dispensing for one-part products. There are different models in use like reciprocation or rotary dispensing units. The following metering principles are used: • • • • • •

Positive rod displacement Piston displacement Double-action piston metering Piston cup metering Dual reciprocating piston metering Gear metering

The metering system which is used depends on the kind of adhesive and the application.

37.8

Transferring the Product

The transferring of the adhesives out of drums and pails is commonly done by using a follower plate combined with a pulping or gear pump. The product is further transferred through pipes and possibly a dozer to a nozzle which applies the adhesive to the area to bond. Gear pumps can have certain disadvantages like:

1078

A. Lutz

• Bringing shear stress into the adhesive, which could negatively influence the rheological properties of the material • Because high-shear stress brings heat into an adhesive mixture, the viscosity of a reactive adhesive can increase considerably through a starting polymerization, and, as a result, pipes and dozers can be blocked Depending on the adhesive’s viscosity, the material has to be heated to offer a good material flow. For this purpose the follower plate and each section of the application equipment can be heated. The heat should be switched off in case of longer breaks. Heated drums or pail jackets are not recommended to be used because if the temperature is not carefully controlled, too much heat can be transferred into the material which eventually can lead to a burning of the material as a result of an exothermic reaction. In simple application equipments, the adhesive flows through a pipe to a gun and is manually applied. More sophisticated equipments use a dozer which controls the applied volume and the pressure and applies the adhesive very accurately through a nozzle. If temperature is required for a good material flow, ideally a temperature gradient, with a gradually increasing temperature from the follower plate over the dozer to the nozzle, is applied. The dozer can vary in volume and in pressure which can be used. The amount of necessary pressure depends on the rheological adhesive properties and on the application. Ideally, first in-first out dozer is used to avoid an adhesive plaque building in the dozer through an increase of the adhesive’s viscosity over time, which eventually can block the dozer after a certain time of operation. Figure 24 shows a heated application equipment which is commonly used to apply one-part adhesives, like in the automotive industry for car body bonding. Figure 25 shows the flow of the material from the pumping of the material out of the pail or drum, over the material transfer in the hoses to the dozer and the final

Dosing station

6 bar Electronic switch board

Ext. steering unit 400 V

SCA SCA

SCA SCA

Gun

SCA SCA

CAN-bus

Hydraulic unit

System cabinet Robot control unit

CAN-bus

Heating cabinet

6 bar Double drum pump

Fig. 24 Application equipment as supplied by SCA Schucker and which is used to apply one-part adhesives

37

Preparation for Bonding

1079

Heating hose T = 40–45°C Pump T = 35–40°C

Follower plate T = 30–35°C

Dozer T = 50–55°C T = 45–50°C Nozzle T = 55–65°C Fig. 25 Material flow and applied heat per section Fig. 26 Gun system to apply one-part adhesives

transfer to the nozzle. If heating is required, the temperature is gradually increased as shown. Figure 26 shows a dozer and gun system which is used to apply one-part adhesives. The dozer controls the applied adhesive volume, and the gun finally

1080

A. Lutz

Fig. 27 Hydraulic dosing station

applies the adhesive to the bonding area. Hydraulic, pneumatic, or electronic dosing systems are in use. Hydraulic and electronic dozers are preferred over pneumatic systems because more pressure can be applied and the application can be run faster. Electronic systems have an advantage over hydraulic systems because they do not require space for an aggregate. Figure 27 shows a hydraulic dosing station. In blue, incoming material into the dozer and in red material which is transferred out of the dozer and to the nozzle. The application out of the nozzle to the bonding area can happen in different forms. It depends on the applied pressure and the dimension of the nozzle. Since the application part is discussed in more detail in ▶ Chap. 38, “Equipment for Adhesive Bonding,” only a few automatic application methods are briefly described hereunder. There are some application methods which are frequently used in the industry: bead, swirl, spray, jet-spray application, or extrusion. Which one is used depends on the application and on the rheological properties of the material. Higher viscous materials are not easy to spray but applicable as a bead or by extrusion. Lower viscous materials are applicable faster by using methods like jet spraying or swirling, which are commonly used in the application of structural adhesives to bond car bodies. Table 1 summarizes the applied conditions for each application technology. The adhesive can be applied much faster, if using the air-assisted swirl or the jet-spray method compared to a simple bead application which saves a significant amount of process time. The swirl technique is limited with regard to the applicable volume, which is less compared to the applied volume of other technologies. Because the application is assisted by air, entrapped air can lead to the occurrence of bubbles at the bond surface after the curing of the adhesive. The nozzle diameters and the applied dozer pressures differ a lot in dependence of the used technology. In particular the smaller nozzle diameters which are used for the

37

Preparation for Bonding

1081

Table 1 Application parameters for different application techniques Application type Bead

Nozzle diameter 1–3 mm

Pressure 10–20 bar

Swirl Jet spray

0.8–1.2 mm 0.5–0.8 mm

10–120 bar 100–200 bar

a

Distance to surface About the diameter of the bead: 2–5 mm 20–40 mm 5–20 mm

Application speed 150–300 mm/s 200–600 mm/s 300–600 mm/s

b

c Application with swirl nozzle

Fig. 28 Different application technologies: bead application (a), jet-spray application (b), and swirl application (c)

jet-spray application can cause issues with blocked nozzles, if the fillers which are used for the adhesive formulation are not below a certain particle diameter or if they tend to form agglomerates. Figure 28 shows the principles of the discussed application technologies.

37.9

Substrate Preparation

The preparation of the substrate before the application of the adhesive is very important for the quality of the bond and the performance of the bond during operation (Metals Handbook 1982). A good preparation is a guaranty for a maximum bonding strength, best fatigue durability, and best resistance to environmental factors like the exposure of the bond to humidity, heat, UV light, and salt. The required surface preparation depends on which substrate will be bonded and which kind of adhesive is applied. Some metals, for example, have oxide layers like aluminum, magnesium, or titanium. In this case, the surface preparation focuses on stabilizing the oxide layer. Depending on the chemistry of the adhesive, the adhesion to different substrates and the tolerance for surface contamination are different. Epoxy- and acrylic-based adhesives bond well to a variety of different surfaces like metals, aluminum, or composites and tolerate some degree of surface contamination like oil if cured by

1082

A. Lutz

heat. They are not adhering well to low-energy surfaces like plastic. Two-part epoxy formulations need a cleaning of the surface to remove any contamination. Special two-part acrylic-epoxy hybrid adhesives are able to bond even on oily polar surfaces like metals. Newly developed purely epoxy-based two-component adhesives have been developed which are capable to absorb surface oils. But there is a time limitation after which the adhesive needs to be subjected to heat to facilitate the oil absorption into the adhesive. The application of heat at temperatures typically above 100  C and below 140  C should be done after several hours. Some uniquely catalyzed acrylic two-part formulations bond well to nonpolar surfaces like plastic. Polyurethane-based adhesives adhere well to coated or painted surfaces like on ceramic-coated glass, electro-coated or painted metal, or to plasma-treated low-energy surfaces. Usually a primer or an activator is applied to the surface before adding the adhesive. This section gives a brief overview of the substrate preparation and is treated in more detail in ▶ Chap. 8, “Surface Treatments of Selected Materials.”

37.9.1 Preparation of Steel Substrates Beside stainless steel, most adhesives, especially epoxy- and acrylic-based ones, bond well to steel surfaces. The steel substrates can differ in strength, surface topography, and their composition. The preparation for bonding is different whether one-part heat-curing adhesives or two-part formulations are used. In the case of one-part epoxy adhesives, the surface just needs to be cleaned from any contamination like dust. Oil is tolerated in an amount lower than about 3 g/m2. Higher viscous oils like dry lubes are tolerated up to about 2 g/m2. In case of applying two-part formulations, with the exception of acrylic-epoxy hybrids, oil and any other contaminations need to be carefully removed. For a better long-term durability of the bond, for example, a better resistance against galvanic corrosion, a coating of the steel surface is recommended. This coating is usually a zinc coating (galvanized steel), either hot-dipped zinc or electro-galvanized zinc coating, and is used in many industries like in the automotive industry. The quality of the zinc layer has to be high to offer a good durability and corrosion resistance of the adhesive bond. The achievable mechanical strength values are high and often even higher than the strength values for the non-coated steel. An insufficient quality of the zinc layer often results in a delaminating of the zinc layer under loading of the bond. The adhesion, especially of the e-coat, the filler, and the paint, can be further optimized by applying crystals, like zinc and iron phosphates, to the metal surface. This treatment, called phosphating, is often used, for example, in the automotive industry. In addition to the above-described metal coating, organic methods can be used like epoxy coating, e-coating, or the application of paint. In this case, the achievable bond strength values strongly relate to the quality of the coating and are generally

37

Preparation for Bonding

1083

lower than the strength values for the non-coated steel. Not all adhesives adhere well to painted surfaces, and their adhesion has to be tested in advance (Dettner 1991).

37.9.2 Preparation of Aluminum Substrates Aluminum substrates are available in different alloys and surface topographies as sheet panels or casting parts. To prevent the aluminum oxide layer from undefined growth, a special surface treatment of the aluminum is necessary. The cleaning of the surface can be done thermally, chemically, or electrochemically. It is necessary for the following reasons: • • • •

Remove oil and surface contamination. Remove aluminum fines. Remove surface enrichments of trace elements. Remove the not perfect natural oxide layer.

Thermal cleaning is done by coil annealing in a furnace. The chemical cleaning can be done by immersion of the coil into an acidic or alkaline solution. Alternatively, the solution can be sprayed on the coil. The electrochemical cleaning is done in sulfuric or phosphoric solutions. In addition to the cleaning of the surface, a chemical conversion treatment or an anodizing process can be applied which offers more homogenous surface properties. The oxide layer is stabilized, and the adhesion of the adhesive, the primer, or the paint is improved. Anodizing is commonly used to prepare aluminum for bonding in the aerospace industry. The chemical conversion of the aluminum surface happens in the presence of titanium and fluoride ions in a sulfuric solution. Some treatments use zirconium ions together with titanium ions. The sulfuric acid and the fluoride ions are pickling the surface, and the metal salts are responsible for the passivation of the surface. The titanium and zirconium ions react with the aluminum oxide on the surface and form a mixed oxide layer. Like the cleaning process, it can be applied by an immersion or by a spray process. Figure 29 shows the simplified chemical formula for the mixed oxide layer. Figure 30 shows a microscope picture of a cut image of an aluminum sample which has been prepared using titanium and zirconium surface treatment.

Zr H

O

O Al

O Al

Ti O

O Al

O Al

Fig. 29 Simplified chemical formula of the mixed oxide layer

O

O Al

O Al

H

1084

A. Lutz

Fig. 30 Cut image of an aluminum sample passivated by using titanium and zirconium surface treatment

37.9.3 Preparation of Magnesium Substrates Chromate conversion coatings are effective treatments for magnesium and are applied before the adhesive bonding or the paint, but the chemistry used is harmful and restricted by regulations. The alternatives to the chromate treatment and the treatment process for magnesium are discussed below in further detail (Skar and Albright (2002)). Generally the pretreatment of magnesium includes the following steps: 1. 2. 3. 4. 5. 6. 7. 8. 9. 10.

Treatment of the magnesium cast: e.g., machining, deburring, or grinding. Rinse with water (omitted in case of a no-rinse process). Cleaning with alkaline solution. Rinse with water. Optional: activation by acidic pickling. Applying the conversion coating. Rinse with water (omitted in case of a no-rinse process). Rinse with deionized water. Drying at 90–105  C. Applying the adhesive or the paint.

After the machining and the deburring, the die cast part is cleaned by using an alkaline solution. In some cases there can be an etching or pickling step applied before the conversion treatment. The conversion coating can be applied like for aluminum by immersion into a bath or by spraying. The cleaning of magnesium casts happens in an alkaline solution. In the case of magnesium, there is an optimum pH range of about 10–11 in which the etching rate shows the highest value. Below or above that range, the etching is less effective because of a lower etching rate. For aluminum, the etching rate increases with an increasing pH.

37

Preparation for Bonding

1085

The acidic pickling can be done by using acidic solutions of different acids like sulfuric acid, phosphoric acid, hydrofluoric acid, nitric acid, or chromic acid. The etching depth can vary from a few micrometers only to about 50 μm. The etching rate depends on which acid is used; it is high for sulfuric acid and low for nitric acid. In the case of chromate treatment, the magnesium oxide on the surface reacts together with chromate ions to form a chemical complex and establishes a defined surface film of magnesium and chromium chromate. As alternative to a chromate treatment (Gehmecker 1994), other chemical coatings can be chosen like: • Similar to aluminum treatment: titanium and zirconium fluorides • Phosphate or phosphate-permanganate solutions • Solutions containing vanadate, manganate, and molybdate The chromate treatment alternatives offer a similar good adhesion of the adhesive if, in addition, a cleaning and pickling process is used. If pickling is not done, the chromate treatment offers the best adhesion performance. A proposed combination is using a cleaner with a pH of about 9–11 and an acidic pickling. The alternative treatments are missing the self-healing and inhibiting effect of the chromate layer, which offers an additional corrosion protection. An exception is the mixed metal treatment. The performance of the different treatments was tested by performing accelerated corrosion tests. The loss in bonding strength was by far lower than the loss of strength of not treated lap-shear specimen. Table 2 lists some commercially available conversion treatments for magnesium. Beside the chemical treatment, anodizing is an effective treatment method too. The magnesium part is the anode in an electrochemical cell and dissolves into magnesium ions. Together with electrolyte, a defined oxide and hydroxide layer is Table 2 Commercially available chromate-free conversion treatments Supplier AHC Oberflächentechnik GmBH Chemetall GmbH

Henkel

Treatment MagPass™

Type Vanadate, molybdate, and manganate solution

Gardobond (TM)X4707 Gardobond (TM)X4740 Alodine(TM) 5200 Alodine(TM) 4850 Alodine(TM) 400 Alodine(TM) 160

Titanium-zirconium hexafluorides Zirconium hexafluorides Titanium and/or zirconium oxyfluoro compounds with polymer additions Same as above Same as above Zirconium fluorides and magnesium oxides

1086

A. Lutz

formed on the magnesium surface. If phosphates, fluorides, silicates, or chromates are present in the electrolyte, they will be incorporated into the surface layer.

37.9.4 Preparation of Low-Energy Substrates Plastics can be either thermoset or thermoplastic materials. Examples for thermoset materials are polymerized epoxy, polyester, or phenolic resins, which do not melt after being cross-linked. Thermoplastics are materials like polyolefins, for example, polyethylene or polypropylene, polycarbonates, polyamides (e.g., nylon), or acrylonitrile butadiene styrene (ABS). Thermoplastic materials are increasingly used in industries like in automotive for reasons like weight saving. Thermoplastic materials like polyethylene or polypropylene exhibit low surface energies and are by far less polar than other substrates like steel or aluminum. Polar functions like oxides or hydroxides are missing at the surface. For good adhesion of low-energy surfaces, a good surface preparation is necessary to offer a good initial and also durable adhesion (Korane 2000). For the following reasons, a surface preparation is strongly recommended: • Removal of possible contaminations like release agents or oils • Increase the surface energy of the substrates, by adding more polarity to the surface • Adding texturing to the surface, to increase the available bonding surface The preparation of plastic parts can be done by different methods (Bienak 2009). Those methods are summarized below in more detail. Cleaning of the surface can be done by using solvents or special pastes which remove possible contaminations like oils or release agents. The surface can be abraded to remove contaminations but also to offer a textured surface, which increases the bonding area. Different abrasion methods can be used. A flame treatment oxidizes the surface and creates polar oxide groups on the surface. The surface energy is increased and the adhesion improved. Figure 31 shows the plasma treatment of a polypropylene surface. Plasma is generally generated by adding energy into a gas, so that it becomes ionized. It can be used for the cleaning of the surface but also can create polar groups on plastic substrates similar to flame treatment. Low-pressure or atmospheric pressure plasma is generated by applying a high voltage. The plasma contains free electrons, which are accelerated and create ions from the gas. In contact with any kind of surface, the energy of the plasma is transferred and causes reactions on the surface in a way that radicals are formed on the substrate surface. Those radicals can react with oxygen or different gases, like nitrogen or argon, and form functional groups on the surface. Polar functional groups like carbonyl, carboxyl, or hydroxyl are formed. The surface can be additionally activated through the cleavage of carbon bonds by UV radiation.

37

Preparation for Bonding

1087

Fig. 31 Atmospheric plasma treatment of a polypropylene surface (Source Plasmatreat GmbH, Germany)

Corona treatment uses a high-frequency discharge at about 15–25 kHz. The energy transfer to the substrate is high in this particular case due to oscillating electrons between the electrodes.

37.10 Quality Control The quality of the bond needs to be audited in order to guarantee its performance. This is very important because adhesives are applied today in critical areas where a long lifetime of the bond is required. The application itself can be controlled by accurate metering and dispensing and by using dozing systems. The right equipment makes certain that the correct amount and mixing ratio are used and that the adhesive is applied exactly there, where it should be applied. In case of any incorrect dosing, the system will stop the application and will give a warning. The application can be controlled further by using vision systems, which are monitoring the application and control the positioning of the adhesive and any inconsistency within the adhesive bead. Such systems are, for example, in use in the automotive industry to control sensitive adhesive applications in the body shop. Figure 32 shows a vision system which is installed directly at the application nozzle and controls the bead structure as well as the positioning of the bead on the metal sheet. Any failure, like bead disruptions, will be corrected in a second application step. Similar systems are used, for example, in the packaging industry to control the application of the glue to corrugated paper. Another, but deconstructive, method is the control of the bond by dismantling the part and by visible inspection of the bond. This part’s audit can be done regularly like once a day, a week, or a month depending on the importance of the application. Car bodies or parts of a body, which are using adhesive bonding in addition to older methods like spot welding, are regularly put apart for inspection purposes.

1088

A. Lutz

Fig. 32 Vision systems as supplied by QUISS GmbH to the automotive industry

Quality control is discussed in more detail in ▶ Chap. 41, “Processing Quality Control.”

37.11 Conclusion Adhesives are widely used in many industries for various different applications like in the automotive industry for bonding entire car bodies, or in the packaging industry for attaching labels, or for various applications in households. Many more examples could be listed. The adhesive bonding technology offers a lot of advantages over other more conventional joining methods. Because one finds so many examples where the performance of the final product relies much on the performance of the adhesive bond, like in glued car bodies, the process of preparation before the application of the adhesives contributes a lot to the final quality of the bond. Failure in mixing or in metering, as well as in dispensing, can result in a disastrous performance of the adhesive bond during usage. Each step of the described process chain needs a careful preparation and execution. The people who are working either in the developing or in the manufacturing stage of the adhesive bonding process or are applying the adhesives should be trained how to correctly use the material and how to prepare the substrates to achieve the best initial and long-term bonding performance. The precautions which should be taken to work safely should be well understood.

37

Preparation for Bonding

1089

References Alhofen B (2005) Get into the thick things. www.chemicalprocessing.com Angos C (2007) Vacuum mixing. www.processingmagazine.com Bienak D (2009) Preparing polymeric substrates for bonding. www.reinforcedplastics.com Dettner HW (1991) Pocket dictionary for metal surface treatment. Leuze Verlag, D-7968 Saulgau Gehmecker H (1994) Chrome free pre-treatment of magnesium parts. In: 51st world conference on magnesium Habenicht G (1990) Kleben. Springer-Verlag Berlin Heidelberg Houwink R (1967) Adhesion and Adhesives, Elsevier Science & Technology, Oxford, United Kingdom Korane KJ (2000) Guidelines for bonding plastics, machine design. http://www.machinedesign. com/adhesives/guidelines-bonding-plastics Lees WA (1997) Adhesives in engineering design. Springer GmbH, Berlin Lutz A (2009) Crash-resistant two-component adhesives for repairs, adhesion adhesives & sealants. Adhäsion Kleben & Dichten, Springer Fachmedien Wiesbaden Lutz A, Brändli C (2010) Optimal haftend, temperaturstabil und länger lagerfähig, adhesion adhesives & sealants. Adhäsion Kleben & Dichten, Springer Fachmedien Wiesbaden Metals handbook (1982) 9th edn, vol 5, Surface cleaning, finishing and coating, ASM National Research Council (1976) Structural adhesives with emphasis on aerospace applications. Dekker. Adhesion and Adhesives, Elsevier Science & Technology, Oxford, United Kingdom Satriana J (1974) Hot melt adhesives: manufacture and applications. Noyes Data Corporation. Adhesion and Adhesives, Elsevier Science & Technology, Oxford, United Kingdom Skar JI Albright D (2002) Emerging trends in corrosion protection of magnesium die castings. Magnesium Technology., Springer, Cham pp. 585–591 Symietz D, Lutz A (2007) Structural bonding in automotive manufacturing. Verlag moderne Industrie Verlag Moderne Industrie, Die Bibliothek der Technik, D-80992, Munich, Germany Symietz D, Lutz A (5/2008) Strukturkleben von hochfestem Stahl, leightweightdesign. Vieweg +Teubner Verlag, Germany

Equipment for Adhesive Bonding

38

Manfred Peschka

Contents 38.1 38.2 38.3 38.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manual Processing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General Design of a Metering and Mixing Machine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Components Required for the Machine-Processing of Adhesives . . . . . . . . . . . . . . . . . . . . . 38.4.1 Drum Presses (RAM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.4.2 Pressure Containers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.4.3 Pumps . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.4.4 Valves . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.4.5 Mixing Devices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.5 Computational Fluid Dynamics (CFD) as a Tool for the Further Development of Metering Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.6 Automation and Robotics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.6.1 Prerequisites for Automating Adhesive Application . . . . . . . . . . . . . . . . . . . . . . . . . 38.6.2 Parameters Affecting Accuracy and Methods of Compensation . . . . . . . . . . . . . 38.6.3 Dosing Accuracy (Dynamic Behavior of a Two-Component Metering and Mixing Machine) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.7 Interface Between the Robot Control and Metering Control . . . . . . . . . . . . . . . . . . . . . . . . . . 38.7.1 Accelerated Curing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.7.2 UV-Radiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.7.3 Induction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1092 1092 1093 1094 1095 1095 1095 1099 1102 1104 1106 1106 1106 1108 1113 1113 1113 1114 1115 1116

Abstract

This chapter describes the equipment required for manual adhesive bonding processes. For some readers, the comments about equipment may seem trivial,

M. Peschka (*) Fraunhofer Institute, Bremen, Germany e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_38

1091

1092

M. Peschka

but they are important as nonobservance of these points can in practice lead to flawed bonded joints. The individual components which make up mixing and dosing systems have been described and their special features outlined in an attempt to quantify the hitherto qualitative descriptions that have prevailed in the literature. Thereafter, some basics of automation and robotics are outlined with a special excursion on the topic parameters affecting accuracy. Terminal opportunities of accelerated curing are explained using the examples of UV radiation and inductive heating.

38.1

Introduction

The equipment required for manual adhesive bonding processes will first of all be described. For some readers the comments about equipment may seem trivial, but they are important. This is because nonobservance of these points can in practice lead to flawed bonded joints. The individual components which make up mixing and dosing systems will then be described and their special features outlined. This is an attempt to quantify the hitherto qualitative descriptions that have prevailed in the literature. Thereafter some basics of automation and robotics are outlined with a special excursion on the topic parameters affecting accuracy. Finally the opportunities of accelerated curing are explained using the examples of UV radiation and inductive heating.

38.2

Manual Processing

The properties of a bonded joint are created during a production process. Chemical reactions convert the starting materials (monomers, prepolymers) into a cured adhesive film. The adhesion to the substrates and the cohesion can, however, easily be inadvertently influenced. This is the reason why a bonding process must be strictly carried out in accordance with prespecified conditions. In practice, the individual workers are a significant variable, and this is a further reason why industrial bonding processes are being increasingly automated. The manual processing of adhesive systems therefore puts very high requirements on the persons undertaking the work and on their understanding of adhesive bonding processes. Not least for this reason, a group of experts under leadership of Fraunhofer IFAM developed a series of training courses for training personnel in all aspects of adhesive bonding. These courses have been accredited Europewide and successfully running for more than 10 years. Three course levels are offered, reflecting the differing responsibilities of the relevant personnel:

38

Equipment for Adhesive Bonding

1093

• European Adhesive Bonder (course duration: 1 week) • European Adhesive Specialist (course duration: 3
1 week) • European Adhesive Engineer (course duration: 8
1 week) The courses conclude with an examination before an independent examination committee. Successful completion of a course ensures the participants are able to effectively apply adhesive bonding technology in their work area. All persons who are involved in adhesive bonding technology are highly recommended to take the relevant training courses. In order to be able to keep the workplace clean, it is sensible to work on a base mat made of strong paper or foil. This can be disposed of after completing the work, meaning that a clean workplace is once again available for the next bonding task. Before starting the actual bonding work, all the required auxiliary processing materials including cleaning agents should be gathered together. In order to weigh out the components of a two-component adhesive, a suitable balance must be used. The resolution of the balance must be an order of magnitude higher than the required accuracy of the weighing. It must be ensured that the balance is horizontally positioned. Beakers that are used for mixing components must have smooth edges and be resistant to the adhesives and cleaning agents. Experience has shown that Polypropylene beakers are ideal. Spatulas made of high-grade steel are particularly good for mixing. When mixing, ensure that material on the edge and base of the beaker is constantly included in the mixing process. A mixture is homogenous when it has a uniform color and there are no streaks. Cartridge guns are suitable for applying one-component adhesives. These guns can be mechanically or pnematically operated. Similar cartridge guns are available for processing 2-C adhesives which are packaged in special twin cartridges. In combination with static mixers, this means that possible faults due to an incorrect mixing ratio and inhomogeneous mixing are avoided. The following equipment is required to monitor the ambient conditions and time: Thermometer Hygrometer Laboratory clock

38.3

General Design of a Metering and Mixing Machine

The machine-processing of adhesives involves a number of different tasks: – Maintaining a stock of adhesive (or the components). – Removal of the adhesive (or components) from a storage container. – Transporting the adhesive (or the components).

1094

M. Peschka

Storage containers • Pressure containers • Supplied containers

M

Dosing pump • Piston • Gear • Eccentric screw • Diaphragm

Pump drive • Electric • Pneumatic • Hydraulic

Valve • Needle • Snuff-back • Pinch

M

Mixing system • Static • Dynamic

Fig. 1 2-C dosing system (schematic representation)

– Dosing the adhesive (or the components). – Mixing the adhesive – if the adhesive consists of several components. – Applying the adhesive. The word “dosing” here has different meanings. Sometimes it refers to the measuring out of quantities of the two components to get the correct mixing ratio, other times it refers to the measuring out of the total quantity of adhesive to be applied. To avoid ambiguity, a proposal is made here: Dosing will be used to refer to the measuring out of the components (2-C) prior to mixing. Applying is used to refer to the application of the correct amount of adhesive to the substrate. The simplest conceivable machine dosing system consists of a storage container under a defined pressure. When an opening is released, the adhesive flows, and the flow rate depends on the viscosity of the adhesive and the geometry of the opening (diameter and length). This type of application system is often used for microproduction, whereby defined pressure impulses are passed to a cartridge using a pressure/time controlled unit. Additional components such as pumps and valves are necessary when applying larger volumes of adhesive. The following figure shows the schematic representation of a dosing and mixing system (Fig. 1).

38.4

Components Required for the Machine-Processing of Adhesives

This section attempts to specify the features of the various machine components. In some cases, however, the indicated ranges are exceeded.

38

Equipment for Adhesive Bonding

Fig. 2 Fundamental structure of a RAM

1095

Traverse

Pneumatic drive

Pneumatic ram

Pump

Follower plate Drum Medium to be pumped

38.4.1 Drum Presses (RAM) The adhesive is compressed using a drum follower plate. Via a telescope frame, pressure can be exerted on the surface of the adhesive by the follower plate, so that the adhesive flows into the center to the pump. The plate is pressed against the surface of the adhesive. In the car manufacturing industry, and indeed anywhere where uninterrupted production must be guaranteed, the feed containers are present in duplicate. When a storage container becomes empty, the second container is automatically brought into operation. Whilst the production continues, the empty container can be changed (Fig. 2).

38.4.2 Pressure Containers These can also act as intermediate containers to allow uninterrupted operation of the plant. If equipped with a riser pressure, containers cover the original barrel ensuring a contamination-free operation. They are not suitable for all viscosities.

38.4.3 Pumps Pumps are required for transporting the adhesives or their components over large distances (= feed pumps) and for exact volumetric dosing (= dosing pumps). In this section, the key features of different types of pumps will be outlined.

1096

M. Peschka

In

Out

• There must always be pressure on the inlet side(!), not self-priming • Virtually pulsation-free adhesive metering • Permanent metering • Use of filler-containing adhesives possible (Mohs’ hardness < 5!) • Operating pressure up to max. 100 bar • Metering output proportional to the drive speed • Heatable systems available • Viscosities: low(≈101 mPas) to high (≈105 mPas) • Slip increases with duration of use, not detectable from outside

Fig. 3 Gear pump – design, function, and characteristics

Gear Pumps This type of pump is also used in many other technical areas. It transports a fluid by creating defined partial volumes between the sled wall and one tooth respectively of the two gear wheels. The rotation of the gears carries these partial volumes from the intake side to the discharge side. Gear pumps are used as both feed pumps and dosing pumps. They are usually driven by an electric motor. As the speed of an electric drive can be very easily controlled, very precise dosing is possible using gear pumps. The features are summarized in Fig. 3. Double Acting Piston Pumps These pumps are often used for processing highly viscous adhesives. The design of a twin piston pump is schematically shown in Fig. 4. The special feature of a twin piston pump is that it transports materials in both movement directions. This is achieved due to the fact that the two chambers, which are separated by the piston, are kept in phase with each other by a nonreturn valve, and the cross-section of the lower chamber is twice as large as the cross-section of the upper chamber. When the pump is at the bottom dead center, both nonreturn valves close (= piston valve and foot valve). If the piston now moves upwards, the foot valve opens and the lower chamber is filled. Simultaneously, the adhesive in the upper chamber is transported via the outlet into the connected pipe system. After reaching the top dead center and reversal of movement of the piston, the piston valve opens and the foot valve closes.

38

Equipment for Adhesive Bonding

Out

1097

• Self-priming • Slight pressure lost on movement reversal

Pump rod Non-return valve (piston) here: open Piston Non-return valve (inlet) here: closed

(ca. 5–10%) • Permanent metering • Abrasive filler-containing adhesives can be processed (Mohs’ hardness >5) • Operating pressure up to 400 bar • Metering proportional to the speed of the piston • Heatable systems available • Viscosity range: ca. 101–106 mPas • Wear cannot be detected from outside in

Non-return valve (piston) here: closed

every case

Non-return valve (inlet) here: open In

Fig. 4 Double acting piston pump – design, function, and characteristics

The upwards movement of the piston forces the volume of adhesive from the lower chamber into the upper chamber. As the cross-section of the upper chamber is only half that of the lower chamber, not only the upper chamber is filled but the excess adhesive is transported via the outlet into the pipe system. At the moment of reversal of movement at the dead centers of the piston pump, there are short periods of lower pressure at the outlet, which is due to the switching time of the nonreturn valves. These pumps are used as both feed pumps and dosing pumps. The drive system in most cases is pneumatic. Hydraulic drive systems are used when there are very demanding requirements. When used as a dosing pump, it must be remembered that the fleeting low pressure at the points of reversal is accompanied by a decrease in the applied quantity. The accuracy of gear pumps cannot be achieved.

Single Acting Piston Pumps These pumps are used as very accurate dosing pumps for high viscosity and/or abrasive adhesives (adhesives with abrasive fillers). The design and operation are explained in Fig. 5. When the inlet valve opens, the pump is filled due to the pressure to which the material is exposed. On reaching the top dead center, the inlet valve closes and the piston is moved until the desired dosing pressure is reached. The opening of the outlet valve starts the application process, which must be completed before the bottom dead center is reached. Before starting the next application process, the pump has to be filled again. The drive is pneumatic when the accuracy

1098

M. Peschka

Plunger

• Self-priming • Pulsation-free metering • Limitated metering volume

Inlet valve

• Abrasive filler-containing adhesives can be processed (Mohs’ hardness >5) • Operating pressure up to 400 bar • Metering proportional to the speed of the plunger • Heatable systems available • Viscosity range: ca. 101–106 mPas • Wear can be detected from outside

Outlet valve

Fig. 5 Single acting piston pump – design, function, and characteristics

requirements are low. For higher accuracy applications, hydraulic or electric drives are used. Since about 2005, electric drive systems, with a three-phase servo motor and a recirculating ball screw, have rapidly driven hydraulic drives out of the marketplace. Their advantages are their cleanliness (no droplets of hydraulic oil!) and favorable cost. Electric drives are available as standard modules. Single action piston pumps are the standard for chassis production (body in white).

Eccentric Screw Pumps These pumps are widely used for dispensing a wide variety of chemical and food media. They have been used for applying adhesives since the start of the third century AD. The design and operation are explained in Fig. 6. The advantage of these pumps is that they can be operated without valves. Defined partial volumes form between the rotor and stator, and these are transported along the axis of the rotor by rotation. At the end of the adhesive application, a short reversal of the rotor promotes breaking of the adhesive thread. These pumps are driven by electric motors. When operating these pumps, special attention must be paid to a number of points: – If possible, the first filling of the pump should be undertaken manually. Dry running of the rotor destroys the stator. – When cleaning the pump, ensure that the solvent being used for the cleaning is compatible with the material from which the stator is made. Depending on the application, the manufacturers provide stators made of different elastomers.

The simple mechanical design and the relatively small pressure loss when pumping highly viscous adhesives mean that these pumps are particularly attractive for highly integrated production lines.

38

Equipment for Adhesive Bonding Not shown: Electric drive

1099

• Self priming • Pulsation-free metering

In

• Permanent metering • Abrasive filler-containing adhesives can be processed (Mohs’ hardness >5) • Operating pressure up to 100 bar • Metering proportional to the drive speed • Viscosity range: ca. 101–106 mPas

Rotor

• Wear can not be detected from outside

Stator

Out Figure courtesy of viscotec

Fig. 6 Eccentric screw pump – design, function, and characteristics

Other Types of Pumps In addition to the pump types which have been mentioned, other pumps are also used for transporting and dosing adhesives for special areas of application. Only brief mention of these pumps is therefore appropriate here. Diaphragm pumps are mainly used as feed pumps for low-viscosity and mediumviscosity adhesives. In these pumps, the pump chamber is defined by a diaphragm. The fluid is transported via two nonreturn valves, which act as input and output valves, and the oscillating motion of the diaphragm. Peristaltic pumps are usually used for process engineering applications. These pumps are occasionally used in cases where small amounts of low-viscosity adhesives have to be applied. The flexible tube is pinched by cams or rollers and the adhesive therein becomes trapped (“occluded”). Due to the rotation of the cams or rollers, the adhesive is transported from the inlet to the outlet of the pump. The benefit of peristaltic pumps is that the actual pump mechanism does not come into contact with the medium being pumped.

38.4.4 Valves Valves are used to control the flow of an adhesive in a pipe system. They are present as inlet and outlet valves on pumps. Application valves are often the last component of the adhesive delivery section.

Needle valves This type of valve is by far the most common type of valve in use. A circular valve seat is closed by a suitable valve needle (Fig. 7). By movement of the valve needle in

1100

M. Peschka

Pneumatic drive

On

• Abrasive-filled adhesives can be processed (Mohs’ hardness >5)

In

• Operating pressure up to 400 bar • Viscosity range: ca. 100–106 mPas

Off

• Closing motion displaces an additional amount of adhesive; can be minimized with stroke limiter • Heatable systems available

Out

Fig. 7 Needle valve – design, function, and characteristics

the direction of the longitudinal axis, the cross-section of the valve seat is opened and the adhesive is able to flow. The drive for the opening and closing movement can be pneumatic, pneumatic/spring force, hydraulic, or electromagnetic. Depending on the nature of the adhesive being processed (viscosity, fillers) and the boundary conditions for the process (pressure, temperature), there is a wide choice of cross-sectional area and materials (elastomers or hard metal) for the valve seat.

Snuff-Back Valve The kinematics of the opening and closing motion of a snuff-back valve is precisely the reverse of that of a needle seat valve (Fig. 8). To open the valve the piston is moved forward and this opens the valve cross-section. To close the valve, the piston is retracted until the free cross-section of the valve seat is closed. This generates a slight vacuum in the outlet needle, allowing effective thread break of the adhesive at the end of the application. The sealing of the valve on the outer surface of a cylinder is more technically demanding than sealing a needle seat. The pressure region in which absolute sealing can be guaranteed is therefore limited. In industrial practice this type of valve is only used in a few very special applications. Pinch Valve Pinch valves are used for low-viscosity adhesives which are processed in the pressure range up to 10 bar (Fig. 9). The tube cross-section is pinched in such a way by a pneumatically driven lever that the flow of adhesive is stopped. The

38

Equipment for Adhesive Bonding

1101

Pneumatic drive Off • Filler-containing adhesives can be processed (Mohs’ hardness “static” mixer • Mixing energy is removed from the flow • Alternating clockwise and anticlockwise mixing elements split the flow of adhesive into layers until it is homogenous • Number of layers: 2n (n –number of mixing elements) • For poorly miscible adhesive components, the process of introducing the two components into the mixer has increased importance • Favorable-cost mixer, maintenance-free • Cannot be adjusted • Many variants and sizes for adapting to the mixing task • Mixing ratio range: ca 100:100 … 100:3 • Viscosity differences of the components up to ca. 10

Figure courtesy of Nordson EFD

Fig. 10 Static mixing device – design, function, and characteristics

38

Equipment for Adhesive Bonding

1103

volume of waste. Static mixers are normally not rinsed and are discarded at the end of a production run. It is claimed in the literature that static mixers can only be used for mixing ratios of the order of 1:1 and for components having similar viscosities. In practice, this has long been disproved. Adhesives with mixing ratios of 10:1 and with a similarly large difference in the viscosities of the components have been successfully mixed, demonstrating the effectiveness of static mixers.

Dynamic Mixing Devices A dynamic mixer automates the process of manual mixing in a beaker (Fig. 11). The adhesive components are introduced into the mixing chamber in the correct mixing ratio. An externally driven mixing element mixes the two components until they are homogenous at the outlet nozzle. The name indicates that a moving (rotating) mixing element is used for the mixing process. The energy required for mixing is supplied from outside. Depending on the operating conditions, this leads to a temperature increase in the adhesive and a concomitant decrease in the pot life. There is a vast array of different sizes and designs available for different applications. When starting up the system, the speed of the mixing element is increased until the components are homogenously mixed. A further increase in the speed does not improve the homogeneity of the mixture. Moreover, the increased energy input unnecessarily shortens the pot life. A dynamic mixer must also be rinsed at intervals during the ongoing production, in order to prevent closing of the outlet opening. This produces

• Rotation of a stirring element in a closed mixing chamber => “Dynamic” mixer

Motor

• Mixing energy is supplied from outside • Mixer speed is a key adjusting parameter (speed limit) In

In

Comp. A

Comp. B

• Intermediate rinsing necessary (=> recycling of the rinsing liquid) • Technically complex, more expensive • Many variants and sizes available for adapting to the mixing task • Mixing ratio range: ca. 100 : 100 … 100 : 1 • Viscosity difference up to 100 : 1

Out

Fig. 11 Dynamic mixing device – design, function, and characteristics

1104

M. Peschka

hazardous waste which must be disposed of accordingly. The limits on the mixing ratio which can be processed and the viscosity differences between the adhesive components are broader than for static mixers.

38.5

Computational Fluid Dynamics (CFD) as a Tool for the Further Development of Metering Components

Starting in the 1960s, industrial adhesive bonding technology began to switch over from purely manual processing to machine processing. The current state of technology today has largely developed in an empirical way. Dosing technology and the adhesives themselves have largely developed during the past 50 years. On reflection it is clear that: – The complexity of adhesives has continuously increased. The formulations are composed of an increasing number of individual substances, and this is placing ever high requirements on the dosing technology. – It is also evident that industrial processes are becoming shorter, with an increased degree of automation. Both these trends will probably continue in the future, meaning that an understanding of the interplay between the adhesives and the processing machinery is absolutely vital. The challenges involved here can no longer be solved using empirical approaches. Accurate analysis of the flow conditions inside dosing machinery and customization of the rheological properties for special plant components will determine whether a system (adhesive plus processing plant) can be operated in an effective way. One tool that can help solve these challenges is to employ Computational Fluid Dynamics simulation. This gives an insight into the inside of dosing and mixing plants. The processes involved can be studied in quasi slow motion and this provides valuable information about the various flow anomalies which are constantly observed in practice. The underlying Finite Volume Method divides the system geometry into small (linked) partial volumes of the same magnitude analogue to the well-known finite elements in CAD (Computer Aided Design). The differential equations that describe the flow are converted in this way into difference equations (integrals to be summed). These can then be solved as a linear equation system using a high-performance Solver. If this calculation is repeated for each time step, the result is a description of the time-dependent transient flow. Figure 12 shows, by way of example, the inside of a needle valve which is commonly used as an application valve for hot melts. As the flow channel is rotation-symmetric, it suffices to consider half the cross-section in order to describe the flow.

Fig. 12 Needle valve – cross section and visualization of the flow

Real component CAD

Figures courtesy of UES AG

CFD

Viskosity in Pa s

1.5E+000 1.3E+000 1.2E+000 1.0E+000 8.4E–001 6.7E–001 5.0E–001 3.3E–001 1.7E–001 0.0E+000

1.5E+000 1.3E+000 1.2E+000 1.0E+000 8.4E–001 6.7E–001 5.0E–001 3.3E–001 1.7E–001 0.0E+000

38 Equipment for Adhesive Bonding 1105

1106

38.6

M. Peschka

Automation and Robotics

38.6.1 Prerequisites for Automating Adhesive Application When an adhesive is manually processed and manually applied, there is a control system with closed loop between the dosing and mixing plant on the one hand and the worker on the other hand. The speed with which the worker moves the application nozzle across the workpiece determines, together with the machine settings, the cross-section of the applied adhesive bead. Small machine deviations are recognized by the worker automatically and are unconsciously corrected. This control system with closed loop is absent if the application of the adhesive is also automated. An automatic moving system solely travels a programmed distance at a programmed speed. Inaccuracies of the dosing and mixing plant are not automatically corrected. The moving system does not recognize these errors. In order for the desired results to be achieved with automatic application, a number of measures must therefore be taken: • The accuracy of the dosing and mixing plant must be adapted to the requirements of the automatic application (in most cases the accuracy must be increased). • The control system of the dosing plant must contain control loops that allow monitoring of the dosing and mixing processes, so that the control system recognizes deviations from the set value and can identify these as errors. • The control systems of the dosing and mixing plant and the robotic system must communicate with each other via an interface. A further consideration is what variant is most suitable for the planned production: 1. The robot moves the adhesive nozzle across a stationary component. 2. The robot moves the component across a stationary nozzle. In the second case, the robot can also join the component after the adhesive application. It must be noted that this type of production has longer cycle times than application with a moving nozzle. Certainly, the number of robots and handling robots is lower. This is ultimately a question of economics, which should be clarified at an early stage of the project.

38.6.2 Parameters Affecting Accuracy and Methods of Compensation In order to discuss the factors which affect the dosing accuracy assume that a pneumatically driven piston dosing pump is being used. The free cross-sections of the piston dosing pump and pneumatic motor define the transmission ratio. For example a transmission ratio of 20:1, setting the air pressure at the inlet of the

38

Equipment for Adhesive Bonding

1107

pneumatic motor to 2 bar gives a material pressure in the piston dosing pump of 40 bar. This material pressure is constant in the whole system, including the downstream pipe system. If the outlet valve is now opened, there is a pressure drop across the pipe system, which to a first approximation can be calculated according to the Bernoulli’s principle. For simplicity here it is assumed that the adhesive is a Newtonian fluid and that there is laminar flow. According to Bernoulli, the pressure drop for laminar tubular flow is given by: Δp ¼

128  l  η  Q π  d4

where: η = viscosity of the fluid Q = volume flow rate l = length of the tube d = diameter of the tube The viscosity is the only variable here. The viscosity in turn depends on the temperature. Independent by of the absolute viscosities of the different adhesives, the temperature-dependent change is ca. 5% per degree Celsius. For the dosing plant under consideration with a pneumatically driven piston dosing pump, this means that fluctuations in the adhesive temperature directly cause fluctuations in the discharge rate. A temperature change of 1  C leads to a ca. 5% change in the applied amount of adhesive. For a pneumatic drive, automatic adjustment of the compressed air supply to the accuracy demanded by a specific application requires disproportionately high complexity. In practice, a pneumatic piston dosing pump does not possess such temperature compensation. This type of plant is only used with a robot in situations where fluctuations of the application rate are not a criterion for exclusion. For hydraulic dosing plants, the dosing pressure can be controlled very precisely via a servo valve in the hydraulic system. If the adhesive temperature is measured and if the hydraulic pressure, and hence the dosing pressure, is adapted to the measured values, then good temperature compensation for the temperaturedependent flow properties of the adhesive can be achieved. Rotatory dosing pumps such as gear pumps or eccentric screw pumps which are electrically driven do not have the aforementioned drawbacks. The control occurs internally in the drive system. The pump size (in ccm/revolution), the transmission ratio and the individually determined slip-factor allow exact correlation between the speed of the drive and the resulting volume flow. The internal control loop of the electrical drive adjusts the set value. The control loop usually consists of an incremental counter, which monitors the speed within the work cycle (2–5 ms). If a difference between the actual increment and the set increment is recorded (contouring error), the power uptake of the motor is adjusted with the aim of no longer having a difference at the end of the next cycle (cycle time). Factors such as viscosity changes in the adhesive result in altered power uptake by the drive.

1108

M. Peschka

The pressure in the plant is not specified, rather it independently sets itself based on the resistance to flow. A widely used means to improve the accuracy of dosing plants is to minimize the length of pipe between the dosing plant and the application valve. In the extreme case that means dosing plants having an application valve flanged onto its outlet. This design describes the current state of technology for manufacturing automobile bodies (body in white).

38.6.3 Dosing Accuracy (Dynamic Behavior of a Two-Component Metering and Mixing Machine) The processing of two-component (2-C) adhesives requires synchronous flow of the two components. The spotlight here generally falls on the accuracy of the dosing unit. The dosing unit is, however, only one part of a total system in which the adhesive components and their flow behavior play a major role. The dosing process is an intermittent process. Features of this process are that the adhesive components undergo an abrupt change from a state of non-flow to a state of flow at the start of the dosing, and vice-versa at the end of the dosing process. Consistent and synchronous dosing is only possible if the two components of the adhesive have identical pressures at the point where they flow together in the processing plant. Only then will there be an even flow of the components and compliance with the prescribed mixing ratio. If this condition is not met, one of the two components will be over-dosed. The point at which the two components flow together is not accessible experimentally. Up until now, no statements have been able to be made about how synchronous the two components flow, in particular at the start of the dosing process. It is known from the literature (Sauer 1989) that the period of time from the start of the dosing to the moment of establishment of a steady state of flow is the order of a second. A second is a short time. However, when one considers the typical application rates of up to 400 mm/s currently used or highly dynamic processes with application times of just a few seconds then that one second becomes very significant. For this reason, detailed information is desired about the interactions between an electricallydriven – and hence high-precision – adhesive processing unit and nonlinear fluids (adhesive components), in particular at the start of the dosing. The main difficulty in trying to evaluate the quality of the dosing unit (quality = accuracy with which the prescribed mixing ratio is complied with over the duration of the dosing process) is the lack of accessibility to take measurements. In practice the dosed components flow into a mixer at whose outlet the mixing ratio cannot be measured. Therefore the dosing unit has to be separated from the mixer. By a special designed nozzle, access to the outlet can be gained (see Fig. 13). The specially designed nozzle allows separate application of the two components and also allows the pressures in the line system of the dosing unit to be set the same, as is the case for operation with a directly flanged mixing unit. This so ensures there

38

Equipment for Adhesive Bonding

1109

System ready for operation Rinsing position reached Requirements – rinsing Automatic operation Start/end metering Set target value ( 0–10 V) Metering fault

Metering control system

Robot control system

Pot-life elapsed

Emergency shutdown …..

Fig. 13 Schematic representation of the experimental access to the interface between the dosing unit and mixer

is no falsification due to different pressures between the start of the operation and taking measurements. The measurements themselves were performed contactless using a laser scanner (see schematic representation in Fig. 14). As part of the studies to determine the suitability of the measuring equipment, the effects of various parameters were evaluated: – Shadow; – Undercut of the applied bead; – Color of the adhesive components (possibly require a modified exposure time)

These preliminary studies used a cured 1-C adhesive to determine the accuracy of the measuring method using an analytical microscope. Following the contactless measurements, thin discs were taken from the adhesive bead to accurately examine the cross-section under the microscope. This showed up the following deviations: • Effect of shadow: Not measurable • Effect of the undercut: +0.82% . . . +1.15% • Effect of the color: 1.55% . . . 2.64% (see also Fig. 15)

The effect of the undercut on the mixing ratio measurement will be smaller in practice than the individually measured values because the deviations for the two components are of the same order of magnitude and so cancel each other out. Only for mixing ratios far from 1:1 would there be an effect on the overall result, and even this would be less than 1%. Depending on the color of the adhesive bead, its cross-section is underestimated to a differing extent. Considering, however, the practically common combination of black and white (having deviations of 2.36% and 2.64%, respectively), these

1110

M. Peschka

Fig. 14 Schematic representation of the measuring method (Source: Mikro-Epsilon)

deviations cancel each other out. To achieve good measurement results with minimal scatter, different exposure times must, however, be chosen for these two colors. As such, this color combination represents a challenge because in order to determine the mixing ratio both beads (black and white) must be measured simultaneously. A compromise must be found here for the exposure time.

Measurement of the Mixing Ratio Figure 16 below shows the experimental set-up. The equipment was firmly secured on a coordinate table. The application base was moved at a constant speed below the special nozzle. Figure 17 shows an example signal plot, with time resolution of the dynamic behavior of the two individual components compared with the machine operation. It can be seen that after 0.1 s, the metering machine achieves steady state. There is a difference in time (delay) of 0.4 s until the components are applied. A small difference in time (delay) between the onsets of dosing of the two components is without significance. It also can be seen that a modification of the diameters of

38

Equipment for Adhesive Bonding

1111

Fig. 15 Effect of the color of the adhesive beads on the measurement results

Fig. 16 Schematic representation of the dosing unit for time-resolved measurement of the mixing ratio (left) and taking actual measurements (right)

sensor matrix receiver measuring range z-axis (profile height)

measuring range x-axis (profile width)

light source

1112

M. Peschka

glue

colour

deviation [%]

3M Scotch-Weld 9323

orange

-2,61 %

Weicon Flex 310M

grey

-2,43 %

Betamate 1496V

blue

-1,86 %

Betamate 1480 203

grey-purple

-1,55 %

Sika Sikaflex 265

black

-2,36 %

Collano BM810

white

-2,64 %

Fig. 17 Example signal plot for adhesive flow (mixing ratio 2:1) compared with dosing pump motor operation

internal piping (3.5 mm vs. 4.5 mm) in this case has no influence on components flow. After ca. 0.35 s, a steady state of flow is established for component B. Steady state of flow (component A) is established after ca. 0.9 s. This means that the prescribed mixed ratio is established ca. 0.9 s after the start of application and ca. 1.3 s after start of dosing. A very interesting aspect can be seen as well: While preparing the test the calibration of the metering pumps was not performed properly. This led to an improper function of pump A delivering ca. 20% less than required. Without the new measuring method this mistake would not be able to detect! Further detailed measurements using different adhesives have resulted in the following findings: • The non-steady state at the start of the dosing (= a state in which the mixing ratio lies outside the range prescribed by the adhesive manufacturer) can last up to 1.3 s. • The duration of the non-steady state is largely determined by the differing flow behavior of the two adhesive components. The overall system can be optimized by having detailed knowledge of that system, which comprises the processing units and adhesive components. The differing flow behavior of the adhesive components can be compensated by controlling the dosing unit. The signal plot provides accurate data about the required delay of the dosing pumps in order to guarantee synchronous dosing of the two components. The acceleration rates to achieve the prescribed mixing ratio were realized by altering the acceleration of the drive motors such that there was equivalent acceleration of the two components. The new measuring method has a number of different uses: • Verification of a dosing rate agreed between an equipment manufacturer and user • Identification of the causes of production faults • Development of customized dosing technology and specific adaptation of an overall system

38

Equipment for Adhesive Bonding

38.7

1113

Interface Between the Robot Control and Metering Control

An interface allows communication between the control systems of the mixing and dosing plant and the robot. The exchange of signals replaces the person-machine control loop for manual application. In addition to general signals such as “emergency-stop” or “automatic operation,” individual signals adapted for specific task are exchanged. An example is shown in Fig. 18. The directions of the arrows indicate the directions of the signals. For example, the signal “ready for operation” is sent in one direction: from the dosing plant control system to the robot control system. For the robot control system this means that only when this signal is received can the automatic process be started. The “emergency-stop” signal is sent in both directions. If the emergency-stop button on the dosing control is pressed, not only must the dosing plant stop immediately but also the robot. An analogous situation arises when the emergency-stop button is pressed on the robot control.

38.7.1 Accelerated Curing A demand of the industrial practice is that the time for curing an adhesive should be as short as possible. But often different requirements of the complete process anticipate the use of a quick curing adhesive. For this reason, methods were developed enabling the user to handle the adhesive as long as needed and to cure within seconds or minutes after having joint the adherends.

38.7.2 UV-Radiation The intensity of electromagnetic radiation increases with decreasing wavelength. High-energy, short-wave radiation in the UVA and UVB regions can excite electrons

material supply cartridge

M

M

pressure sensor

pump with motor

mixer inlet device special nozzle

Fig. 18 Example of an interface

1114

M. Peschka

in the outer shells of atoms and hence result in rapid chemical reactions. This mechanism is utilized for curing specially formulated adhesives. At least one of the two substrates must be permeable for the radiation. Note that materials which are transparent to the human eye are not necessarily permeable for UV radiation. Curing can take place in three different ways: Via radicals: UV acrylates cure via a radical reaction within seconds. The reaction is inhibited by oxygen. This means that adhesive which emanates from the joint forms a sticky surface. To avoid this, the curing can be carried out under a protective gas atmosphere. The UV absorption normally takes place in the wavelength region between 320 and 450 nm. Via cations: UV epoxides cure rapidly via a cationic reaction (ca. 30–90 s). Oxygen does not inhibit the reaction. Adhesive which emanates from the joint forms a dry surface. The UV absorption is usually in the wavelength region between 320 and 450 nm. Dual cure: A combination of UV curing and another curing mechanism (anaerobic, thermal, moisture-activated) is able to cure the adhesive film at points which are not accessible to the UV radiation (“shadow curing”).The UV absorption normally takes place in the wavelength region between 320 and 450 nm. UV radiation is invisible to the human eye but can irritate and damage the eyes. There is a risk of blindness and also skin irritation (sunburn). In addition to UV radiation, UV emitters also emit a large amount of visible light. The resulting brightness can also lead to eye damage if emitters are not properly used. For this reason, UV emitters which are incorporated into industrial plants must be shielded such that no UV radiation is emitted externally. Despite this protection, safety glasses must be worn at all times when working with UV emitters. During operation, most industrial UV emitters produce a small amount of ozone. This must be removed by suitable extraction, to avoid any danger to personnel working with the equipment. The complex measures required for UV protection were the impulse for the development of rapidly curing adhesives which do not require UV radiation. Initiators are now available which absorb in the visible region (400–560 nm). Adhesives formulated with these initiators cure under the influence of special blue light emitters.

38.7.3 Induction Induction is a process in which heat is induced in electrically conducting workpieces. When a low-frequency alternating current flows through a coil (= inductor), an alternating magnetic field is produced. This induces eddy currents in ferromagnetic materials and produces magnetic hysteresis losses. The heat is generated in the component itself. The effectiveness, the transferrable energy, and the comparatively accurate temperature control are major advantages over other heating methods. For adhesive bonding, the induction is used for indirect heating of the adhesive. The inductor must be adapted to the shape and desired temperature profile. Likewise,

38

Equipment for Adhesive Bonding

1115

comparison acceleration Primary pressure

1,6

Δt ≈ 0,1 s

1,4 flow rate [ml/s]

Δt ≈ 0,9 s

1,2

5 bar (A), 5 bar (B) Flow rate

Δt ≈ 0,35 s

1

2,5 ml/s

0,8 0,6

Mixing ratio

0,4

2:1

0,2

Acceleration of

0 0

0,2

0,4 A 3,5mm B 4,5mm

0,6

0,8 time [s]

1

A 4,5mm A acceleration Viscotec

1,2

1,4

1,6

dosing motors A, B: 15,8 ml/s²

B 3,5mm B acceleration Viscotec

Fig. 19 Saw blade with adhesively bonded cutting elements through induction heating

the frequency has to be adapted to the size and conductivity of the component. The reproducibility of the heating depends on the uniformity of thickness of the adhesive film, which in turn depends on the allowable variation of a component, and requires precise control of the temperature in the sensitive region for adhesive curing. A saw blade is shown in the Fig. 19. This is used for sawing materials such as marble and granite. It consists of a steel body with sintered diamond cutting segments which are normally brazed onto the saw body. The temperatures which arise during this process cause internal stress in the saw body. After production, the saw body has considerable unevenness and it has to be straightened. If the cutting segments are bonded rather than brazed, the joints are subjected to a maximum temperature of only 200  C. There is much lower internal stress in the saw body and no straightening is required after production. The engineering properties, namely, the cutting performance and wear performance are equivalent for the two manufacturing methods.

38.8

Conclusions

A large array of auxiliary equipment is available for processing adhesives, regardless of whether that is carried out manually or automatically. To describe all such equipment in detail here would be require far more space than is available in this chapter. It has therefore been attempted to describe particular pieces of equipment, such as pumps, which are widely used, covering how they function, areas of use, and also quantitative aspects. It can be stated in summary that the machine processing and the automatic application of adhesives is a highly developed technical discipline. Acknowledgments The presented findings were made within ZeDeMAB. Project 18155 N ZeDeMAB (project term: 01.05.2014 – 30.04.2016) of the Forschungsvereinigung Schweißen und verwandte Verfahren e.V. of the DVS, Aachener Straße 172, 40223 Düsseldorf, was funded via the AiF under the CORNET program of the Federal Ministry of Economics and Technology following a decision by the German Parliament. We are most grateful for this financial support.

1116

M. Peschka

References With the exception of articles in a few technical journals, little has been published about industrial adhesive bonding technology. Anyone wishing to acquire a deeper understanding of this topic is recommended to refer to the two books listed below Cognard P (ed) (2006) Adhesive and sealants, handbook of adhesives and sealants, vol 2. Elsevier, Amsterdam Endlich W (1998) Kleb- und Dichtstoffe in der modernen Technik; ein Praxishandbuch der Klebund Dichtstoffanwendung. Vulkan-Verlag, Essen Sauer J (1989) Neue Aspekte der Qualitätssicherung bei der Herstellung von Klebeverbindungen mit zweikomponentigen Klebstoffen. Heinrich Vogel Fachzeitschrift GmbH, Aachen

Environment and Safety

39

Ansgar van Halteren

Contents 39.1 39.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Health Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.1 Work Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.2 Consumer Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.3 Health Risks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.4 Health Protection When Using Adhesives: Examples of Toxicological Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.2.5 Adhesive Selection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3 Environmental Protection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3.1 Air . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3.2 Water . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3.3 Soil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39.3.4 Examples of Assessing the Environmental Impact of Adhesives . . . . . . . . . . . . 39.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1118 1118 1119 1119 1119 1120 1123 1123 1124 1124 1124 1125 1125 1126

Abstract

The handling of adhesives in a responsible way throughout their lifecycle, namely from their manufacture via the usage stage right through to recycling and disposal, is a generally recognized principle. The development and manufacture of adhesives is carried out following the principle of Responsible Care ® and the subsequent rules for sustainable development defined by the International Council of Chemical Associations. This specifically means that health protection and environmental compatibility considerations are taken into account when developing and manufacturing new adhesives. This has consequences for the composition of the adhesives, the product design, the recommendations for A. van Halteren (*) Industrieverband Klebstoffe e.V. RWI-Haus, Düsseldorf, Germany e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_39

1117

1118

A. van Halteren

application of adhesives, and the purpose of use and for the recycling of bonded products after they have been used. This chapter deals with the different aspects of consumer, work, and environmental protection, including health and safety information, related to the use of adhesives in industrial and private areas.

39.1

Introduction

The handling of raw materials and products in a responsible way throughout their lifecycle, namely from their manufacture via the usage stage right through to recycling and disposal, is nowadays a generally recognized principle. For many years now, the development and manufacture of adhesives has been carried out following the principle of Responsible Care ® and the subsequent rules for sustainable development as defined by the International Council of Chemical Associations (Fig. 1) (ICCA 2008). This specifically means that health protection and environmental compatibility considerations are taken into account when developing and manufacturing new adhesives. This has consequences for the composition of the adhesives, the product design, the recommendations for application of the adhesive and the purpose of use, and for the recycling of bonded products after they have been used.

39.2

Health Protection

The aim of health protection is to protect people against hazards and exposure. “Work protection” is used to describe protection in commercial and industrial work environments while “consumer protection” refers to private users. The three dimensions of sustainable development

Sustainable development is a concept which mankind must adopt to ensure a responsible future. The concept has environmental, economic and social dimensions. We pledge today to leave behind a world worth living in for future generations.

Environmental

Social

Fig. 1 #Industrieverband Klebstoffe e.V., Düsseldorf, Germany

Economic

39

Environment and Safety

1119

39.2.1 Work Protection Industrial and commercial users of adhesives require special protection because they work on a daily basis with adhesives, often for many hours a day. In addition to hazards associated with the physical properties of adhesives such as flammability, explosiveness, and burns when using hot melts, it is in particular chemical effects such as toxicity, skin irritation, acid burns, and allergies that have to be avoided. This is achieved by equipping the workplace (workplace guidelines) with air replenishment systems and local extractors, personal protection equipment such as work clothing, gloves, and safety glasses and by ensuring that hygiene regulations are adhered to, for example, thorough washing of the skin before breaks and at the end of work and refraining from eating, drinking, and smoking in the workplace. Safety officers monitor compliance with these safety measures. Personnel who work with adhesives undergo regular training on matters relating to work safety and environmental protection. If special risks are involved, then personnel undergo regular medical examinations.

39.2.2 Consumer Protection Although in a professional work environment possible hazards can be efficiently managed using suitable protective measures, even in situations where there is longterm use, the situation is different for private individuals using adhesives at home where no special protective measures are taken. Different requirements are hence put on these adhesives. This is also so because private users, unlike industrial users, generally have no knowledge of the properties and potential hazards of products. That is why only relatively few of all the known types of adhesives are available to private users, and even then only small amounts are made available in the form of tubes, cartridges, and tins. On the other hand, private users do not use adhesives every day, but rather only occasionally, and even then only in limited quantities and for limited periods of time. As such, the protective measures described for industrial users are generally neither possible nor necessary. It is nevertheless essential that the safety information given on the small packages is observed, as well as basic principles of work hygiene.

39.2.3 Health Risks The area of toxicology is concerned with issues relating to the effects of chemical compounds and mixtures of chemicals. Assessing the health risk is a multistep process. The first step is sound assessment of possible undesirable properties of substances based on recorded data. In the second step, the quantities of material involved and the nature and degree of any possible contact are determined. It is then investigated whether an undesired effect of a material can be caused as a result of this contact. There is no health risk for people if there is no contact with the undesired

1120

A. van Halteren

Risk assessment

Toxic properties of the material

Exposure Contact with the material

Hazard potential

level of exposure

Level of risk

Fig. 2 #Industrieverband Klebstoffe e.V., Düsseldorf, Germany

material. The risk assessment determines whether and to what extent there is a health risk to people as a result of the relevant hazard potential of the substance and the nature and degree of exposure (Fig. 2). Although the exposure can for example be reduced by automated processing (e.g., robots) or by taking suitable protective measures (e.g., gloves, extraction of vapors, safety glasses, etc.), the hazard potential is a fundamental property of a substance or formulation (mixture of substances). The hazard potential generally decreases when the fraction of hazardous material present is lower. Small amounts of a hazardous substance can often be present without the product having to be accordingly labeled. The physicochemical properties of the substance, the nature and degree of exposure, and the ability of the substance to get into the body determine the relevant amount and the resulting dose that can be taken up by the body. In general, there is no undesired effect below a certain dose. The dose level determines whether and to what degree a substance can harm a person’s health. The risk to human health is determined from the hazard potential of the substance and the exposure (opportunity for contact).

39.2.4 Health Protection When Using Adhesives: Examples of Toxicological Assessment Of the many different types of adhesives, those discussed below are those that are also commonly used in the household:

Physically Hardening Adhesives The active components are mainly solid polymers and resins. For application these must be converted to a liquid form. This can either be carried out by users by means

39

Environment and Safety

1121

of heating (hot melts) or can be carried out by adhesive manufacturers by dissolving the components in organic solvents or dispersing the components in water. The dry, fully aired adhesive is generally unreactive and biologically inert. There is hence usually no health hazard. In contrast, there is a potential health hazard from auxiliary components, such as organic solvents, that are present in some of these adhesives. These can make up as much as 80% of the weight of the adhesive product.

Hot Melts These adhesives contain largely polymers and resins and only small amounts of auxiliary materials. There is virtually never a health hazard. When applying these adhesives manually, there is the risk of burns and users must protect themselves against this. During heating, small amounts of auxiliary materials, contaminants, and cleavage products can be liberated, but these are insignificant when small amounts of adhesive are being processed. In an industrial or commercial environment, an extraction system is recommended due to the larger quantities being used and the longer working times with the adhesives.

Solvent Containing Adhesives (e.g., Contact Adhesives) In these adhesives, the polymers and resins are dissolved in organic solvents. The hazard potential is determined by the nature of the solvent (e.g., flammability, irritation potential). Due to the high volatility of the solvents, exposure by inhalation of the vapors is the biggest problem. For most solvents, the maximum concentration at the workplace and limiting factors are laid down (workplace limit value). Due to the small amounts of adhesive used by private users, these limit values are generally not reached or are only exceeded for a very short time. Dispersion Adhesives (e.g., PVAc – Polyvinylacetate/Wood Adhesives) In the adhesives, the organic solvents are replaced by water and suitable polymers are dispersed in the water. There are hence no potential health hazards from organic solvents. However, water-based adhesives are sensitive to attack by microorganisms (e.g., mold formation). For that reason, dispersion adhesives contain small amounts of preservatives for protection purposes. The potential health hazard is the triggering of allergic skin reactions, for example, allergic reactions triggered by natural polymers such as natural rubber and non-modified colophony resins. The risk of sensitization in nonallergic people is generally extremely low due to the very small amounts of preservatives in the adhesives. Skin contact is here the exposure issue. Depending on the mode of application, skin contact may be unavoidable, as, for example, when using wallpaper pastes. However, here the concentration of preservatives is reduced as a result of mixing with water. As even wearing protective gloves for a long time can lead to skin irritation (e.g., caused by constant sweating), it is worth considering whether the very low risk of direct skin contact causing an allergic reaction justifies wearing protective gloves.

1122

A. van Halteren

Pressure Sensitive Adhesives Private users only come into contact with these adhesives in the form of selfadhesive articles such as labels, adhesive tape, etc. As such, these adhesives represent no hazard to private users in practice. Such articles are manufactured industrially using adhesives in the form of solutions, but mostly using dispersions and melts.

Chemically Curing Adhesives The chemically reactive monomers/oligomers and hardeners and cross-linking agents determine the potential health hazard of these products. Once fully cured, the adhesive polymers are in general nonhazardous. Exposure and risk considerations hence only apply for the time period up until the adhesives have fully cured. Cyanoacrylate Adhesives (Superglues) This group of adhesives reacts with water from the surroundings or water on the substrate. As private individuals only usually use small drops when applying the adhesive, the potential health hazard here is that if there is contact with the eyes or splashes of adhesive enter the eye then undesired bonding can take place (can gradually be dissolved using a soap solution). When being used industrially and commercially, possible irritation caused by the cyanoacrylate monomer, thermal effects, and the rapid polymerization reaction have to be taken into consideration. Increasing the humidity prevents irritation of the respiratory tract. Therefore, cyanoacrylate adhesives must be specifically labeled. When carrying out major bonding tasks, not only is it recommended to adjust the humidity of the air but also to wear safety glasses and protective gloves. Silicone Adhesives Silicones react with water. Depending on the type of silicone, this reaction releases either acetic acid or alcohols. The acetic acid can be clearly sensed by the nose before any irritation begins. In general, the slowly released amounts are so small that they present no health risk, especially in the case of private users. Old formulations of neutral silicones that release butanone oxime must be labeled, but they are only used nowadays for special applications. Epoxy Resin Adhesives Epoxy resins are widely used by hobbyists and in home do-it-yourself applications because of their good bonding properties, for example, for bonding glass, ceramics, or metals. This is particularly so because the supplied twin-nozzle container allows even the smallest quantities to be dispensed and then mixed. Liquid epoxies resins, and also cross-linking agents, irritate the skin and eyes and cause skin hypersensitivity. As such, hobbyists should avoid epoxy resin adhesives contacting the skin. For the industrial application of epoxy resins, safety glasses and suitable gloves must be worn. Epoxy resins having a molecular weight up to 700 are irritating to eyes and skin and they bear the potential risk of causing skin hypersensitivity. Therefore, epoxy resin systems must be labeled accordingly.

39

Environment and Safety

1123

39.2.5 Adhesive Selection For the private use of adhesives, there are no legal requirements. The selection of an adhesive is left largely to the experience of users and the recommendations of manufacturers. For commercial users, and in particular for industry, the selection of a suitable adhesive system for a particular application must be based on the technical requirements and the potential health hazard. In a situation where two adhesives are equally suitable, the one providing the lower potential health hazard must be chosen. If using an adhesive that does represent a potential health hazard is unavoidable for technical reasons, the exposure must be limited by taking suitable measures, which may even include automating the process, such that a risk to health is prevented. Safe working with adhesives is guaranteed by observing this regulation and the other regulations described in this section. Adhesive manufacturers provide support here by providing information in the form of technical data sheets and safety data sheets. If necessary, users should contact the adhesive manufacturers directly. As is clear from the iterations in this section, adhesive manufacturers make considerable efforts to protect the health of both private and professional users of their products. It must be mentioned here that from a quantity point of view, only a small number of adhesives have to be classified as hazardous formulations. Private users complying with the safety instructions given on containers and industrial workers complying with relevant company regulations in their production facilities serve to protect people in accordance with the motto: “prevention is better than cure.”

39.3

Environmental Protection

Environmental issues have gained increasing importance over recent decades. The following is an example of a positive contribution to the environment: The car manufacturing industry uses up to 40 kg of adhesive sealants in each vehicle. The main objective of this is to save energy in the form of fuel by means of so-called “lightweight design,” whereby less weight has to be moved over the lifetime of the car and so resources are saved (e.g., 3-l car). On the other hand, the adhesives in this same example could burden the environment if there are residues left over from the application and any cleaning. At the end of the lifecycle of the car, the adhesive that was used must not adversely affect the recycling process and must be disposed of in accordance with regulations. In order to determine the effects on the environment, an environmental assessment is carried out. The expected concentration in the environment is calculated using a model. Simultaneously, the concentration at which no harm is expected to environmental organisms is determined. The calculation models and determination methods are laid down in international standards (European Chemicals Bureau 2003; European Chemicals Agency 2008). An adverse effect on or harm to the environment can be excluded with certainty if the Predicted Environmental Concentration (PEC) is less than the predicted concentration for which no

1124

A. van Halteren

harm to environmental organisms occurs (Predicted No-Effect Concentration – PNEC), taking into account safety factors.

39.3.1 Air The emission of organic solvents is detrimental to air quality. As such, considerable efforts have been made by the adhesives industry over many years to switch to low-solvent adhesives and where possible to solvent-free adhesive systems. There has been considerable success, and this has also been beneficial for health protection. Today, adhesive bonding is only responsible for less than 3% of all the solvent emissions in Europe. The few large-scale industrial applications operate with solvent-recovery systems. One typical example for a significant reduction of the use of solvent-based adhesives is the construction industry: While, for example, in Germany in 1985 the fraction of solvent-containing adhesives was still about 50%, it is in 2010 less than 10% with a still decreasing tendency. Dispersion adhesives and reactive systems have largely replaced these solvent-containing adhesives (Adhesion KLEBEN and DICHTEN 2010).

39.3.2 Water With dispersion adhesives there is a particular risk of contaminating surface waters. Organic polymers and pre-polymers, although not easy to biodegrade, can be removed in biological treatment plants with the excess sludge. The use of these adhesives in handicraft work, in the home, and for DIY work normally results in only relatively small amounts of adhesive ending up in large amounts of wastewater (communal treatment works). As polymers usually have low toxicity for water organisms and due to the very low concentrations of adhesives in the wastewater, an adverse effect on the treatment plant and adjacent surface waters (outfall) is not expected in this case. In industry, residual adhesive and rinse water must be disposed of in accordance with waste legislation.

39.3.3 Soil When used by private individuals, liquid adhesive residues are usually collected in the hazardous waste collection boxes, but are also often disposed of with the rest of the household waste. The amounts involved here are, however, small. For industrial and commercial users, disposal must be carried out in accordance with the so-called waste code numbers (to be found on the safety data sheets), which determine how the waste is disposed. Adhesive residues are normally disposed of as landfill or incinerated. Regarding the latter, specific contamination of the air by the incineration plants is not expected. The energy used in the manufacture of the adhesives can be partially recovered here.

39

Environment and Safety

1125

39.3.4 Examples of Assessing the Environmental Impact of Adhesives Production and Industrial Use of Adhesives In industry, residues of cured adhesives and also non-cross-linked and liquid adhesives occur as waste. The former, as solid materials are either directly passed on as waste for disposal, or are preferably passed on for recycling. In the least favorable environmental scenario, they are disposed of as landfill. They are, however, usually recycled, either via composting or for energy recovery via incineration. Depending on how they have been treated, liquid adhesives are discharged with the wastewater and are either directly passed to a treatment plant or, if necessary, passed to a treatment plant after undergoing a specific pretreatment. There they are biodegraded or removed. Do-It-Yourself/Private Use of Adhesives Adhesives used in private household/Do-It-Yourself have to be considered in different groups: • Solid adhesive waste from the home such as hardened paper or wood glue is usually disposed of with the rest of the household waste. • Water is usually used to remove water-soluble, hardened adhesives in the home, for example, for removing wallpaper paste. This adhesive-containing water is disposed of with the household wastewater via the water treatment plant. The adhesive components, for example, cellulose derivatives, which are dissolved to varying extents, can easily be eliminated under real environmental conditions, meaning that only very small amounts enter the outfall (river). When composting is used for recycling, cellulose decomposes slowly but thoroughly – as known for plant materials. • Adhesives used in the home, for example, roofing adhesives, or around the home, for example, for the car, are exposed to processes that can wash out the adhesive, for example, rain. This means that a small amount of adhesive ends up in drain water and then directly enters surface waters. Considering all the adhesives that are in use, the expected environmental concentration (PEC) is determined. The most unfavorable scenario is assumed in order to ensure that all conceivable possibilities have been taken into account (Fig. 3).

39.4

Conclusion

Adhesives as a whole do not represent a major environmental problem. Nevertheless, cured and nonrecyclable residues, as well as excess adhesive from applications are waste materials. These materials not only have to be disposed of, but also represent an unnecessary use of materials and resources. The principle of recyclable design, made possible by having detachable bonded joints, will be important in the future. Besides

1126

A. van Halteren

Adhesives and the end-user Rain Adhesives on the home

Washed out adhesive Adhesives on the home Adhesive in the environment Solid adhesive

Residual adhesive Adhesive

Disposal

Fig. 3 #Industrieverband Klebstoffe e.V., Düsseldorf, Germany

technologies for separating bonded materials, a further challenge of bonding technology will be resource-friendly optimization of production and application processes. It will be necessary to incorporate non-removable adhesive into future recycling processes. This will mean that the adhesive to be used for manufacturing a component will have to be customized to the recycling process at the end of the component’s lifetime.

References Adhesion KLEBEN and DICHTEN (2010) Industrieverband Klebstoffe, Handbuch Klebtechnik 2010/2011, Wiesbaden ECB – European Chemicals Bureau (2003) Technical guidance document on risk assessment of chemical substances following European regulations and directives. European Communities, Ispra, Apr 2003 ECHA – European Chemicals Agency (2008) Guidance on information requirements and chemical safety assessment, Helsinki, May 2008 ICCA – International Council of Chemical Associations (2008) ICCA Responsible Care Report 2008, Brussels

Part VIII Quality Control

Quality Control of Raw Materials

40

Kazutami Wakabayashi

Contents 40.1 40.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Types and Roles of Raw Materials for Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.1 Main Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.2 Solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.3 Tackifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.4 Plasticizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.5 Fillers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.6 Thickeners . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.7 Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.8 Other Important Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.9 Functional Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.2.10 Functional Fillers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.3 Quality Control Procedures for Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.3.1 Acceptance of Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.3.2 Inspection for Acceptance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.3.3 Quality Control of Raw Materials in Storage . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.3.4 Inspection for Shipment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4 Material Standards . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.1 Main Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.2 Solvents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.3 Tackifiers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.4 Plasticizers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.5 Fillers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.6 Thickening Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.7 Pigments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.4.8 Other Raw Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1130 1130 1131 1133 1133 1134 1134 1134 1134 1134 1135 1138 1139 1139 1140 1143 1143 1145 1146 1146 1146 1147 1147 1148 1148 1148 1149 1150

K. Wakabayashi (*) APS Research Co Ltd, Osaka, Japan e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_40

1129

1130

K. Wakabayashi

Abstract

Since raw materials for adhesives influence the performance of final adhesive products, the selection, storage, handling, and testing of the raw materials are very important. In this chapter, types of raw materials including basic resins, hardeners, fillers, and functional additives for adhesives are firstly described. In addition, the roles of the materials are also shown. Practices for the quality control of the raw materials consisting of acceptance inspections and testing are also explained. Other tests conducted when adhesive products are shipped are explained too. These tests should be carried out following test standards. The standards for testing raw materials, and standards for materials are also mentioned in this chapter.

40.1

Introduction

Quality control of products can be done considering only four elements called “4M.” Adhesives are not an exception to that concept. 4M is the abbreviation of “Materials,” “Method,” “Machine,” and “Man.” Managing the four elements, the quality of products can be ensured. Another word, “5M,” which includes “Management” is sometimes used instead of 4M (Ozawa 1987). What is the definition of “Management” in terms of quality control? Generally speaking, it can be defined as the actual conducts to carry out the whole cycle of “Plan,” “Do,” “Check,” and “Action” (PDCA) for all items. The PDCA cycle is called the Shewhart cycle or Deming cycle (Deming 2000). Note that the management of row materials is the most important to control the quality of adhesives. Commercial products of adhesive are varied and the number of raw materials involved in them is almost infinite. The reason to employ so many raw materials is that adhesives need special compositions to meet the different demands or conditions such as adherends, applications, conditions of use, environments, and laws such as regulation concerning the registration, evaluation, authorization, and restriction of chemicals (REACH). Figures 1 and 2 show the main composition of adhesives commercially available. Since adhesives are varied, the raw materials also vary to a great extent. In this chapter, the description is concentrated in synthesized organic adhesives which are commercially available and widely used, because this reduces the number of combinations.

40.2

Types and Roles of Raw Materials for Adhesives

An ordinary process of adhesion is performed as shown in Fig. 3 (see Part G). The first step of the process is application of adhesives. Every adhesive must be liquid just before the application. The liquefaction can provide fluidity to the adhesive and help it to enter the micro structure of an adherend surface. The phenomenon is called “wetting.” Methods to induce the liquefaction are as follows:

40

Quality Control of Raw Materials

1131 Sodium silicate Cement

Inorganic Adhesive

Adhesive

Portland cement Mortar Plaster Magnesia cement Litharge cement Dental cement

Ceramics Organic Adhesive

Natural Materials

Synthetic resin

Starch Protein Natural rubber

Dextrin Rice paste Glue Casein Soy protein

Asphalt

Fig. 1 Classification of adhesives by categories of materials

1. Use of solvents such as water or organic solvents 2. Dispersion of micro particles in water, i.e., emulsion adhesives or latex dispersion adhesives 3. Heating, i.e., hot-melt adhesives 4. Use of liquid monomers or oligomers that can be polymerized

Water, solvents, surface-active agents, liquid monomers, or oligomers are the key components for liquefaction of adhesives. The performance of adhesives depends on the mechanical properties of the cured bulk adhesive. To emphasize the performance, many kinds of additives are blended to an adhesive, i.e., tackifiers, plasticizers, antioxidants (age inhibitors), curing agents (vulcanizing agents), adhesion prompters, and fillers. Adhesives consisting of mono component are very rare. Functional adhesives have to have functions such as flameproofness or conductivity besides the intrinsic function to bond adherends together. Flame retardants or conductive fillers are used for the purpose. They are also the components of the adhesives. More information on composition of adhesives is given in ▶ Chap. 13, “Composition of Adhesives.”

40.2.1 Main Components The main components of adhesives are polymeric materials such as elastomers or synthetic polymers. The main component type for a certain adhesive is decided by the designer considering the type of adherends and use conditions. Control items to

1132

K. Wakabayashi

Fig. 2 Classification of synthetic resins for adhesives

Fig. 3 Adhesive and adherend preparation in bonding processes

Adherends

Adhesives

Degrease or Washing

Measurement & Mixing

Surface treatment

Application

Bonding

Pressurizing & fixing

Curing

(This process should be done within the open time of adhesive)

40

Quality Control of Raw Materials

1133

be considered is as follows: solubility parameters, average molecular weights, Mooney viscosity, crystallinity, and existence or nonexistence of functional groups. Materials that can be used as main components of elastomer adhesives are polychloroprene rubbers, nitrile rubbers (butadiene-acrylonitrile copolymer), SBR (styrene-butadiene copolymer), thermoplastic elastomers (SBS, SIS, SEBS, SEPS etc.), butyl rubbers, acrylic rubbers, silicone rubbers, polysulfides, modified polysulfides, modified silicone rubbers, and silyl urethanes. Materials that can be used as main components of thermoplastic adhesives are vinyl acetate, polyvinyl alcohol, polyvinyl acetal, ethylene-vinyl acetate copolymer (EVA), vinyl chloride, acrylic resin, polyamide, cellulose, alfa-olefin resin, and waterborne isocyanate resin. Materials that can be used as main components of thermosetting adhesives are urea resin, melamine resin, phenolic resin, resorcinol resin, epoxy resin, structural acrylic resin, polyester, polyurethane, polyimide, polybenzimidazole, etc.

40.2.2 Solvents The role of solvents is to reduce the viscosity of adhesives and to improve fluidity. That can provide the adhesives wettability to create an intimate contact with the surface of adherends. Solvents must be able to dissolve the components of adhesives. Solubility parameter is an index to show the soliditivity of solvents. A solvent can dissolve a high amount of materials whose solubility parameters are close to that of the solvent. Water, alcohols, aromatic hydrocarbons (e.g., toluene and xylene), ketones (e.g., methyl ethyl ketone and cyclohexanone), acetate esters (e.g., ethyl acetate and butyl acetate), n-hexane, cyclohexane, methylene chloride are used due to their solubility, dehydration rate, noncombustibility, and workability. To meet the demands concerning environmental issues, the use of some solvents such as toluene, xylene, ethylbenzene, and styrene is restricted by laws such as the air pollution control law legislated by Ministry of the Environment in Japan (The Ministry of the Environment 1996).

40.2.3 Tackifiers A tackifier is a component to provide initial tackiness to adhesives. The compatibility of the tackifier to the main components is very important to increase the tackiness without degradation of the adhesive mixture. For that purpose, natural resins such as rosin and dammar, modified rosins such as coumarone-indene resin, polymerized rosin, hydrogenated rosin and rosin ester, and polyterpene resin are used. Recently, alkylphenol resins, terpene phenol resins, and xylene resins are out of use because of their hazardous nature inducing sick building syndrome.

1134

K. Wakabayashi

40.2.4 Plasticizers Plasticizers can give flexibility to resins and are often used for vinyl resins. For instance, phthalic esters such as dioctylphthalate (DOP) and dibutyl phthalate (DBP), aliphatic diprotic acid esters, glycol esters, resin acid esters, and phosphoric esters are used as plasticizers. The use of phthalic esters has been decreasing because of environmental regulations.

40.2.5 Fillers The purposes to mix fillers with adhesives are (1) viscosity control, (2) suppression of resin invasion into porous adherends, and (3) strengthen the cured adhesive layer. Commonly used materials for fillers are minerals such as calcium carbonate, clay, talc, and diatom earth. Cellulose powders and recycled rubber powders are also used for the purpose.

40.2.6 Thickeners Thickeners are water-soluble polymers used for increasing the viscosity of emulsion, dispersion, and latex adhesives. Typical materials for thickeners are casein, methylcellulose, carboxymethyl cellulose, polyvinyl alcohol, and sodium polyacrylate.

40.2.7 Pigments Pigments such as titanium white, carbon black, and organic colors are used for coloring adhesives. Mixing of pigments in adhesives can make them less noticeable. In addition, the method has another function in which adhesives mixed with pigments can be visualized by color. Pigments must not react with the other components of an adhesive. Preservation stability of the mixture has to be checked experimentally.

40.2.8 Other Important Components Antiseptic agents, antiaging agents, antifoam agents, adhesion promoters (e.g., aminosilane coupling agents, alkyltitanate, etc.), or fragrances can be applied to adhesives as necessary.

40

Quality Control of Raw Materials

1135

40.2.9 Functional Components Recent adhesives may include new types of components that can be used as main components or modifiers.

Silane Coupling Agents It is widely known that aminosilane coupling agents can be added to adhesives to improve their adhesion strength. Aminosilane coupling agents have organic groups and alkoxyl groups that can modify the main resin components of adhesives. Alkoxyl groups have compatibility with inorganic materials. They also condense with each other and produce cross-linked structures, which leads to desirable properties such as resistance to climate, acid tolerance, heat resistance, resistance to solvent attack, etc. Aminosilane coupling agents have recently been introduced to many types of resins or elastomers. Special coupling agents suitable for particular materials have also been developed. The methods to apply aminosilane coupling agents to resin or elastomers are as follows (Yanagisawa 2007): 1. Graft reaction 2. Chemical reaction 3. Copolymerization These methods are shown in Fig. 4.

Fig. 4 Methods to introduce silane coupling agents to molecules

1136

K. Wakabayashi

Functional Acrylic Materials Oxazoline acrylic polymers have an oxazoline group, which is a kind of cyclic imino ether, as side chains. Ethylenimine acrylic polymers are modified with ethylenimine. Both materials have been used as main components of acrylic adhesives (Noda 2005). Oxazoline acrylic polymers can be applied to paints, coats, adhesives or pressure-sensitive adhesives (PSAs), and textile treating agents as water-base crosslinkers. Applications of oxazoline acrylic polymers are as follows: 1. Water-base cross-linkers: Although carboxyl groups in water-base resins are neutralized by amines or ammonia, oxazoline groups do not react with carboxyl groups and are stable in liquid conditions. When the water-base resin dehydrates, neutralizing agents are evaporated, and the carboxyl groups start to react with the oxazoline groups. 2. Primers for plastics: Oxazoline acrylic polymers have good adhesion compatibility to plastics. The reason seems to be the presence of polarity, affinity, and interaction caused by oxazoline groups that are a kind of cyclic imino ethers. The carboxyl groups in the polymer react with the oxazoline groups and they can also give the polymers good adhesion to plastics. 3. Film laminating adhesives: Although EVA emulsion is often used for film lamination as adhesives, the bonding strength to polyvinyl chloride (PVC) and polyolefin firms is not high. The defect can be corrected adding oxazoline acrylic polymers to EVA emulsions having carboxyl groups. Ethylenimine acrylic polymers have high bond strength to many types of materials because ethylenimine groups are very reactive. The reactions caused by ethylenimine groups can be categorized as (a) polymerization by ring-opening ionic polymerization, (b) introduction of primary amine group in side-chain of acrylic polymers by ring-opening addition of ethylenimine groups to carboxyl groups, and (c) aziridinyl reaction (modification with preservation of aziridinyl groups). The good reactivity of ethylenimine acrylic polymers enables them to be cross-linkers for water-base polymers having carboxyl groups. When the polymers are used as a part of two-part cross-linking systems, use of a double-headed gun is necessary.

Oxetane Polymers Oxetane polymers are polymers having oxetane rings that are a kind of fourmembered cyclic ether. The ring has high strain and basicity, and the polymer can cause cation ring-opening polymerization, as shown in Fig. 5. The characteristic of oxetane polymers are as follows (Sasaki and Crivello 1992; Crivello and Sasaki 1993a, b; Sasaki et al. 1995; Sasaki 2000): 1. The result of the Ames test is negative even in the condition of low molecular weight monomers. 2. Quick curing to mix to epoxy materials. 3. Relatively high molecular weight of generated polymers. 4. Diversity of synthesis methods because of oxetane ring stability under alkali conditions. 5. Volume shrinkage due to curing is not large and almost equal to epoxy resins.

40

Quality Control of Raw Materials

1137

Fig. 5 Cationic ring-opening polymerization of cyclic ethers

Such characteristics are suitable for using the material for modifiers of UV curing epoxy adhesives applied to electronic parts.

Amorphous Polyalphaolefin Amorphous polyalphaolefins are olefinpolymers that are in amorphous condition and have relatively low molecular weights. They are synthesized by copolymerization of propylene alone, propylene and ethylene, or butane-1. The characteristics of amorphous polyalphaolefins are as follows: 1. Good adhesion compatibility for polyolefins such as polyethylene (PE) and polypropylene (PP). 2. They can be used alone as hot-melt adhesives. 3. Mixture with EVA hot-melt adhesives may be sprayed to adherends. 4. They can be mixed with large amounts of mineral fillers. 5. They can be mixed in asphalt with a good dispersion. 6. Safe and harmless (compatible with the standard: US Food and Drug Administration (FDA) 21CFR.175.105).

1138

K. Wakabayashi

Fig. 6 Deformation mechanism of thermally expandable microcapsules (TEMs)

These characteristics lead to applications such as hot-melt adhesives, modifiers for hot-melt adhesives, modifiers for resins or elastomers, modifiers to prevent corrosion for waxes, and modifiers for asphalts.

Thermally Expandable Microcapsules Thermally expandable microcapsules (TEMs) are micro spheres of 10–40 μm in diameter, having thermoplastic shells filled with liquid hydrocarbon of low boiling points. As their temperature rise, the inner pressure increases and the shell is softened. That leads to huge expansion of the sphere up to 20–40 times in volume, as shown in Fig. 6. The TEMs can be applied to dismantlable adhesives, thermally expandable adhesives, or sealants. Core-Shell Rubber Core-shell rubbers have been developed to improve the ductility of epoxy resins without heat resistance reduction. Core-shell technology is a kind of capsulation process in which hydrophobic core resins can be covered with hydrophilic shell resins. There are many combinations of core and shell materials and wide varieties of dimensions, 40–500 nm in diameter, for instance. A method to pre-react liquid carboxylated nitrile butadiene rubber (C-NBR) and an epoxy resin and mix the product into the epoxy resin has been used for long time and it is a typical approach to design structural adhesives recently. Mixture of core-shell rubber is much more effective than those by the C-NBR method.

40.2.10 Functional Fillers Carbon Nanotubes Carbon nanotubes are hollow carbon structures as a graphite sheet wound cylindrically. Their diameter is few Å to 100 nm. Since the aspect ratio is very large, the material has a fibrous form. Carbon nanotubes having one graphite layer are called SWCNT (Single-Walled Cabon Nanotube). Other carbon nanotubes are called

40

Quality Control of Raw Materials

1139

MWCT (Multiwalled Carbon Nanotube) (Iijima 1991; Saito 2003; Fan et al. 1999; Nakayama 2002; Akita et al. 2001). The characteristics of carbon nanotubes are as follows: 1. 2. 3. 4.

Chemical stability High electric and thermal conductivity High mechanical properties High electron emission property

The carbon nanotubes are very promising to apply for adhesives as fillers to emphasis the strength and electric or thermal conductivity. Mixing of carbon nanotubes is not easy because of their strong cohesive attraction. Dispersion technology has to be established to obtain good characteristics.

Fine Calcium Carbonate Fine calcium carbonates, coated with aminosilane coupling agents, have been developed. Their diameter is about 20 nm. The material is promising as fillers to reinforce structural adhesives or elastomer adhesives (Tutui 2006). Micro Balloons Micro balloons are convenient to reduce the weight of adhesives because the density of the balloons is very low. Typical micro balloons are 5–300 μm in diameter, and 0.7 g/cm3 in density. They have hollow structures comprising of silica (60–65 wt%), alumina (27–33 wt%), and ferric oxide (less than 4 wt%) (Tozaki 2006). The mixture should be carried out so that the micro balloons could not be broken. Since the macro balloons are difficult to be dispersed into resins, they are mixed with solvents and dispersed. After the mixture, the solvents are extracted by evaporation.

40.3

Quality Control Procedures for Raw Materials

40.3.1 Acceptance of Raw Materials Acceptance inspections are necessary when raw materials are received. The name, type, and quantity of a material should be verified. The delivery conditions of raw materials, for instance, chilled transport, are also important in some cases and should be checked. After the acceptance inspections, there are several inspection tests to check the quality of the materials, which should fit the requirements. When these tests are carried out by suppliers of materials, users should check the specification documents indicating the test results. Recently, suppliers tend to contract for these tests. However, tests by users are still important to make sure the quality of their products. Capabilities of users to conduct tests are indispensable to decide the types and sequences of tests, which should be discussed and agreed with the suppliers testing the materials. Types of inspection tests are explained in the following section. Specification documents often include information for workers’ health protection

1140

K. Wakabayashi

such as material safety data sheet (MSDS), which should be always available to the worker. When a raw material is accepted, the storage term of the material starts. Thus, the inventory list of a warehouse should be updated. Stored raw materials should be used “first in, first out.” Too long storage term than the shelf life of a material is prohibited. On the package of each material, tag labels indicating necessary information such as receiving date should be stuck. Inventory managements are important not only for the performance proof of materials and final products, but also for accounting managements and chemical material managements for acting in accordance with pollutant release and transfer register (PRTR), which is linked to raw material purchase, production, and shipping of final products. Nowadays, the updating of inventory lists can be automatically done with that of the PRTR database using computers.

40.3.2 Inspection for Acceptance In inspection tests for receiving raw materials, the following properties should be investigated: chemical equivalent, impurity content rate, molar weight, viscosity, specific weight (density), color, transparency, conductivity, permittivity, etc. Among these properties, chemical equivalent, impurity content rate, molar weight, and viscosity are important.

Chemical Equivalent For raw materials of reactive adhesives, the chemical equivalent values are very important to specify the reactivity, and often tested. Epoxy equivalents for epoxy resins, and functional group equivalents for hardeners are experimentally determined. Methods to measure the epoxy equivalents are described by industrial standards as follows: International organization for standardization (ISO) 3001, Japan industrial standards (JIS) K 7236, and American society for testing and materials (ASTM) D 1652. For instance, JIS K 7236 indicates the methods using titration. In this process, an epoxy sample is solved with chloroform, mixed with acetic acid, tetraethylammonium bromide, and crystal violet solution. The epoxy sample solution is finally titrated with perchloric acid–acetic acid solution. If automatic titrators can be used, the epoxy sample solution is titrated with perchloric acid–acetic acid solution by the machines. Functional group equivalents for hardeners should also be measured because these are important parameters for adhesive curing. Impurity Content Rates Adhesives for electronic applications are sensitive to impurities. For instance, inclusion of conductive particles often induces problems of insulation breakdown. Chlorine inclusion can cause corrosion of metal parts such as electrodes made of oxygen-free copper foils on flexible circuit boards that can react with chlorine in adhesives for bonding the foils and the boards, yield dendrites, and lead to insulation

40

Quality Control of Raw Materials

1141

breakdown. Since the phenomena occur at very low density of chlorine, the content rates should be checked by standards such as ISO 21627 and JIS K 7243.

Molecular Weight Since main components of adhesives are usually monomers of low molecular weight, the performance of the adhesives is not influenced by the variation of the molecular weight. Thus, it is not important to measure the molecular weights of the components. In contrast, for pressure-sensitive adhesives (PSAs), raw materials such as rubbers and tackifiers have relatively higher molecular weight than those of adhesives, and the variations lead to deviations of the final performances. Therefore, the molecular weights of the raw materials for PSAs are often measured by experimental methods such as end group determination, colligative property measurements, light scattering, ultracentrifugation, viscosity measurements, and gel permeation chromatography (GPC). These methods should be selected properly for objective materials. Viscosity Viscosities of raw materials are very important because they can influence not only the properties of final adhesive products, but also the production process. Methods for measuring viscosities of liquid raw materials are included in the following standards: ISO 2555, ISO 3219, JIS K 7117, and ASTM D1084. Viscosities of liquids can be described as: τ ¼ μγ_

(1)

where, τ, μ, and γ_ signify shear stress, viscosity, and shear strain rate, respectively. Consider the situation shown in Fig. 7, in which there are two parallel plates of area A with a separation distance of h, where the plates translate in relation to each other with a velocity v. The shearing force Fs is given by the following equation. Fs ¼ μA

v h

(2)

Although the international system of units (SI unit) for viscosity is Pas, cps is also used often for raw materials of adhesives and PSAs. For viscous liquids, BrookfieldFig. 7 Relation between shear force and shearing velocity in viscous fluid

1142

K. Wakabayashi

Fig. 8 Types of viscometers for a viscous fluid

type viscometers or coaxial-cylinder-type viscometers, shown in Fig. 8, are usually used. For higher viscous materials, cone-plane-type viscometers (Fig. 8), Mooney viscometer, and rheometers are used. For example, a Brookfield-type viscometer has a spindle rotating in the liquid with a constant rotational velocity, and the resultant torque is measured. The viscosity of the liquid can be calculated from the torque and calibration results of standard viscous materials commercially available. For elastomers and rubbers, Mooney viscometers are used, which are documented in standards ISO 289 and JIS K 6300. Brookfield-type viscometers can also be used for elastomers. Elastomers should be dissolved in a solvent in content of 10–20 wt% before measuring with Brookfield-type viscometer.

Density (Specific Weight) Densities and specific weights are basic material properties and should be measured to check errors of adhesive production. The measurement methods of specific weights are defined by standards ISO1183, JIS K7112, and ASTM D792. Densities are basically equivalent to specific weights, and the selection depends on the nature of the material. Densities are suitable for liquids, and specific weights are for solids. The density of a liquid can be measured with pycnometers or specific gravity cups, whose inner volumes are known. The density is calculated dividing the mass of the liquid by the volume. Since volumes of solids are quite difficult to measure, specific weights are appropriate for solids (ISO 3675, JIS-K2249). Specific gravity balances are usually used for the measurements. The weights in air and water of a solid are measured using balances. The difference between the weights in air and water is equivalent to the buoyancy indicating the volume of the solid. Dividing the weight of the solid by the buoyancy yields the specific weight. The density of the solid can be calculated multiplying the specific weight and the density of water. The method has a problem

40

Quality Control of Raw Materials

1143

that it cannot be applied to water-soluble materials. In this case, volumes of the materials should be measured by different methods. For liquids such as solvents, areometers can be used. They are convenient because specific weights of liquids can be measured in the containers.

40.3.3 Quality Control of Raw Materials in Storage Circumstance controls are very important for raw materials in storage. The most influential condition on material degradation is temperature. Air conditioners for warehouses, refrigerators, and freezers should be appropriately selected for the required temperature conditions of raw materials, and used. Although humidity and ultraviolet irradiation are lesser influential than temperature, protection methods have to be taken not only in storage, but also in processing of adhesive products. For adhesive production, particular amounts of raw materials are extracted by pumps from the containers such as metallic barrels, and transferred to reaction vessels. Some materials such as isocyanate resins are very sensitive to humidity, and solidify when the containers are kept open. For the materials, time limits of opening containers should be specified and kept by workers. Consuming up materials soon after the containers are opened is a good idea. Therefore, proper plans for the production of adhesives can contribute to the quality control of raw materials. Adhesives for electronic applications are nowadays processed in clean rooms. If containers for raw materials are dirty and contaminated, the cleanliness of the clean rooms may decrease. Therefore, the containers should be stored in clean conditions in order to prevent them from contaminations. More information on storage is given in ▶ Chap. 36, “Storage of Adhesives.” Education of raw material managers is another point of importance. No manager is the worst. Irresponsible managers are almost the same as no manager. In these situations, the performance of raw materials cannot be proved, and possibilities of defective products increase. Indispensable rules such as keeping materials in proper conditions of storage, and making overdue materials unusable are obeyed strictly by the managers. Disposals of overdue materials may occur due to production plans. Even in the cases, managers must not be blamed.

40.3.4 Inspection for Shipment When adhesives are shipped after the processing, users of raw materials become supplier of products. Inspection should be done to the adhesive products to prove the performances. Specification sheets should be issued for users if required. Sampling tests of industrial adhesives are rare unlike household adhesives. Tests of each production lot are usual. Inspection tests before shipments of adhesive products are as follows: adhesion strength, tackiness, hardness, viscosity, density, volume shrinkage, thermal expansion, shelf life, pot life, gelation time, water resistance, chemical resistance, weather

1144

K. Wakabayashi

resistance, flame retardance, electric conductivity, color, optical properties, amount of volatile organic compounds (VOC) emergence, etc. Some of these tests are the same as those tests for acceptance. Before shipment, strength tests of joints are very important for adhesives in terms of performance proof (see ▶ Chap. 19, “Failure Strength Tests”). In contrast, for PSAs, tackiness is important (see ▶ Chap. 15, “Pressure-Sensitive Adhesives (PSAs)”). Other important properties that should be tested are shown below.

Shelf Life Shelf life is a term during which adhesive can be stored without undesirable physical and chemical changes. The term depends on the storing condition of the adhesive. Thus, the shelf life of an adhesive should be indicated with the storage condition such as temperature, humidity, and light. Since some adhesives are sensitive to oxygen or humidity, the shelf lives of the adhesives should be stored with unopened container during a term, then the viscosities or joint strengths are experimentally measured. The procedure is denoted in test standards such as ASTM D 1337 (see ▶ Chap. 18, “Thermal Properties of Adhesives”). Pot Life (Working Life) Pot life (working life) is the time length in which an adhesive can be applied and used to join after the adhesive comes to an applicable condition. For instance, two-part epoxy adhesives gradually cure and harden after the mixing of the parts, and become too stiff to be applied after their pot lives, decreasing the bond strength. For moisture curable adhesives, the pot lives start when the adhesives are extracted from the containers. Pot lives are determined measuring the viscosities or joint strengths of adhesives with respect to time. The procedures are denoted in some standards, i.e., ISO10364, JIS K 6870, and ASTM D 1338 (see ▶ Chap. 18, “Thermal Properties of Adhesives”). Curing Shrinkage Reactive adhesives such as epoxy and acrylic resins may induce shrinkage stresses that reduce often the bonding strengths. Therefore, the volume shrinkage should be measured by standards such as JIS K6911, in which adhesives are cured in a mold and the dimension of the adhesive is measured to calculate the shrinkage ratio. Curing Rate Curing rate is important information for users of adhesives. However, ordinary adhesives cure gradually and sometimes reach to the final strength within a few days. Hence, the time to the final strength is less important than the time when de-molding or handling of the joints can be possible, because the total time of bonding processes highly depends on the de-molding or handling time. In contrast, a too short curing time, which is equivalent to a short pot life, spoils the workability. Therefore, curing rates are very important and should be measured using many specimens such as lap shear joints with respect to time.

40

Quality Control of Raw Materials

1145

Strength and Hardness Although the strength of the adhesive in bulk is not so important to be often measured, sometimes it is necessary to verify the specification. For the strength test of adhesives in bulk, dog bone specimens for tension or rectangular specimens for bending are used. In contrast, hardness of adhesives in bulk can be used as an index of curing condition, and is measured in terms of quality control. Hardness measurements include scratch tests and indentation tests, and the latter, such as the Shore test, is often utilized for adhesives in bulk. A Shore hardness tester has a small diamond half sphere equipped at the tip of a hammer, and the hammer hits the surface of a specimen. Hardness is calculated from the height of the hammer rebounding after the collision. A too soft adhesive implies under-cure or mixture of air bubbles, and a too hard adhesive means over curing. Glass Transition Temperature Grass transition temperature (Tg) is an important parameter for the quality control of adhesive formulation. The temperature is also related to the maximum usable temperature of adhesives. Adhesives are softened and become viscoelastic around the Tg. The Tg of an adhesive can be measured with differential scanning calorimetry (DSC), thermo mechanical analyzer (TMA), and dynamic mechanical analyzer (DMA). The Tg is calculated from the absorbed or emitted energy measured with DSC, deformation with TMA, and fluctuation of viscoelastic parameters with DMA, although the values from the methods are slightly different. Therefore, the definition of Tg is quite ambiguous. There is another similar temperature called softening point measured by penetration or bending tests using TMA. This is very close to Tg in value (see ▶ Chap. 18, “Thermal Properties of Adhesives”). Other Properties Nowadays, additional functions become important and the performance should be tested. For instance, since the use of Pb solder is nowadays prohibited by the restriction of the use of certain hazardous substances in electrical and electronic equipment (RoHS), conductive adhesives are used instead. For the adhesives, electric conductivity and shear resistance are vital, and their tests, besides other conventional tests, should be conducted. Thus, some adhesives developed in the future may have novel functions that conventional adhesives do not have, and new test methods and standards should be required to check the functions. From the environmental point of view, measurements of volatile organic compound (VOC) emergence and dismantlability of adhesives will become more important in the future (see ▶ Chap. 59, “Recycling and Environmental Aspects”).

40.4

Material Standards

Since qualities of raw materials influence considerably the performances of adhesive products, the materials should be tested appropriately by test standards depending on the types of materials. The material standards are discussed below.

1146

K. Wakabayashi

40.4.1 Main Components The main components of adhesives are polymers such as elastomers and synthetic resins, and monomers and oligomers for reactive adhesives. As shown in Sect. 3.2, properties that should be tested are chemical equivalent, molecular weight, viscosity, etc. In addition, solubility parameters, crystallinity, and presence of functional groups are also indispensable check points for the design of adhesives. 1. Solubility parameters: Solubility parameters are values indicating the solubility of plastic materials. Materials having similar solubility parameters can be dissolved and adhere to each other very well. Therefore, the value is very convenient to select the main component of an adhesive applied to a particular material. Solubility parameters of ordinary plastic materials may be shown in handbooks (e.g., Handbook of Adhesives edited by Skeist (1990)). Suppliers should be asked about solubility parameters of special plastics. 2. Crystallinity and presence of functional groups: Ordinary polymers except amorphous polymers have crystalline parts. Crystallinity of polymers affects the properties. Polymers of high crystallinity have high cohesive attraction. In contrast, low crystallinity leads to good contact to the adherend surfaces. Therefore, attention should be paid to the crystallinity of polymers used for adhesives. For instance, polychloroprene rubber has several types of low, medium, and high crystallinity, which are shown in the technical data provided by the suppliers. Type selection should be done carefully to design adhesives. The presence of functional groups (e.g., hydroxyl, carboxyl, amino, and epoxy groups) is also important to design adhesives, and the information is also indicated in technical data.

40.4.2 Solvents Properties that should be controlled for solvents are molecular weight, specific weight (density), boiling temperature, ignition temperature, aqueous solubility, initial boiling point, dry point, sulfur content, bromine index, residue on evaporation, moisture content, etc. Each solvent has a material standard (Table 1) and the properties are controlled by the suppliers.

40.4.3 Tackifiers Properties controlled for tackifiers are molecular weight, specific weight (density), melting temperature, softening temperature, aqueous solubility, solvent solubility, melt viscosity, burning point, etc. Molecular weights of tackifiers are measured by vapor pressure osmometry (VPO) methods. Melting temperatures and softening temperatures are measured by the method shown in a standard JIS K 2207. The Cleveland method is widely used to measure the burning point of tackifiers.

40

Quality Control of Raw Materials

1147

Table 1 Material standard for solvents Solvent n-Hexane (very pure) Mineral spirit Toluene, xylene Industrial methylated alcohol Isopropanol Methyl cellosolve Ethyl acetate n-Butyl acetate Isobutyl acetate Acetone Methyl ethyl ketone Methyl isobutyl ketone Trichloroethylene Perchloroethylene 1,1,1-Trychloroethane

ISO

ISO 5272,5280

ISO 1836 ISO 1836 ISO1090 ISO 1245 ISO 2212 ISO 2213

JIS JIS K0504 JIS K2201 JIS K2435 JIS K1505 JIS K1522 JIS K 8895 JIS K 1513 JIS K 1514 JIS K 8377 JIS K 1503 JIS K 1524 JIS K 1508 JIS K 1521 JIS K 1600

ISO International Organization for Standardization, JIS Japan Industrial Standards

Table 2 Material standard for tackifiers Solvent Dibutyl phthalate (DBP) Dioctyl phthalate (DOP) Di-isodecyl phthalate (DIDP) Tritolyl phosphate (TCP)

ISO

ISO 2520, 21, 22

JIS JIS K6752 JIS K6753 JIS K6750 JIS K6752

ISO International Organization for Standardization, JIS Japan Industrial Standards

40.4.4 Plasticizers Properties controlled for plasticizers are molecular weight, specific weight, oxidation degree, ester number, refraction index, loss on heating, oxidation degree after heating, boiling temperature, melting temperature, viscosity, etc. Material standards for plasticizers are shown in Table 2.

40.4.5 Fillers Important characteristics for fillers are low density, uniform dispersibility, workability of cured bulk including fillers, minimum decrease of properties of cured bulk (e.g., mechanical properties, resistance for water, chemicals heat, and weather) by inclusion of fillers, large maximum content, low cost, etc. Controlled properties are average grain size, specific surface area, pH, etc. Since the suppliers usually provide experimental reliable data, the users do not test again by themselves.

1148

K. Wakabayashi

40.4.6 Thickening Agents Casein, which is a type of protein included in milk, is used as a thickening agent for adhesives. There are three types of casein: acid casein, lactic casein, and rennet casein, depending on the production process. Acid casein is used for latex adhesives as thickening agents. Since casein is basically insoluble in water, alkali water solutions (e.g., with ammonia, potassium hydrate, sodium hydrate, borax, etc.) are used after water immersion to swell. Quality check of casein is conducted to measure the viscosity of water solution of 1–2%. Methylcellulose, carboxymethyl cellulose, polyvinyl alcohol, sodium polyacrylic, and polyacrylamide copolymer are also used for thickening agents. The qualities of methylcellulose and carboxymethyl cellulose are checked measuring the viscosity of water solution of 2–5%. For polyvinyl alcohol, viscosity (4% at 20  C), saponification number, and polymerization degree are indexes of the quality.

40.4.7 Pigments Color pigments are often mixed in adhesives. Since there are many types of pigments, which can affect the performance of adhesives, thorough discussion between users and suppliers of pigments is indispensable. Standards for pigments are shown in Table 3.

40.4.8 Other Raw Materials Other raw materials such as antiseptic agent, antiaging agent, antioxidizing agent, antifoam agent, and adhesion promoter are also mixed to adhesives. Tests by users of these materials are very rear. Specifications provided by suppliers should be checked and usable conditions should be kept.

Table 3 Material standard for pigments Solvent Titanium oxide (titanium white) Zinc oxide (zinc flower) Carbon black Ferric oxide (colcothar) First yellow G (Arylide yellow G) Phthalocyanine blue

ISO ISO 591

ISO 1248

ISO International Organization for Standardization, JIS Japan Industrial Standards

JIS JIS K 5116 JIS K 5102 JIS K 5107 JIS K 5109 JIS K 5217 JIS K 5241

40

Quality Control of Raw Materials

40.5

1149

Conclusion

The performance of adhesives depends on the quality of raw materials used for the adhesives. Therefore, it is easy to understand that the quality control of the raw materials is very important. As adhesive manufacturers design and provide adhesives which meet the demand of the market, raw material makers should provide appropriate products to adhesive makers, understanding and considering their needs. To keep the quality of raw materials, the specifications of materials and their testing methods should be determined and agreed by adhesive makers and raw material makers. Standards of materials and testing methods can be used to decide the specifications. All the purchase and shipping of raw materials should be done following the specifications. For instance, inspection of raw materials are carried out based on their specifications before shipping by raw materials makers, and specification sheets of each product lot are issued. For recent novel raw materials, such as functional fillers for conductivity or flame retardancy, new standards should be established. Since testing methods are varied, selection of necessary and sufficient tests is important in terms of efficiency and economy. Storage condition of raw materials is also very important to proof the performance of final products. Temperature and humidity of storages or warehouses should be within the range shown in the specification of materials. No material over the shelf life can be used to produce adhesives. First-in first-out sequence is a good manner to manage the inventory of raw materials. Education for workers who treat raw materials is also an effective investment to maintain high quality of adhesive products. Raw materials for adhesives have to be selected to meet the demands such as high performances, additional functions, and durability including environmental resistance. Traditional techniques to realize the desirable performance is to blend raw materials. Other methods, which have been introduced recently, are as follows: 1. Polymer alloying: This can be applied not only to mutually compatible materials, but also to noncompatible ones using compatibilizing agents. 2. Copolymerization, graft polymerization: They can combine immiscible materials, which leads to hybrid effects and synergies. 3. Pre-reaction: This can introduce sea-island structures in cured adhesive, and useful to reinforce adhesive layers. The technique has been applied to structural adhesives already. 4. IPN (inter penetrate network) structures: Developing of molecular chains penetrating each other can improve the resistance of cured bulk materials due to the stress relaxation. For instance, elastic adhesives have IPN structures comprising an epoxy backbone and modified silicone molecular chains. 5. Nanotechnology: This can be applied to surface modification. 6. Microcapsules. This can be used to realize novel functions such as dismantlability.

1150

K. Wakabayashi

These methods are very promising to establish new materials for futures adhesion technology. The other important requirement for ensuring the quality of raw materials is frequent and close communication between material providers and users. Information-sharing between providers and users is very crucial and indispensable for total quality control of adhesive products.

References Akita S et al (2001) Appl Phys Lett 79:1691 Crivello JV, Sasaki H (1993a) J Macromol Sci Pure Appl Chem A30:173 Crivello JV, Sasaki H (1993b) J Macromol Sci Pure Appl Chem A30:189 Deming WE (2000) Out of the crisis. The MIT Press, Cambridge (reprinted version) 88 Fan S et al (1999) Science 283:512 Iijima S (1991) Nature 354:56 Nakayama Y (2002) Ultramicroscopy 91:49 Noda N (2005) Adhes Technol Jpn 25:29 Ozawa M (1987) TQC and management, Japan Standards Association, p 116 Saito Y (2003) J Nanosci Nanotechnol 3:39 Sasaki H (2000) J Photopolym Sci Technol 13:119 Sasaki H, Crivello JV (1992) J Macromol Sci Pure Appl Chem A29:915 Sasaki H, Rudzinski JM et al (1995) J Polym Sci A 33:1807 Skeist I (1990) Handbook of adhesives, 3rd edn. Springer, New York, p 13 The Ministry of the Environment (1996) Government of Japan, Air pollution control law No. 32 Tozaki I (2006) Adhes Technol Jpn 26:52 Tutui S (2006) Adhes Technol Jpn 26:25 Yanagisawa S (2007) Adhes Technol Jpn 27:26

Processing Quality Control

41

Kosuke Haraga

Contents 41.1 41.2 41.3 41.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mathematical Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Basic Requirements for Highly Durable Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Important Points for Adhesion Process Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.4.1 Dimension Allowance of Parts Bonded Adhesively . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5 Management and Preparation of the Bonding Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5.1 Wettability Measurement of the Bonding Surface . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5.2 Surface Treatment to Improve Adhesion Adequateness . . . . . . . . . . . . . . . . . . . . . . 41.5.3 Primer Coating . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5.4 Pot Life of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5.5 Verification of Adhesive Spreading and Curing After the Bonding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5.6 Inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.5.7 Process Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.6 Additional Efforts to Ease Process Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.6.1 Reduction of Operator’s Tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.6.2 Foolproof . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.6.3 Use of Adhesives with an Easy Process Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.7 Education and Training . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.7.1 Operator’s Education and Qualification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41.7.2 Thorough Exchange of the Information About Process Changeovers . . . . . . . 41.8 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1152 1153 1155 1158 1158 1161 1161 1161 1162 1162 1163 1164 1165 1165 1165 1165 1166 1168 1168 1168 1168 1169

K. Haraga (*) Tapes and Adhesives Department, Electronic Materials Division, Electronic Materials Business Unit, Denki Kagaku Kogyo Kabushiki Kaisha, Chuo-ku, Tokyo, Japan e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_41

1151

1152

K. Haraga

Abstract

The processing quality control of adhesively bonded joints in actual production lines is discussed in this chapter. To improve the yield ratio of products, a scientific approach based on both probabilistic design and statistical treatment is indispensable. In addition, control of environmental conditions and materials is also very important. For the purpose, a realistic trial-and-error approach should be done to meet any demands or to solve problems occurring in the production. Investigation of reasons decreasing joint strength is another key point that should be done. The strength decrease occurs often due to mechanical or chemical mismatch of adherends and adhesives. The surface treatment of adherends is important too. Inspection, which can be carried out easily, is desirable in actual production lines. Process design should be done appropriately to improve the efficiency and the economy of the lines and to increase the yield ratio of products. Parallel processes should be avoided if sequential processes can be adopted. Reduction of operator’s tasks is another important issue. The concept of foolproof should be introduced in the process design.` Proper adhesive selection is vitally important, not only for good strength but also for increasing operation efficiency. New types of adhesives have been available to meet the demands to control the process easily. The education and training of operators are also an essential issue. Operators should learn standard procedures to perform their work appropriately. The necessary information should be exchanged properly.

41.1

Introduction

It is quite difficult to check visually the adequateness of adhesion bonding after the joining process. Therefore, design of bonded joints, process optimization, and its control are very important. In this chapter, some important aspects are discussed to obtain practically good adhesion in the fabrication lines of products having adhesively bonded joints. The quality control in adhesives production is not treated here. This aspect is, for example, treated in Petrie (2007). The quality control of adhesively bonded joints in fabrication lines includes control of adhesion processes, appropriate design of joints to ease the process control, and education and direction for works to conduct the manufacturing. Many efforts have been made to improve the quality stabilities (Petrie 2007; Espie 1995; Roberts 1990; DeFrayne 1983; Bandaruk 1962). Due to the invention of new types of adhesives such as acrylics whose strengths are less sensitive to processing conditions than those of conventional adhesives, the quality control can be done more easily and precisely in these days than the previous as shown by Haraga et al. (2009). Statistical approach is essential to control the quality of adhesively bonded joints in products. Both probabilistic designs of the joints and precision enhancement in each process to increase the yield ratio of useable joints are important. Statistical

41

Processing Quality Control

1153

approaches, in which the average and standard deviation of joint strengths are obtained and used for joint design or predicting the expected strength range, have been performed in several investigations (Straalen et al. 1998; Stroud et al. 2001; Hadj-Ahmed et al. 2001; Gorbatkina Yu and Ivanova-Mumjieva 2001; Seo and Lim 2005; Vallee et al. 2005a, b, 2006; Arenas et al. 2010). The idea can be applied to actual production processes in industries. Additionally, particular practical methods to improve the yield ratio of usable joints in each stage should be established.

41.2

Mathematical Background

Sample number increasing

Attribution value

Frequency (sample numbers)

Frequency (sample numbers)

Statistical methods are indispensable for ensuring the reliability of adhesively bonded joints. The mathematical background of statistical methods is very briefly explained in this section. See other textbooks (e.g., Montgomery 2008) for more details. Imagine the case in which there are many samples having their own attributions. For instance, adhesively bonded specimens can be regarded as the samples, and they have strengths as the attributions. The attributions can be often treated with numerical values. A graph, in which the number of samples within a particular range of the numerical value is plotted, as shown in Fig. 1, is called histogram. In this figure, the abscissa axis denotes the range of strength, and the axis of ordinate signifies the number of samples. The summation of the histogram is obviously equal to the total number of the samples. It is empirically known that a histogram becomes a type of symmetrical shape as shown in Fig. 1 if the sample number is large enough and its attributions are random events such as strength of materials or dimensions of machined parts, for example. If the axis of ordinate is normalized with the total sample number and the attribution range is minimized, the histogram reaches asymptotically a curve called normal distribution or Gaussian distribution as shown in Fig. 2 that is plotted with the function as follows:

Attribution value

Fig. 1 A schematic illustration of histogram and the shape change with respect to sample members’ increase

1154

K. Haraga

Probability density f(x)

f(x) =

1 x–m 1 exp – 2 s s√2π

2

P (– ∞ ≤ x ≤ ∞) =

∫

∞

1 x –m 1 exp – 2 s s √2π –∞

m – 3s

m – 2s

m –s

m

m +s

m + 2s

2

dx = 1

m + 3s

Random variables x

Fig. 2 A schematic illustration of normal distribution

  1 1 x  μ2 f ðxÞ ¼ pffiffiffiffiffi exp  2 σ σ 2π

(1)

where x, μ, and σ denote attribution values, average values, and standard deviations of the sample population that can be defined by the following Eqs. 2 and 3: 1 Xn x i¼1 i n rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 Xn σ¼ ðx  μÞ2 i¼1 i n μ¼

(2) (3)

The abscissa axis of normal distribution curves is called random variables, and its ordinate signifies probabilistic density that is equivalent to the existence probability of samples per unit range of the abscissa axis. The existence probability of samples within a particular range (a  x  b) can be calculated with the following Eq. 4. The integration of existence probability through the whole range is obviously equal to 1: Pð a  x  bÞ ¼

ðb a

f ðxÞdx ¼

ðb a

  1 1 x  μ2 pffiffiffiffiffi exp  dx 2 σ σ 2π

(4)

The existence probability of samples within a symmetrical range of  k from the average μ, which is denoted as P(μ  k  x  μ + k), can be also calculated with the following Eq. 5:

41

Processing Quality Control

Pð μ  k  x  μ þ k Þ ¼

1155

ð μþk μk

  1 1 x  μ2 pffiffiffiffiffi exp  dx 2 σ σ 2π

(5)

For instance, Pðμ  k  x  μ þ 1Þ ¼ 12 f1 þ Pðμ  k  x  μ þ kÞg , and P (μ  3σ  x  μ + 3σ) = 0.9973. Therefore, 99.73% of samples exist within the range of μ  3σ. Imagine a situation where the average strength of joint specimens is 20 MPa and the standard deviation is 3 MPa. In this case, 68.26% of the specimens are within the strength range of 17–23 MPa, 95.45% are within 14–26 MPa, and 99.73% are within 11–29 MPa, respectively. In the final case, the ratio of the samples out of the range (μ  3σ) is very small because it is 0.27%. When the reliability of adhesively bonded joints should be considered, joints too weak in strength are out of use. However, too strong joints are not problems. Therefore, P(μ  k  x  μ + 1) is more crucial than P(μ  k  x  μ + k). The relations between them can be shown as follows: 1 Pð μ  k  x  μ þ 1 Þ ¼ f1 þ Pð μ  k  x  μ þ k Þ g 2

(6)

Some problems happen when the statistical methods derived above are actually applied to the reliability evaluation of products. At first, all products cannot be inspected. The most usual method is sampling inspection. Thus, it is necessary to predict the average and standard deviation of a population from a small number of samples inspected. This process is called “statistical estimation.” For instance, strength tests of adhesively bonded specimens are adapted to this process because relatively small number of specimens such as five or ten are usually used in order to “estimate” the statistical properties of a large population. The process has been investigated very much. Unfortunately, the method is not versatile because obtaining precise information from few specimens is difficult. Enough number of samples is necessary for determining the precise statistical properties.

41.3

Basic Requirements for Highly Durable Adhesion

Highly durable adhesion has to include such capabilities as high strength, low strength scatter, and high resistance to environmental and loading conditions. To realize a highly durable adhesion, an indispensable requirement is cohesive fracture, by which fracture occurs not along the interfaces between the adhesive layer and the adherends but only in the adhesive layer. Table 1 shows the experimental results of acoustic emission (AE) measurement in the case of tensile tests using lap–shear specimens (Haraga 2008). The percentage of the ultimate load corresponding to the first AE occurrence and the total number of AE counts up to fracture are shown. This result implies that interfacial fractures happen at a lower stress than that of cohesive fractures because AE occurs in the case of interfacial fractures much earlier than in

1156

K. Haraga

Table 1 Percentage of the ultimate load corresponding to the first occurrence of acoustic emission (AE) and total counts of AE until the fracture for different failure conditions (interfacial fracture or cohesive fracture) in adhesively bonded joints

Cohesive fracture

Sample no. 1 2 3 Average 1 2 3 Average

Fig. 3 A typical relation between joint strength and cohesive fracture ratio (cohesive fracture area/total adhesion area (C/T)) where the total number of experiments (n) is 1213

Percentage of ultimate load corresponding to the first AE count 7% 8% 31% 15% 51% 76% 100% 76%

Total AE counts until fracture 25 17 117 53 19 11 1 10

n = 1,213

Strength

Fracture condition Interfacial fracture

0

20

40

60

80

100

Cohesive fracture ratio (%)

the case of a cohesive fracture. Since cyclic loads or heating cycles tend to cause stress concentrations around the edges of interfaces due to the loads applied and the thermal stresses, they cause interfacial fracture with a low strength and the joints fail under relatively low loads and short cycles (Haraga 1992). A typical relation between cohesive fracture ratio (cohesive fracture area/total adhesion area (C/T)) and joint strengths is shown in Fig. 3. In this case, the total number of experiments (n) is 1213. There is a threshold in joint strength around 40% of the C/T. Below a C/T of 40%, the number of weak joints increases. Weak joints are

41

Processing Quality Control

Allowable defective ratio (1–R) 5 /100 D = 1–1.6456•CV 1 /100 D = 1–2.3264•CV 1 /1T D = 1–3.0887•CV D = 1–3.7182•CV 1 /10T D = 1–4.2672•CV 1 /100T D = 1–4.7605•CV 1 /1M D = 1–5.2121•CV 1 /10M D = 1–5.6311•CV 1 /100M

1.0

0.8 Dispersion coefficient D

1157

0.6

0.4

CV increasing due to degradation

0.2

0.0

0

0.05

0.2 0.1 0.15 Coefficient of variation CV

0.25

0.3

Fig. 4 Relation between the coefficient of variation CV, the dispersion coefficient D, and the allowable defective ratio (1R) in strengths of adhesively bonded joints

very rare for a C/T over 40%. A highly durable adhesion can be realized if the C/T of a joint is over 40% just after its production process. The other index of adhesion durability is the coefficient of variation (CV) that is defined as the standard deviation divided by the average strength. Adhesively bonded joints show a fairly good correlation between the CV and the C/T. A low CV under 0.15 is necessary, and 0.10 is desirable for highly durable adhesion. Interfacial fracture increases the CV. A large CV condition over 0.2 gives unreliable joints and should be avoided (Haraga 2008). Figure 4 shows the CV and the dispersion coefficient (D) in a case of adhesive joints (Haraga 1992; Teramoto et al. 1993; Nakajima et al. 1997). D is defined as a factor that is multiplied with average joint strength, and this product gives the maximum strength of defectives as shown in Fig. 5, where R is yield ratio of useable joints and 1R is allowable defective ratio. For instance, when 1R is 1/100,000 and CV is 0.15, D can be determined as 0.36 with Fig. 4. This case indicates the condition of a sample having a strength lower than 36% of the average joint strength among 100,000 samples. If CV is 0.2 and 1R is 1/100,000, D becomes 0.15, and this case is obviously not a reliable design. A basic and important rule about reliable design of adhesively bonded joints is the management of C/T under 40% and CV under 0.15.

1158

K. Haraga

Average strength

Small CV

Allowable defective ratio (1–R) Large CV

D-Fm

D-Fm

Fm

Strength

Fig. 5 A schematic illustration of the relation between the average bonding strength Fμ, the allowable defective ratio 1R, and the dispersion coefficient D

41.4

Important Points for Adhesion Process Control

41.4.1 Dimension Allowance of Parts Bonded Adhesively Substrate Deformation A serious error in the adhesion process is to ignore the dimension allowance of the substrate. Some members are often machined beyond their specified dimension allowance indicated in the design drawing. As shown in Fig. 6, a required flat surface Fig. 6a can be deformed as Fig. 6b or c. The hat cross-sectional member shown in Fig. 7a can be finished in the cross section as Fig. 7b or c due to bad bending forming or can be spring-backed as Fig. 7d or e. Such deformations cause strength deviation due to nonuniform adhesive thickness or bad adhesive spreading or defective goods due to the tilt of the bonded parts. Applied Pressure to Parts in the Bonding Process Pressure has to be applied to parts during the curing process of adhesives. In this process, excess pressure applied to the members shown in Fig. 7a can induce undesirable deformation as shown in Fig. 7b. Conversely, in the case of Fig. 7c, the member can deform by excess pressure and take the shape shown in Fig. 7a. The deflected members shown in Fig. 7c, d are flattened by pressurization. In such cases, a spring-back force is induced when the applied pressure is released after the

41

Processing Quality Control

a

1159

b

Flat: normal

c

Concave: defect

Convex: defect

Fig. 6 Undesirable shape of the bonding surface due to machining

b

a

Bonding surface

d

Bonding surface

Bonding surface

Bonding surface

c

Bonding surface

Bonding surface

e

Fig. 7 Undesirable deformation of hat members due to plastic forming (a) no deformation, (b) concave deformation in the flange, (c) convex deformation in the flange, (d) convex deflection, and (e) concave deflection

bonding process because residual stress occurs due to the permanent deformation. The spring-back forces act as creep loads to adhesive joints. Creep rupture often happens not only in the case of weak adhesives such as pressure-sensitive adhesives at ambient temperature but also in the case of strong adhesives if they are subjected to high temperatures during a paint-baking process, for instance.

1160

K. Haraga

Environmental Condition Control Temperature and humidity have to be controlled appropriately during the cure reaction of adhesives. In summer, ambient curing adhesives need a quick operation in bonding process because their gelation time is short. In winter, since gelation time is longer, the time of pressure application is also longer. To avoid the variance of operation and keep the sequence constant, temperature control using air conditioners is highly recommended. Humidity control is also very important, especially when using humidity-cured adhesives or sealants, such as instantaneous adhesives or silicone adhesives. Sometimes, humidity is not “desirable” but rather “indispensable,” and it is necessary to increase humidity when it is very low. On the contrary, in the case of two-part urethane adhesives, the strength decreases due to the foam formation in the adhesive layer, which tends to happen especially in hot–wet conditions, if the operation time from adhesive application to bonding is long, as shown in Fig. 8. Therefore, dehumidification is recommended in such cases. In addition, since polyol components in urethane adhesives tend to absorb water very much, seal-up of adhesive container is also very important. Dew condensation on the surface of substrates before bonding is another problem. The dews can corrode the surface before bonding and inhibit adhesive tack. Just

300

250

Lap shear strength (kgf/cm2)

Fig. 8 Strength variation of joints bonded with a two-part ambient curing urethane adhesive with respect to the foam formation conditions, where open time is the time from adhesive application to lapping adherends, in which the substrates are kept in ambient condition

200

150

100

50

0

Open time : 5 min

Nothing Few Scarce Middle Foam formation condition

Much

41

Processing Quality Control

1161

after the start of room heating in winter, cold members may cause dews on their surfaces. Therefore, humidity control is also necessary in this situation.

41.5

Management and Preparation of the Bonding Surface

41.5.1 Wettability Measurement of the Bonding Surface The surface condition of substrates before bonding depends on the production lot, the shelf condition, and the storage period. Even if the surface is degreased with solvents or chemicals before bonding, it is not necessarily suitable for adhesion. As a method to evaluate the adhesion adequateness easily, wettability tests can be applied using standard liquids commercially available. It is reasonable to measure the wettability of substrates by random inspection once in each morning and afternoon or every time the production lot changes. Experimentally proved minimum value of wettability for proper adhesion is 36–38 mN/m, where the wettability is denoted as surface tension. The use of parts having poor wettability is never accepted. Action to improve the wettability has to be carried out.

41.5.2 Surface Treatment to Improve Adhesion Adequateness

55

100

50

80

Strength (kgf)

Wettabiliy (dyne/cm)

It is frequent that low wettability surfaces influence badly the production line of the goods. It is desirable to introduce surface treatment processes to low wettability materials such as thermoplastics. For instance, short wavelength UV irradiation with low-pressure mercury lamps and ambient pressure plasma irradiation are popular for the purpose. Even shot-term irradiation can improve the wettability of surface very much and the bond strength increases, as shown in Fig. 9.

45 40 35

60 40

PPS/irradiation distance: 5 cm PBT/irradiation distance: 5 cm

30

PPS/irradiation distance: 5 cm

20

PBT/irradiation distance: 5 cm

0 0

1

2 3 4 10 UV irradiation time (min)

0

1

2 3 4 9 UV irradiation time (min)

10

Fig. 9 Variations of wettability and strength of polyphenylene sulfide (PPS) and polybutylene terephthalate (PBT) with respect to UV irradiation

1162

K. Haraga

41.5.3 Primer Coating To improve adhesion strength, primer coating can be applied on the surface of members. Generally speaking, a thin primer layer is advantageous in adhesion compatibility rather than a thick layer, as shown in Fig. 10. Thicker primer layers decrease bond strength. However, too thin primer layers lead to a problem of spreading defect. Applying primer coating properly with a spray or brush is quite difficult and leads to considerable variation in bond strength. Therefore, the use of low-density solution primers is recommended because a light amount of primer is left on the surface even though an excess of solution is applied.

41.5.4 Pot Life of Adhesives Quick-curing adhesives have short pot lives, which are often shorter than the total time of the bonding process including adhesive mixture, application, joining, and pressurization. Therefore, when the total time is too long, the joint strength decreases, as shown in Fig. 11. Bonding time control is very important in such cases. The control can be done, for instance, using a timer having a ring or light signal capability. The timer has to be turned on when adhesive mixing starts. Operators should finish their work by the ring or light signals indicating the end of time tolerance.

Mixed fracture

Adhesion strength

Cohesive fracture Interfacial fracture Interfacial fracture

0

1

2 3 Amount of applied primer coating

4

Fig. 10 Relation between adhesion strength and applied amount of primer coating

Processing Quality Control

Fig. 11 A typical strength variation of lap–shear joints bonded with a two-part ambient curing modified acrylic adhesive with respect to time from the mixture to the application of the adhesive

1163

120

100

Lap shear strength retention %

41

80

60

40

20

0

Fig. 12 Verification of adhesive spreading and curing using inspection holes

0

1 2 3 4 5 6 7 8 9 Time from mixture to application (min)

10

Adhesive spew Adherend 2

Inspection hole

Adherend 1

41.5.5 Verification of Adhesive Spreading and Curing After the Bonding Process It is often requested but not easy to verify the spreading and curing conditions of adhesives in joints after the bonding process. An usual method is to check the amount of spew and its hardness. Another one, which is more sophisticated, is the inspection hole method shown in Fig. 12. In this method, a hole is machined in one adherend to confirm the amount of adhesive and its curing condition. If

1164

K. Haraga

enough amount of adhesive is applied and is spread properly along the surface, a spew can be found from the hole, and the curing condition of the adhesive can be checked by pushing the spew. This method is available even when side spews do not form.

41.5.6 Inspection The performance of structures bonded adhesively is assessed with strength tests of samples selected randomly or dummy samples. If testing machines are not available in factories, alternative test methods that should be easy to carry out are necessary. Unfortunately, such cases are not rare because testing machines are expensive. Figure 13 shows an example: a simple peel test using a manual testing machine of reasonable price, consisting of a roller and a lever (Haraga et al. 2009). The curing situation of adhesive and the C/T are used as the criterion of the test, and C/T should be greater than 40% to pass the inspection. When testing machines are available, the strength of samples can be used as the criterion.

Fig. 13 Manual testing machine consisting of a peel roller and a handle lever for peel tests of specimens

41

Processing Quality Control

1165

41.5.7 Process Design Optimum Process Layout of Automatic Devices and Manual Operation Although a fully automatic bonding process including adhesive application and joining of adherends is possible, it is hard to check visually the mixture condition, the applied amount of adhesive, and its application patterns. Imaging technologies are, of course, possible to use recently. However, there are several problems in terms of cost and accuracy so far. A combined process of automation and manual work is very efficient for such situations. For instance, adherend joining can remain a manual operation to check the adhesive condition mentioned above, which is helpful to avoid defects. Adoption of a Sequential Process An advantage of manual operation is the availability of a sequential process from adhesive mixture to pressurization. This whole process can be carried out continuously by one person. On the other hand, a parallel process is dangerous because the pot life of the adhesive applied to many queuing parts can be attained if the production line stops before the joining process. Since automation needs a parallel process, a waiting system for the adhesive application, which is activated when the following processes consume too much time, has to be installed.

41.6

Additional Efforts to Ease Process Quality Control

41.6.1 Reduction of Operator’s Tasks The quality of products depends not only on the process control but also on the operator’s attentiveness. Therefore, the reduction of operator’s tasks is very important. Too excess loads are not acceptable. The process should be designed to ease manual works. Refinement of product design is also effective to reduce difficulties in adhesive joining.

41.6.2 Foolproof Mistakes done by operators are another problem. For instance, the adhesive bonding of different parts having similar configurations can cause mistakes. Errors about bonded sides, positions, or alignment of parts must be avoided too. For such cases, it is a good idea to make bumps and dimples (or holes) to facilitate the joining, as shown in Fig. 14. To control the position and the pattern of the combination, side matching, directional alignment, and positioning of parts can be conducted easily.

1166

K. Haraga

b

a Bonded parts

Adhesive layer

Dimples or holes

Bumps

Fig. 14 Foolproof method to prevent operators’ mistakes of bonding sides, positions, or alignment of parts using bumps or dimples (or holes); (a) without this method, (b) with this method

41.6.3 Use of Adhesives with an Easy Process Control Use of Oil Accommodating Adhesives Surface cleaning of adherends is vitally important for good adhesion, and it influences the performance of the joints. Recently, oil-accommodating adhesives, which can be used directly for oily surface, are available. Two-part acrylic adhesives (SGA) and some types of one-part thermosetting adhesives have this capability. Surface cleaning processes can be omitted, and scattering of bond strength can be reduced using this type of adhesives. The use of oil-accommodating adhesives has the advantage that steel parts can be stored in oiled condition and degreased before bonding is no longer necessary. The oil prohibits the stains on the surface of the steel parts that can reduce the bond strength. The whole process is also simplified very much. Use of Adhesives Tolerant to Wide-Range Mixture Ratios Two-part ambient-setting acrylic adhesives (SGA) have wider permissible ranges in terms of mixture ratio than two-part epoxy adhesives, as shown in Fig. 15 (Denki Kagaku Kogyo Co., Ltd 2009). Even if the mixture varies, or the adhesive parts are applied individually on each surface, the bonding strength is not so much influenced because of radical polymerization. The use of adhesives insensitive to mixture ratio is convenient to simplify the control of the mixture process. Use of Visually Checkable Adhesives Mixture ratios of two-part adhesives can be visualized to assign the proper colors to the adhesive parts, as shown in Fig. 15. In actual production lines, sample color tables are helpful for operators to check the mixture ratio. Innovative adhesives, whose color can be changed with respect to cure progress, are recently available. Figure 16 shows the color change of a two-part ambientsetting acrylic adhesive: NS700 series (Denki Kagaku Kogyo Co., Ltd., Japan) (Tomioka 2010). The colors of the two parts of the adhesive are white and black before mixing. The adhesive becomes green after mixing the two parts. The color

41

Processing Quality Control

1167

Lap shear strength (MPa)

35 Strong SGA

30 25

Weak SGA

20 15 10 5 0

Epoxy adhesive Adherends: SPCC 2/8

3/7

4/6

5/5 Mixture ratio

6/4

7/3

8/2

Fig. 15 Relation between mixture ratio of adhesive parts, lap-shear strength, and variation of adhesive colors (SGA second-generation acrylic adhesive, SPCC cold rolled steel sheet)

Fig. 16 Colors change depending on the polymerization time in the case of a two-part ambient curing modified acrylic adhesive: NS700 series (Denki Kagaku Kogyo Co., Ltd., Japan)

A part White

B part Black

+

Mixture

Within pot-life

Green

After pot-life

Brown

changes to brown when gelatification starts and to dark brown after full curing. It is possible for operators to check visually the mixture and the curing conditions of adhesives with this method.

Use of Adhesives Including Spacers The adhesives shown in Figs. 15 and 16 include plastic fillers as spacers to keep the bond-line thickness constant. The use of such adhesives is very helpful to control the thickness without particular attentiveness.

1168

41.7

K. Haraga

Education and Training

41.7.1 Operator’s Education and Qualification Appropriate education for operators is indispensable for a work of good quality in adhesion processes. At first, a process manual book has to be written. Apparatus, materials including adhesives and solvents, procedures, and inspection methods have to be indicated comprehensibly using pictures or photos. Not only the standard working methods and conditions but also the maximum and minimum values of permissible allowance for each condition must be indicated in the procedure description. Determination criteria to pass the inspection should be indicated with specific samples. Prohibited actions have to be indicated explicitly. Lectures and training for operators should be given using the process manual book. The reason for minimum and maximum criteria must be explained appropriately so that operators can understand them entirely, and there should be no “black box” in the processes. For operators providing adhesion works at vitally important sites, theory, practice, and examinations are mandatory. The operators passing the examinations can be approved as qualified operators. Only the qualified operators have to be appointed to carry out adhesion processes at important sites.

41.7.2 Thorough Exchange of the Information About Process Changeovers Processing procedures, methods, and conditions must not be altered by each operator. If any processes shown in the manual book need to be changed to improve the performance, the operator has to inform the manager in advance. The suggestion must be investigated by the manager with the process designer. If the suggestion is worth to be employed, the process can be changed after the authorization of the process and the revision of the process manual book. This rule is indispensable, and all operators must be obligated to comply it.

41.8

Conclusion

For the processing quality control of adhesively bonded joints, scientific approach based on both probabilistic design and statistical treatments is necessary. For decreasing the ratio of defective joints, the coefficient of variation (CV) of the samples should be kept small. The cohesive fracture ratio (C/T) of fractured joint surfaces is vitally important to keep CV small. Since the cohesive fracture ratio decreases often due to mechanical or chemical mismatch of adherends and adhesives, the surface treatment of adherends is important. Process design should be done appropriately to improve the efficiency and the economy of the lines and to increase the yield ratio of products. The use of sequential

41

Processing Quality Control

1169

processes, reduction of operator’s tasks, and foolproof for workers are also very important and should be introduced in the production process. The condition of adherend members, adhesives, and process environments are changing continuously with time. Care should be taken to the change and establish flexible adhesion processes and procedures that can meet the change. Therefore, adhesion process should have high adaptivity to the conditional change. The process design should be done as such. The methods shown in this chapter are just typical examples, and other particular efforts must be necessary for each practical process. Reexamination of actual processes in terms of increasing the yield ratio of products having adhesively bonded joints is also essential. The education and training of operators are effective investments to avoid making defectives and stopping the production line. Inspection of goods is not only the final barrier of shipping defective products but also the first stage of quality control actions.

References Arenas JM, Narbon JJ et al (2010) Int J Adhes Adhes 160:165 Bandaruk W (1962) J Appl Polym Sci 217:220 DeFrayne GO (1983) Adhesive specification and quality control. In: Schneberger GL (ed) Adhesives in manufacturing. Marcel Dekker, New York Denki Kagaku Kogyo Co., Ltd. (2009) HARDLOC brochure Espie AW (1995) Int J Adhes Adhes 81:85 Gorbatkina Yu A, Ivanova-Mumjieva VG (2001) Int J Adhes Adhes 41:48 Hadj-Ahmed R, Foret G, Ehrlacher A (2001) Mech Mater 77:84 Haraga K (1992) Adhes Technol 28:33 Haraga K (2008) Kagaku Kogyo 271:277 Haraga K, Ganryu Y et al (2009) Mitsubishi Denki Giho 19:23 Montgomery DC (2008) Introduction to statistical quality control, 6th edn. Wiley, Hoboken Nakajima Y, Taguchi K et al (1997) Eng Mats 87:104 Petrie EM (2007) Handbook of adhesives and sealants. McGraw-Hill, New York, pp 203–226 Roberts RW (1990) Processing quality control. In: Adhesives and sealants, vol 3. Engineered material handbook. ASM International, pp 735–742 Seo DW, Lim JK (2005) Compos Sci Technol 1421:1427 Straalen IJJ, Wardenier J et al (1998) Int J Adhes Adhes 41:49 Stroud WJ, Krishnamurthy T et al (2001) AIAA2001– 1238 1:12 Teramoto K, Okajima T et al (1993) J Adhes Soc Jpn 37:44 Tomioka T (2010) Nikkei Monozukuri 59:67 Vallee T, Correia JR et al (2005a) Compos Sci Technol 331:340 Vallee T, Riberio J et al (2005b) Proc Int Symp Bond Behav FRP Struc 149:155 Vallee T, Correia JR et al (2006) Compos Sci Technol 1915:1930

Nondestructive Testing

42

Robert D. Adams

Contents 42.1 42.2 42.3 42.4

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Nature of Defects in a Joint . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tests Carried Out Before Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Testing After Bonding and During Service . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.1 Ultrasonics (Pulsed) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.2 Sonic and Ultrasonic Vibrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.3 Spectroscopic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.4 Other Acoustic Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.5 Acoustic Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.6 Thermal Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.7 Optical Holography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.4.8 Visual . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.5 Future Developments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1172 1175 1178 1180 1180 1184 1186 1187 1188 1188 1190 1190 1190 1191 1191

Abstract

The objective of any system of nondestructively examining an adhesive joint is to obtain a direct correlation between the strength of the joint (howsoever defined) and some mechanical, physical, or chemical parameter which can readily be measured without causing damage to the joint. Faults may be defined as anything which could adversely affect the short- or long-term strength of a joint. There are two basic areas for examination in properly made joints, the cohesive strength of

R. D. Adams (*) Department of Mechanical Engineering, University of Bristol, Bristol, UK Department of Engineering Science, University of Oxford, Oxford, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_42

1171

1172

R. D. Adams

the polymeric adhesive and the adhesive strength of the bond between the polymer and the substrate. In addition, voids, disbonds, and porosities create an additional issue for inspection. During the production phase, and also in service with critical structures, it is essential to use nondestructive tests to assess the quality and fitness for purpose of the product. The nondestructive test will not measure strength directly but will measure a parameter which can be correlated to strength. It is, therefore, essential that a suitable nondestructive test is chosen and that its results are correctly interpreted. Typical defects found in adhesive joints are described and an indication given of their significance. The limits and likely success of current physical nondestructive tests will be described and future trends outlined. It is shown that a variety of techniques are available for disbond detection, ultrasonics, and different types of bond tester being the most commonly used. These techniques are very time-consuming, especially if large bond areas are to be tested. Monitoring interfacial properties is much more difficult, and there is currently no reliable test after the joint is made although there are some indications of a possible way forward using high-powered lasers.

42.1

Introduction

Adhesive bonding offers the designer many advantages, especially in being able to join dissimilar materials without bolts, rivets, etc. (Adams et al. 1997). However, despite its potential advantages, the use of adhesive bonding in safety critical structures has been limited by a lack of adequate nondestructive testing (NDT) procedures: Without such procedures, the reliability of a structure cannot be guaranteed. When using advanced composites, adhesive bonding is often preferred because these materials have low transverse tensile and local bearing strength. The problem comes when it is found to be necessary to define how strong the joint is. While there are certain situations when it is possible to be reasonably confident of the initial strength of a joint if it has been carefully made by a well-established process and there is a record of previous destructive tests, this cannot be guaranteed for the lifetime of the joint. If we add to this the possibility of joints being made incorrectly, we quickly come to the conclusion that there is a need for a technology to determine the strength of a joint without breaking it. This is called nondestructive testing, evaluation, or inspection, usually abbreviated to NDT, NDE, or NDI. Such testing will usually be performed after manufacture or at stages during manufacture. However, in more stringent applications, inspection during service may also be required. Ideally, a nondestructive test would predict the strength of the bonded joint and its susceptibility to environmental attack. This is very difficult to achieve, partly because a direct measurement of strength cannot be nondestructive, so it is necessary to correlate strength with other properties, such as bond area, interfacial stiffness, etc. Also, the stress distribution in a typical adhesive joint is far from uniform, and so the strength is much more sensitive to the integrity of some areas of the joint than to

42

Nondestructive Testing

Fig. 1 A simplified schematic diagram showing the various layers in an adhesive joint

1173

1

Adherend

2 3 4 3 2

Adhesive

1

Adherend

Typical thicknesses 1 Adherend

0.5 + mm

2 Surface Oxide

4 – 300 × 10–6 mm

3 Primer

0.02 – 0.1 mm

4 Adhesive layer

0.1 – 0.5 mm

others. Measuring bond area, stiffness, and so on will not necessarily give good correlations with strength. Changes in these properties do, however, give an indication that a joint may be defective. In order to understand the nondestructive evaluation of the strength of an adhesive joint, we need to understand the mechanics, physics, and chemistry which are involved in the production and stressing of such a joint. The details of these factors are given elsewhere in this book, so only a brief description will be given here. A bonded joint generally consists of two adherends or substrates, connected by an adhesive layer. However, it is vital to understand that there are also two surfaces and that adjacent to these is an interphase region such that the adhesive properties do not change abruptly to those of the adherend. These different zones are shown schematically in Fig. 1. The most commonly used form of adhesively bonded joint is in tensile lap shear, as shown schematically in Fig. 2. Many theories have been used to analyze the stresses in bonded joints (see ▶ Chaps. 24, “Analytical Approach,” and ▶ 25, “Numerical Approach: Finite Element Analysis” in this book), but the essence is shown in Figs. 3 and 4. Here, it can be seen that the shear and peel (transverse tensile) stresses are a maximum at the joint ends, while the middle of the joint generally has low stresses. Failure is therefore going to begin at these joint ends and then propagate toward the interior of the joint until total failure occurs. Another important class of adhesively bonded components is in repair patches, particularly for aircraft. It is often cost-effective and convenient to apply a patch rather than to replace an aircraft component, particularly if the need to repair is remote from the source of the component (Pyles 2003; Baker et al. 2009). Since the repair is often

1174

R. D. Adams

P t b

t l P

Fig. 2 Schematic diagram of a single-lap joint bonding adherends of thickness t and width b with an overlap l

Fig. 3 Variation of shear stress with distance along the overlap according to Volkersen’s theory

No load

Loaded

P

P

τ

τav

P τav = — bl

1

Fig. 4 Variation of transverse tensile stress (peel) with distance along the overlap according to Goland and Reissner’s theory

σ

made in non-factory conditions, it is important to be able to carry out some form of nondestructive inspection before releasing the vehicle back into service. There exists a wide variety of nondestructive testing techniques for interrogating structures. These range from well-known methods such as ultrasonics and X-rays to a variety of specialist methods such as shearography, tapping, and thermal. However, it must be stressed that many of the techniques used with metallic structures are quite unsuitable for use with bonded joints. Various authors have reviewed the literature on nondestructive testing of bonded joints. The reader is referred to Hagemaier (1985), Guyott et al. (1986), Adams et al. (1987), the special edition of NDT International which was edited by Adams and Cawley (1988), and recently by

42

Nondestructive Testing

1175

Cawley (2005). Crane and Dillingham (2008) have provided an excellent review of adhesive bonding with composite materials. The objective of this chapter is to describe the mechanics and physics of the main NDE methods for adhesively bonded joints and to indicate what is possible and not possible.

42.2

The Nature of Defects in a Joint

It is instructive at this stage to examine the types of defect which may occur in bonded structures. First, let us consider lap joints in which the bonded area is a few square centimeters or more. Figure 5 illustrates many of the possible defects. Porosity is caused by volatiles and entrained air in the adhesive. The voids may also be due to products of the chemical reaction[s] involved in curing the adhesive. Porosity is therefore present in most bond lines to some extent. Adhesive cracks are due to problems with curing (cure and/or thermal shrinkage) or to large applied stresses, either one-off or repeated (fatigue). Local areas of poor cure are due to incorrect mixing of the adhesive system. Larger areas, possibly extending through the whole bond line, are either due to incorrect mixing, incorrect formulation, or insufficient thermal exposure. Sometimes, poor cure is self-correcting with time in that the chemical reaction continues, albeit slowly. Voids are often due to air becoming trapped by the pattern of laying the adhesive or to insufficient adhesive being applied. Voids can also be caused by relative movement of the adherends

Adherend Disbond or zero - volume unbond

unbond

Cracks

Adhesive Void Porosity

Poor cure

Adherend

Fig. 5 An illustration of the various defects which might be present in an adhesive bond line

1176

R. D. Adams

Fig. 6 Void in adhesive bond line most probably caused by moving the joint before the adhesive had reached a sufficient state of cure

during the curing process. Large voids cannot be caused by volatiles, unless something is very wrong with the adhesive system. Figure 6 shows what can happen in a tubular joint when the adherends are moved before the adhesive has reached a stage in its cure cycle when it has hardened sufficiently. The voids offer a path for the ingress of water or other fluids and fatigue crack initiation (see ▶ Chap. 33, “Fatigue Load Conditions” in this book). Surface unbonds are an alternative form of void, often caused when adhesive is applied to one adherend only and unevenly. The most dangerous form of these can occur where the adhesive and adherend are in contact, but there is no significant bond strength between them. Such defects may be caused by poor surface preparation, failure to remove the manufacturer’s backing film completely, a loose substrate, and so on. Such bonds are called zero-volume unbonds or kissing bonds. While the latter term is commonly used today, it implies there is a bond, whereas, in fact, there may be no bond at all, or only a very weak one. Jeenjitkaew et al. (2010) and Marty et al. (2004) have given an excellent description of the causes and morphology of kissing bonds. Galy et al. (2017) and Jeenjitkaew and Guild (2017) feature among various authors who have tried to make realistic kissing bonds and then to devise methods to test them. The main problem is to be able to make these joints so that they are representative of true structures. Only then can the process of testing them move forward. The presence of defects in lap joints is more indicative of poor joint manufacture than of impending failure, especially for short-term loading. Over the long term, defects may allow faster ingress of water or aggressive substances or provide sites for fatigue crack nucleation. It still, therefore, remains necessary to look for these defects. Karachalios et al. (2013) have shown that the strength of lap joints may or may not be affected significantly by large voids in the interior of a lap joint. This is because fracture initiation in the adhesive occurs due to the failure strains in many

42

Nondestructive Testing

1177

35 Hard Steel Gauge Steel Mild Steel

30

Failure load (kN)

25 20 15 Total Bondable area for 25 mm overlap

10 5 0 0

100

200

300

400

500

600

700

Defective area (mm2)

Fig. 7 Variation of failure load with defective area for single-lap joints bonded with AV119 adhesive showing how the lap-shear strength is a function of adherend type and remaining bond area (From Karachalios et al. 2013)

Fig. 8 A sketch of honeycomb bonded to a skin and showing the fillet (From Adams et al. 1997)

adhesives being much less than the yield strains of structural steel or aluminum. However, if high-strength alloys are used with modern rubber-toughened epoxies, the lap joint strength is more or less proportional to the bonded area. This is shown in Fig. 7. The other major form of joint used with structural adhesives is the T joint used in bonding honeycomb to the skins as shown in Fig. 8. The skin and core are bonded by a large number of these lightly stressed joints. It is essential that the adhesive forms a generous fillet, as shown in Fig. 9a, and not the apparently more economical but weaker joint shown in Fig. 9b. Some years ago, the author tried to stiffen an

1178

R. D. Adams

Fig. 9 Good and bad bonds between honeycomb and skin (From Adams et al. 1997)

a Skin Adhesive

Honeycomb Good

b Skin Adhesive

Honeycomb Bad

aluminum alloy honeycomb core by filling with a polyurethane foam. This was then scraped down to the level of the honeycomb and the bond made. Unfortunately, the foam prevented the formation of the fillet, and the joint failed at a low load. With bonded honeycomb structure, the major defects consist of a lack of attachment between the core and the skin. This may be due to several causes such as locally crushed honeycomb (Fig. 10a), skin defects (Fig. 10b), or lack of adhesive (Fig. 10c). In themselves, none of these defects may reduce the short-term joint strength. However, as for the lap joint, they may show poor preparation and may provide sites for fatigue crack propagation.

42.3

Tests Carried Out Before Bonding

In order to make a successful adhesive joint, it is vitally important that the surfaces to be bonded should be properly prepared. Usually, this means clean, chemically active, and sufficiently strong. We therefore need surface inspection tests to ensure that the surfaces do not contain excessive amounts of water vapor, hydrocarbons, or other contaminants. Today, there exist many powerful laboratory tools for assessing surface properties. However, these are rarely used in manufacturing environments since they are slow and expensive. A simple test involves the wettability of the surface, by measuring the contact angle. A full discussion of surface preparation and wettability is given in Part B in this book. If the surface is clean, it is readily wetted, and a drop of water will spread

42

Nondestructive Testing

Fig. 10 Possible defects with honeycomb core and skin system (From Adams et al. 1997)

1179

a Disbond-core damage

b

c

Disbond-skin imperfection

Disbond-lack of adhesive

over a large area. However, if it is contaminated, the water will remain as droplets. A useful procedure with carbon-reinforced or glass fiber-reinforced plastics is to abrade the surface with a fine “wet and dry” carborundum paper until a thin film of water can be spread over it. A more quantifiable test involves measuring the spread of a liquid drop of constant volume through a transparent gauge placed over the drop or by looking at the “contact angle” between the drop and the surface. Commercial products are also available which will indicate the suitability of the surface for bonding by showing as colored. These self-indicating primers also contain substances to enhance the bond between the surface and the adhesive. The main problem is the uncertain life expectancy and stability of these primers. The Fokker Contamination Tester, described by Bijlmmer (1978), uses an oscillating probe to measure the electron emission energy. This varies greatly with the degree of surface contamination and can even be used to detect residues from alkaline cleaning operations. None of these methods is totally satisfactory, and the best means of ensuring that a “good” surface exists prior to bonding is carefully to control the processes leading to its preparation, having previously established “good practice.” Maintaining good control over the manufacturing process will increase the probability of achieving a defect-free joint, and this is particularly important in the case of adhesive strength for which no satisfactory nondestructive testing method currently exists (Schliekelmann 1972; Kim and Sutliff 1978).

1180

R. D. Adams

Crane and Dillingham (2008) claim that the tape peel test (in which a PSA tape is pressed on to a surface and then peeled off) offers an approach that is potentially inexpensive which could be implemented in a fast-paced production environment. Unfortunately, there is insufficient data at this time on the wide range of possible contaminants and the possible levels of concentration for this approach to be applied without further testing. However, it does offer an approach to providing the NDE engineer an important tool to access the adhesion potential of a composite surface before the consolidation of a component.

42.4

Testing After Bonding and During Service

By far the majority of nondestructive testing (NDT) techniques associated with adhesive bonds are applied after the joint has been made and are usually carried out at the manufacturing location or during service. Most techniques are void detectors, and, although it is claimed that the cohesive strength of the adhesive is being assessed, this is unlikely to be the case.

42.4.1 Ultrasonics (Pulsed) The monitoring of ultrasonic echoes forms one of the most widely used methods of nondestructive testing for bonded joints and composites. The method is commonly used for the detection of disbonds, voids, and porosity. An incident pulse of ultrasound will be reflected and transmitted (assuming normal incidence and hence no refraction) at each interface of the joint. The amplitudes of the reflected (R) and transmitted (T ) pulses for a unit amplitude incident signal are dependent on the acoustic impedances of the materials on either side of the interface and are given by R¼

Z2  Z1 Z1 þ Z2

(1)

T¼

2Z 2 Z1 þ Z2

(2)

where Z is the acoustic impedance given by ρ c, ρ is the density, and c is the speed of sound in the material under examination. Subscripts 1 and 2 denote the materials on the incident and remote sides of the interface, respectively. If a defect is assumed to contain air or any other low-density substance, it will have a very low acoustic impedance relative to the adhesive or adherend. A pulse incident at the defect is almost totally reflected, leaving negligible energy to be transmitted through the defect. Measurement of the reflected or transmitted energy may therefore be used to indicate the presence of a defect. The simple ray model of a pulse of ultrasound being specularly reflected at a defect is valid when the defect

42

Nondestructive Testing

Fig. 11 Schematic of ultrasonic pulse in throughtransmission mode (From Adams et al. 1997)

1181

Sender

Wavetrains

Receiver

diameter is about one wavelength or larger. Porosity typically comprises pores which are orders of magnitude smaller than the wavelength. Therefore, porosity does not give simple specular reflection but scatters energy in all directions, thus attenuating the propagating pulse. Various techniques based on ultrasonics are used in connection with adhesive joints. Essentially, a small piezoelectric crystal is pulsed repetitively, causing it to radiate high-frequency sound waves at the natural frequency of the crystal. Normally, these brief wave trains (about five cycles long, at 1–20 MHz and repeated say 1,000 times/s) traverse the joint and can be detected, either by using the same crystal or by using a separate receiver (Fig. 11). By interrogating the change which has taken place in the wave train, it is possible to deduce various characteristics of the structure through which it has traveled. The probes have to be correctly aligned and care must be taken so that the pulse will not reflected or diffracted by curvatures in the surface. If the adhesive contains imperfections such as roughness, cracks, or porosities, the waves will be scattered, and less will be detected by the receiver. If there is a void, the wave can be partially or totally reflected, again reducing what is transmitted. Fine detail can be revealed, if at all, only by using focused or otherwise concentrated “beams” and intelligent analysis. In ultrasonic inspection, there are two basic techniques which are commonly used, through transmission and pulse echo. The through-transmission technique monitors the amplitude of an ultrasonic pulse transmitted through the joint. The amplitude can be reduced effectively to zero when a disbond larger than the beam width is present, smaller amplitude reductions being seen in the case of porosities, microcrackings, disbonds, and voids whose plan area is smaller than the beam width.

1182

R. D. Adams

Fig. 12 Water-jet probe (From Guyott et al. 1986)

Water supply

Input/Output to transmitter and receiver Ultrasonic transducer

Annular gap for water

Water jet Adhesive joint

The technique uses separate transmitting and receiving transducers, positioned on either side of the structure to be tested. Alignment of one transducer with the other is vitally important and can present difficulties when large components are tested. Alignment of the transducer axis perpendicular to the surface to be tested, however, is not as critical as with other techniques. Through transmission is often used for the production inspection of large structures such as aircraft fuselage and wing sections, coupling generally being achieved with water jets, such as shown schematically in Fig. 12. The method is also particularly suited to the inspection of honeycomb structures. Using a pulse-echo technique (see below), only the bonding of the top face to the core can be tested reliably, whereas using through transmission, both top and bottom bonds between skins and core can be inspected in a single test. However, through transmission is not generally suitable for in-service inspection since access to both sides of the structure is often difficult or impossible and the attenuation and reflections even in “good” structure can be sufficiently large to render the technique of little value. Also, while it gives an indication that a defect exists, no information about its depth within the structure can be obtained. The pulse-echo technique generally uses a single transducer as both the transmitter and receiver. Provided the ultrasonic pulses generated by the transducer are short enough, the individual echoes from each interface can be resolved, their position and amplitude being used to detect the presence of a defect. A large fraction of the pulse will be reflected at a defect owing to its large reflection coefficient; echoes from features behind the defect will also be reduced or disappear. It is also possible to use a single transducer in a “double through-transmission” configuration. This is achieved by placing a reflector plate beneath the structure in an immersion tank as shown in Fig. 13 and monitoring the echo from the reflector plate.

42

Nondestructive Testing

Fig. 13 Water bath used for ultrasonic C-scan inspection (From Adams et al. 1997)

1183 Probe (sender and receiver) Water Structure Reflector plate

As in through transmission, this signal is almost completely attenuated by disbonds larger than the beam, smaller amplitude reductions being produced by voids, porosities, and smaller disbonds. In all pulse-echo testings, alignment of the transducer axis perpendicular to the surface of the structure or reflector plate is important if the reflected signal is to be received back at the transducer. Sometimes, a second receiver probe is used instead of the reflector, the two probes being linked by a yoke. This is often necessary with honeycomb structures since the returning signal is very weak and is swamped by the much more powerful reflected signals. The probe is traversed automatically over the structure, and the amplitude of the signal associated with transmission through the structure is indicated on an X-Y or similar recorder. The X and Y coordinates correspond to the position on the surface, while some other parameter, such as gray scales or false color, is used to indicate quality. Because of the inconvenience of immersing the whole component or specimen, special water-jet probes can be used on large (aerospace) structures. Again, probe alignment is a serious problem. Good coupling between the transducers and the structure is essential and can be achieved by pressing them together with a film of glycerine or some similar coupling agent between them. Liquid couplants have the disadvantage of affecting the surface if this is in any way absorbent. Alternatives include laser-induced and monitored ultrasonics (Monchalin 1993) and air-coupled probes (Farlow and Hayward 1994; Schindel and Hutchins 1995). Other non-contacting ultrasonic transducers include electromagnetic acoustic transducers (EMATs) (Thompson 1990). Another method is to use a rubber wheel (Drinkwater and Cawley 1994) which can be rolled over the surface. The mechanical characteristics of the rubber are carefully chosen for maximum impedance matching and minimum attenuation. Several methods of displaying ultrasonic reflections are available, the most common being A- and C-scans. The simplest presentation is an A-scan which shows the amplitude of the signal as a function of time (or distance, if a value for the velocity of sound in the medium is known), as shown in Figs. 14 and 15. In manual testing, an A-scan is obtained at each test point and is interpreted by the operator. If the amplitude of the through-transmitted signal (or of a particular echo in pulseecho testing) is monitored at each point on the surface of the test piece, a C-scan can be produced. Measurements at each point are taken using a scanning mechanism, which produces a plan of the defect positions but gives no information on their depth. The example of Fig. 16 shows a C-scan of an adhesive joint sample roughly 100 mm square that had been subjected to environmental attack. Large edge

1184

R. D. Adams

Signal amplitude

Reflections from top adherend/ adhesive interface

Reflections from bottom adhesive/adherend interface

Top surface reflection 0

1

2

3

4

5 Time/ms

Fig. 14 Schematic of A-scan from a good joint (From Guyott et al. 1986)

disbonds can clearly be seen, together with groups of small disbonds remote from the edges. Corrosion of the aluminum on one side of the sample that was unbonded is also evident. The automatic scanning mechanisms required to produce C-scans usually employ immersion or water-jet coupling, whereas A-scan devices often use the contact technique. Portable scanning frames to facilitate C-scans on site are also available commercially. Other ultrasonic techniques which show promise are oblique incidence waves (Pilarski et al. 1990) and Lamb waves (Bork and Challis 1995). Rose and coworkers (1995, 1996; Puthillath and Rose 2010; Rose and Ditri 1992) have used guided ultrasonic waves to detect defects in adhesively bonded joints. Gowini and Moussa (2016) used surface acoustic waves to investigate the change in surface acoustic wave (SAW) phase velocity due to the change in adhesion of a thin film using an aluminum nitride (AlN)/silicon (Si) SAW sensor. They showed that as the adhesion of the film weakens on the surface of the AlN/Si SAW sensor, the velocity dispersion profiles fluctuate since as the adhesion of the layer improves the stress transmitted increases and the wave is more confined in the film layer. Gauthier et al. (2017) Used Lamb waves and showed that modes could be very sensitive to the nature of the interface, that is to say to the level of adhesion, if properly selected.

42.4.2 Sonic and Ultrasonic Vibrations In the through-thickness direction, an adhesively bonded joint can be regarded as an oscillating system. In its simplest form, the system consists of two masses separated by a spring [the adhesive]. Any variation from a given standard might be thought to indicate the presence of a defect. Unfortunately, the system is more complex than

42

Nondestructive Testing

1185

a Reflection amplitude

Groups of echoes from the bonded joint

Acoustic length of water path between joint and reflector

Time

b

Reflection amplitude

Slowly decaying echoes from disbonded joint

Time

Fig. 15 A-scans from (a) good and (b) disbanded joints (From Guyott et al. 1986)

this. For instance, a reduced stiffness of the adhesive layer might be due to porosity, non-optimum cure, cracks, or simply a thicker layer than usual. On the other hand, if the joint was in a panel which was consistently produced by using a film adhesive to control the bond-line thickness, then any variation in frequency would indicate a defect of some sort, although it would not be easy to ascertain of what kind. Sonic vibration techniques, using frequencies of 1–30 kHz, measure the local stiffness of the structure. Disbonds reduce the local stiffness as measured perpendicular to the surface. Usually, only disbonds or voids can be found, the minimum detectable size depending on the size and depth of the defect. Although the detectable defect size is larger than that obtained using ultrasonic techniques, the tests are often more convenient since they do not require a couplant between the transducer and the test structure. Commercially available mechanical impedance measuring instruments generally take measurements at a single preset frequency. The impedance is highest over good areas and lowest over disbonded areas. The test becomes unreliable when the structure itself becomes more flexible as the impedance of a defective zone can be higher or lower than that of a good zone, depending on the frequency (Cawley 1984,

1186

R. D. Adams Corrosion of unbonded plate

Large edge disbonds

Clusters of small disbonds

Fig. 16 C-scan of joint after environmental attack showing large edge disbonds and smaller disbonds remote from edges (From Vine et al. 2001)

1988). One advantage of this type of test is that, instead of using a couplant, a dry point contact is used between the transducer and structure. This dry contact has a finite stiffness (Lange and Teumin 1971) which must be kept as high as possible, to maximize the sensitivity of the technique. A different principle is used in ultrasonic resonance testing. Here, thickness vibrations are induced in the adherend/adhesive/adherend sandwich in which the two adherends are considered to act as rigid masses, while the adhesive is an almost massless spring. If the type and thickness of the metal sheets are constant, then the resonant frequency is a function of the thickness and bulk modulus of the adhesive. By injecting short pulses of broadband ultrasound (0.5–15 MHz) into a bonded panel and using modern digital techniques to obtain a frequency spectrum in “real time” of the structural response, it is possible to observe differences between good and bad structures. Lloyd and Brown (1978) showed that certain phenomena, such as water absorption, can be readily detected by this method. A general preview of vibration techniques for testing composites and adhesives is given by Cawley and Adams (1987).

42.4.3 Spectroscopic Methods Ultrasonic spectroscopy or the measurement of the through-thickness vibration characteristics of an adhesive joint can be used for the nondestructive testing of

42

Nondestructive Testing

1187

the cohesive properties of the adhesive layer (Cawley and Hodson 1988). The modulus of the adhesive can be determined from measurement of the throughthickness natural frequencies if the thickness of the adhesive layer is known. The Fokker Bond Test Mk II (Schliekelmann 1975), which uses a spectroscopic measurement, is the only commercially available instrument which claims to indicate the cohesive properties of the adhesive in a joint. Frequency and amplitude changes in the first two modes of through-thickness vibration of a system comprising the transducer and the joint are measured. These parameters depend on both the adherend and the bond-line thicknesses as well as the material properties. The range of frequencies over which the instrument operates is typically between 0.3 and 1.0 MHz. Small voids and disbonds at different depths in a multilayer joint can be detected reliably. However, prediction of the cohesive strength of the adhesive is questionable (Guyott et al. 1987) since the frequency shifts, resulting from a change in cohesive properties or bond-line thicknesses, are small and of a similar magnitude to each other. Thus, to obtain a true measure of the cohesive properties with this instrument, the bond-line thickness must be kept constant (or be measured separately). Roth and Giurgiutiu (2017) used piezoelectric wafer active sensors which were permanently bonded to the aircraft structure. Electromechanical impedance spectroscopy, a local vibration method, was used to detect disbonds from the changes in the mechanical impedance of the structure surrounding the disbond. The main problem with this method is the need to bond the sensors in a sufficient number of places to provide a realistic NDE method.

42.4.4 Other Acoustic Methods Tapping the structure with a hammer or coin is one of the oldest methods of nondestructive inspection. This is essentially a subjective method, and there has been considerable uncertainty about the physical principles involved. When a structure is struck with a hammer or coin, the impact depends on the local impedance of the structure and the hammer used. The local change in structural stiffness produced by a defect changes the nature of the impact. This is indicated by the force-time history of the impact on the structure which can be measured by incorporating a force transducer in the hammer. Impacts on good and disbonded areas of an adhesively bonded structure are shown in Fig. 17. Characteristically, the impact on the sound structure has a higher peak force and a shorter duration than that on the damaged area. Thus, a measure of either peak force or duration can be used to locate defects. Since the method uses only measurements of impact force, no transducers need to be attached to the structure, thus avoiding the coupling and alignment problems which arise with ultrasonic techniques (Cawley and Adams 1988, 1989). The sensitivity and reliability can be increased by further processing the force-time histories such as by taking Fourier transforms of the signal to present the data in the frequency domain (force vs. frequency). Hart-Smith and Thrall (1985) recommend the coin-tap test since it is unable to find flaws that are so small that they can be ignored anyway!

1188

R. D. Adams

Force

Force

Time

Time

1.5–2.0ms

1ms

Impact over defect

Impact over good joint

Fig. 17 Tapping with an instrumented hammer over a good and a disbonded joint

42.4.5 Acoustic Emission Application of stress to a material will eventually lead to microscopic fracture or slip (Curtis 1982). This is usually associated with a local release of energy which propagates as a stress wave. The wave has a high-frequency content and is referred to as “acoustic emission” [or stress wave emission] which can be detected either by a high-frequency microphone or by a piezoelectric transducer. Unfortunately, it is necessary to stress the joint to a high proportion of its failing load in order to generate sufficient emissions, and there is doubt whether this is practical in most cases. It is therefore not recommended for general use although there are occasions where it may be effective.

42.4.6 Thermal Methods By heating one surface of a bonded sandwich structure and observing the temperature rise of the opposite face, areas of debond, which resist the transfer of heat, show as cool areas. Alternatively, if the heated face is scanned, debonds will show as hot areas. The temperature rise on the heated surface is governed by the amount of energy deposited and the speed of application, combined with the thermal properties of the surface material. If the thermal pulse is short enough, the ensuing diffusion process is totally controlled by the material itself. The contrast observed at either surface due to the presence of defects depends on the defect size, its depth from the observed surface, the initial surface temperature rise, and the thermal properties of the material. While these parameters change from one specimen to another, the testing technique is always to record the temperature variation of either surface directly after the thermal transient has been applied. Temperature differences caused by the presence of defects may be seen over timescales ranging from sub-millisecond to

42

Nondestructive Testing

1189

several seconds, depending on the material properties and thickness. The useful information is often obtained within 500 ms, so it is necessary to use a system which can acquire many sequential images in this time window. The equipment required to perform transient thermography falls into two categories: the transient heat source and the thermal imaging/analysis system. The heat pulse must have a sufficiently fast rise time to provide a rapid temperature rise since it is the steepness of the temperature gradient which provides the contrast between defective and non-defective areas. In materials of high thermal diffusivity [the ratio of thermal conductivity to mass density] such as metals and carbon composites, this rapid temperature rise can conveniently be provided by discharging several kilojoules of energy from a bank of capacitors through xenon flashtubes which are directed at the area of interest. For poor thermal conductors such as glass fiber composites, the rate of heating produced by hot-air blowers is frequently adequate. Temperature sensing is normally done with a scanning infrared camera. Heat pulses or moving heat sources were used by Reynolds (1988). Temperature-sensitive paints of liquid crystals and thermoluminescent coatings can also be used. Highpower lasers can be used, but care must be taken to avoid causing damage to the surface. Major improvements in thermal imaging equipment have been made in recent years. Early work on the method used a thermal imaging camera whose output was recorded on a video tape, the analysis being performed on the recorded image using slow-motion replay and freeze-frame facilities. (The technique was once called pulse video thermography (Hobbs et al. 1991).) Now the images are stored digitally, and automatic processing routines are available (Shepard 1997). In general, the sensitivity of the method is reduced as the depth of the defect from the surface monitored by the thermal imaging camera is increased. The sensitivity is best expressed as the defect diameter/depth ratio required for the defect to be detectable. This sensitivity is material specific and must be determined experimentally using specially fabricated test plates containing flat-bottomed holes of different diameters and depths. Typical values of the ratio of minimum detectable defect diameter to defect depth are 2–4 for metals (Hobbs et al. 1994). The performance in carbon fiber composite may be somewhat poorer than this because the composite has higher conductivity in the plane of the fibers than in the through-thickness direction. The heat therefore tends to diffuse parallel to the surface, rather than through the thickness, and so is less affected by the presence of disbonds or delaminations that run parallel to the surface. However, there is a wide variety of sensitivities reported in the literature on carbon fiber composites (Hobbs et al. 1994). In general, the best results are obtained if the maximum possible amount of energy is deposited on the surface of the structure, though care must be taken to avoid thermal damage. Conventional X-ray techniques are of little use on metal-to-metal bonded joints since the polymeric adhesive is much less dense than the adherends. Metallic fillers (Segal and Kenig 1989) can be used to enhance the contrast and show tapering or voids. However, the density of fiber-reinforced plastics adherends is of a similar order to the adhesive and so X-rays can be used, by choosing a suitable energy and flux. For honeycomb-cored panels, X-rays are used for checking the position of the

1190

R. D. Adams

core and whether it has been locally crushed or otherwise damaged and to monitor the position of water ingress. A thermal neutron source can also be used (Michaloudaki et al. 2005) since neutrons are absorbed or scattered by hydrogenous materials such as hydrocarbons. Unfortunately, such sources are typically nuclear reactors and accelerators, and so are neither cheap nor readily available! Tighe et al. (2016) interrogated carbon fiber-reinforced plastic (CFRP) single-lap joints to compare the detection of different defect types using pulsed phase thermography (PPT). A PTFE insert, of the type widely used to simulate defects in composite materials, was added to the bond line of the joint. Liquid layer kissing defects were simulated using silicon grease. PPT clearly identified the PTFE insert but not the silicon grease contamination. PPT identified the silicon grease defect when the joint was loaded, probably because loading opens the defect and produces a gap that provides sufficient thermal contrast for detection.

42.4.7 Optical Holography By using holographic interferometry, it is possible to measure surface displacements to less than 0.5 μm. Load is applied to a structure by vibration, vacuum cup, pressure, or heat. Defects show as local perturbations in the holographic interferogram (Rao et al. 1990). The technique has found application with sandwich structures but not lap or similar joints.

42.4.8 Visual Hart-Smith and Thrall (1985) recommend a visual inspection of the adhesive fillet and suggest it is the most effective of all inspection techniques for bonded joints. The state of the fillet is a good indication of the state of the adhesive at the end of the joint. A porous fillet indicates too high a heat-up rate or the presence of moisture in the joint. A lack of fillet indicates a lack of pressure and may indicate internal voids. Just as the water-break test is an indicator of the adherend surface, the contact angle of the cured adhesive at the end of the fillet, formed when the adhesive was liquid, indicates if surface contamination is present.

42.5

Future Developments

The major problem with nondestructively assessing the strength of adhesively bonded joints remains the so-called kissing bond, in which there is continuity, but little or no strength. Many years ago, the author tried to generate a high-intensity tensile stress pulse which would fracture a kissing bond. Using mechanical impact, it proved impossible to achieve a sufficiently short-duration pulse which could be reflected as a tensile pulse in a thin adhesive joint. Gupta and coworkers (1990;

42

Nondestructive Testing

1191

Gupta and Yuan 1993; Yuan and Gupta 1993, 1998) used high-intensity lasers to create tensile shock waves which could fracture weak bonds between a coating and a substrate. Bossi and coworkers at Boeing (Bossi et al. 2002, 2004, 2009, 2011) also tried to generate high-intensity, short-duration stress pulses to fracture these kissing bonds in adhesively bonded joints. After experimenting with flash electron beams, they settled on high-intensity lasers which gave pulses of a few hundred nanosecond duration. With joints made using carbon fiber-reinforced plastic (CFRP) adherends, they were able to distinguish bond quality on surfaces created with different peel ply preparations such as polyester and nylon fabric and surfaces prepared with contamination using a silicone release agent. When examined with conventional ultrasonics, these joints appeared to be identical. The technique developed by Bossi et al. is to use a laser-induced pulse to fracture a weak kissing bond and then to use conventional ultrasonics to detect this fracture. They have developed the technique for application to real aircraft structures. Monchalin 2007 has also worked on this method. The problem is to fracture a weak bond while not fracturing an otherwise good bond and to leave it in such a state so that it can be seen using conventional ultrasonics. The technique is interesting and seems to offer a way forward.

42.6

Conclusions

The types of defect encountered in adhesive joints and the nondestructive testing techniques available to detect them have been reviewed. Several techniques are available for disbond detection, the most commonly used being ultrasonics and different types of bond tester. These techniques are very time-consuming if large bond areas are to be tested; so there is increasing interest in rapid scanning methods such as transient thermography. The detection of poor cohesive properties is more difficult but can be achieved with ultrasonic or dielectric measurements. Monitoring interfacial properties is much more difficult, and there is currently no reliable test after the joint has been made. The high-intensity laser method is the only possible technique at the moment, but this needs a lot of development before it might become commercially acceptable. Finally, we still need to correlate measurable defects with actual strength in the way in which joints are used, that is, in the tensile lap-shear mode where the main stresses are at the joint ends. Since the middle of the joint is relatively lowly loaded, detection of voids in the middle may indicate poor production and a possible leakage path, but it may not necessarily indicate a weak joint.

References Adams R, Cawley P, Guyott CCH (1987) Nondestructive inspection of adhesively-bonded joints. J Adhes 21:279–290 Adams RD, Comyn J, Wake WC (1997) Structural adhesive joints in engineering, 2nd edn. Chapman & Hall, London, 359 pp

1192

R. D. Adams

Baker A, Rajic N, Davis C (2009) Towards a practical structural health monitoring technology for patched cracks in aircraft structure. Compos Part A 40:1340–1352 Bijlmmer PA (1978) In: Allen KW (ed) Adhesion, 2nd edn. London, Applied Science Publishers, 45 p Bork U, Challis R (1995) NDE of the adhesive fillet size in a T-peel joint using ultrasonic lamb waves and a linear network for data discrimination. Meas Sci Technol 6:72–84 Bossi R, Housen K, Shepherd W (2002) Using shock loads to measure bonded joint strength. Mater Eval 60:1333–1338 Bossi R, Housen K, Shepherd W (2004) Application of stress waves to bond inspection. In: SAMPE proceedings, Long Beach Bossi R, Housen K, Shepherd W, Sokol D (2009) Laser bond testing. Mater Eval 67:819–827 Bossi R, Lahrman D, Sokol D, Walters C (2011) Laser bond inspection for adhesive bond strength. In: SAMPE proceedings, Long Beach Cawley P (1984) The impedance method of nondestructive inspection. NDT Int 17:59–65 Cawley P (1988) The sensitivity of the mechanical impedance method of NDT. NDT Int 20:209–215 Cawley P (2005) Nodestructive testing. In: Adhesive bonding: science, technology and applications. Woodhead, Cambridge, UK Cawley P, Adams RD (1987) Vibration techniques. In: Summerscales J (ed) Nondestructive testing of fibre-reinforced plastics. Elsevier Applied Science, London, pp 151–200 Cawley P, Adams RD (1988) The mechanics of the coin-tap method of nondestructive testing. J Sound Vib 122:299–316 Cawley P, Adams RD (1989) The sensitivity of the coin-tap method of non-destructive testing. Mater Eval 47:558–563 Cawley P, Hodson MJ (1988) The NDT of adhesive joints using ultrasonic spectrscopy. In: Thompson DO, Chimenti DE (eds) Review of progress in QNDE 8B. Plenum Press, New York, pp 1377–1384 Crane RL, Dillingham G (2008) Composite bonding. J Mater Sci 43(20):6682–6694 Curtis GJ (1982) Nondestructive testing of adhesively-bonded structures with acoustic methods, ultrasonic testing – non conventional testing techniques. Wiley, Chichester Drinkwater BW, Cawley P (1994) An ultrasonic wheel probe alternative to liquid coupling. Brit J NDT 36:430–433 Farlow R, Hayward G (1994) Real-time ultrasonic techniques suitable for implementing non-contact NDT systems employing piezoceramic composite transducers. Brit J NDT 36:926–935 Galy ME, Moysan J, El Mahi A, Ylla N, Massacret N (2017) Controlled reduced-strength epoxyaluminium joints validated by ultrasonic and mechanical measurements. Int J Adhes Adhes 72:139–146 Gauthier C, El Kettani ME, Galy J, Pedoi M, Leduc D, Izbicki J-L (2017) Lamb waves characterization of adhesion levels in aluminum/epoxy bi-layers with different cohesive and adhesive properties. Int J Adhes Adhes 74:15–20 Gowini MEL, Moussa WA (2016) Investigating the change in surface acoustic wave velocity due to the change in adhesion of a SU-8 thin film using a SU-8/AlN/Si SAW sensor. Int J Adhes Adhes 68:102–114 Gupta V, Yuan J (1993) Measurement of interface strength by the modified laser spallation technique, II applications to metal/ceramic interfaces. J Appl Phys 74:2397–2404 Gupta V, Argon AS, Cornie JA, Parks DM (1990) Measurement of interface strength by laserpulseinduced spallation. Mater Sci Eng A126:105–117 Guyott CCH, Cawley P, Adams RD (1986) The non-destructive testing of adhesively bonded structure: a review. J Adhes 20:129–159 Guyott CCH, Cawley P, Adams RD (1987) Use of the Fokker bond tester on joints with varying adhesive thickness. Proc IMechE 201(B1):41–49 Hagemaier DJ (1985) Adhesive bonding of aluminium alloys. Marcel Dekker, New York

42

Nondestructive Testing

1193

Hart-Smith LJ, Thrall EW (1985) Adhesive bonding of aluminium alloys. Marcel Dekker, New York, pp 241–335 Hobbs CP, Kenway-Jackson D, Milne JM (1991) Quantitative measuremement of thermal parameters over large areas using pulse video thermography proc SPIE, vol 1467. Thermosense XIII, Orlando, pp 264–277 Hobbs CP, Kenway-Jackson D, Judd MD (1994) Proceedings of the international symposium on advanced materials for lightweight structures. ESTEC, Noordwijk, (ESA-WPP-070) Jeenjitkaew C, Guild FJ (2017) The analysis of kissing bonds in adhesive joints. Int J Adhes Adhes 75:101–107 Jeenjitkaew C, Luklinska Z, Guild FJ (2010) Morphology and surface chemistry of kissing bonds in adhesive joints produced by surface contamination. Int J Adhes Adhes 30:643–653 Karachalios EF, Adams RD, da Silva LFM (2013) The strength of single lap joints with artificial defects. Int J Adhes Adhes 45:69–76 Kim DM, Sutliff E (1978) The contact potential difference (CPD) measurement method for prebond nondestructive surface inspection. SAMPE Q 9:59–63 Lange YV, Teumin II (1971) Dynamic flexibility of dry point contact. Sov J NDT 7:157–165 Lloyd EA, Brown AF (1978) In: Allen KW (ed) Adhesion, 2nd edn. London, Applied Science Publishers, 133 pp Marty NP, Desai N, Andersson J (2004) NDT of kissing bond in aeronautical structures. In: Proceedings of the 16th world conference for NDT, Montreal Michaloudaki M, Lehmann E, Kosteas D (2005) Neutron imaging as a tool for the non-destructive evaluation of adhesive joints in aluminium. Int J Adhes Adhes 25:257–267 Monchalin J-P (1993) Progress towards the application of laser ultrasonics in industry. In: Thompson DO, Chimenti DE (eds) Review of progress in QNDE 12. Plenum Press, New York, pp 495–506 Monchalin J-P (2007) Laser-ultrasonics: principles and industrial applications, Chapter 4. In: Chen CH (ed) Ultrasonic and advanced methods for nondestructive testing and material characterization. World Scientific, Singapore, pp 79–115 Pilarski A, Rose JL, Balasubramian K (1990) The angular and frequency characteristics of reflectivity from a solid layer embedded between two solids with imperfect boundary conditions. J Acoust Soc Am 87:532–542 Puthillath P, Rose JL (2010) Ultrasonic guided wave inspection of a titanium repair patch bonded to an aluminum aircraft skin. Int J Adhes Adhes 30:566–573 Pyles R (2003) Aging aircraft: USAF workload and material consumption life cycle patterns. RAND, Pittsburgh Rao MV, Samuel R, Ramesh K (1990) Dual vacuum stressing technique for holographic NDT of honeycomb sandwich panels. NDT Int 23:267–270 Reynolds WN (1988) Inspection of laminates and adhesive bonds by pulse-video thermography. NDT Int 21:229–232 Rose JL, Ditri JJ (1992) Pulse-echo and through transmission lamb wave techniques for adhesive bond inspection. Br J Nondestr Test 34:591–594 Rose JL, Rajana KM, Hansch MKT (1995) Ultrasonic guided waves for NDE of adhesively bonded structures. J Adhes 50:71–82 Rose JL, Rajana KM, Barshinger JM (1996) Guided waves for composite patch repair of aging aircraft. In: Thompson DO, Chimenti DE (eds) Review of progress in QNDE. Plenum Press, New York, pp 1291–1298 Roth W, Giurgiutiu V (2017) Structural health monitoring of an adhesive disbond through electromechanical impedance spectroscopy. Int J Adhes Adhes 73:109–117 Schindel DW, Hutchins DA (1995) Applications of micromachined capacitance transducers in air-coupled ultrasonics and nondestructive evaluation. IEEE Trans Ultrason Ferroelectr 42:51–58 Schliekelmann RJ (1972) The nondestructive testing of adhesively bonded metal-to metal joints. Nondestruct Test 5:79–86

1194

R. D. Adams

Schliekelmann RJ (1975) Nondestructive testing of bonded joints – recent developments in testing systems. Nondestruct Test 8:100–103 Segal E, Kenig S (1989) Acceptance criteria for NDE of adhesively bonded structures. Mater Eval 47:921–927 Shepard SM (1997) Introduction to active thermography for non-destructive evaluation. Anti Corros Methods Mater 44:236–239 Thompson RB (1990) Physical principles of measurements with EMAT transducers. In: Mason WP, Thurston RN (eds) Physical acoustics, vol XIX. Academic, New York, pp 157–200 Tighe RC, Dulieu-Barton JM, Quinn S (2016) Identification of kissing defects in adhesive bonds using infrared thermography. Int J Adhes Adhes 64:168–178 Vine K, Cawley P, Kinloch AJ (2001) The correlation of non-destructive measurements and toughness changes in adhesive joints during environmental attack. J Adhes 77:125–161 Yu A, Gupta V (1998) Measurement of in situ fiber/matrix interface strength in graphite/epoxy composites. Compos Sci Technol 58:1827–1837 Yuan J, Gupta V (1993) Measurement of interface strength by the modified laser spallation technique. I. Experiment and simulation of the spallation process. J Appl Phys 74:2388–2397

Techniques for Post-fracture Analysis

43

Romain Créac’hcadec

Contents 43.1

43.2

43.3

43.4

43.5

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1.1 Structural Bonding, a Plural Science . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1.2 The Need for a Plural Science to Solve the Equation of Bonding . . . . . . . . . . . 43.1.3 Interactions Between These Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1.4 What About the Role of Interfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.1.5 The Importance of Multi-scale and Multi-physics Observations . . . . . . . . . . . . . 43.1.6 The Importance of Observation on the Understanding of the Physics of Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The Difficulty of Identifying an Adhesive or Cohesive Failure . . . . . . . . . . . . . . . . . . . . . . . 43.2.1 Presentation of the Different Modes of Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.2.2 Subjectivity of the Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.2.3 A Multitude of Experimental Setups to Try to Understand the Failure of Bonded Assemblies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Macroscopic and Mesoscopic Optical Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.3.1 Failure Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.3.2 The Importance of Local Geometry on Adhesive Joint Failure . . . . . . . . . . . . . . Microscopic Optical Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.1 One-Dimension Roughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.2 Scanning Electron Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.3 Nanoindentation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.4 Atomic Force Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.5 Confocal Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.4.6 X-Ray Microtomography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Physicochemical Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.5.1 Scanning Electron Microscopy with Energy-Dispersive X-Ray Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.5.2 Scanning Electron Microscopy with Raman Spectroscopy . . . . . . . . . . . . . . . . . . . 43.5.3 X-Ray Photoelectron Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1196 1196 1197 1198 1199 1200 1200 1202 1202 1202 1204 1204 1204 1207 1209 1209 1210 1212 1212 1213 1214 1215 1215 1216 1218

R. Créac’hcadec (*) ENSTA Bretagne, Institut de Recherche Dupuy de Lôme, FRE CNRS 3744, Brest, France Assemblages Multi-Matériaux, Institut de Recherche Dupuy De Lôme, Brest, France e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_43

1195

1196

R. Créac’hcadec 43.5.4 43.5.5 43.5.6

Differential Scanning Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Thermogravimetric Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) Spectroscopy Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.6 Mechanical Observations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.6.1 3D Digital Image Correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.6.2 Digital Volume Correlation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.6.3 Ultrasonic Testing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1219 1223 1223 1224 1224 1226 1228 1228 1229

Abstract

Fracture observation is one of the key ways to understand the behavior of bonded assemblies. The plural science of bonded assemblies includes three main areas such as polymer sciences, physical chemistry of surfaces and interfaces, and mechanics. Solving the equation of adhesion needs intersection of all these aspects. Moreover, this multi-physics science is also a multi-scale one. Understanding surface fracture deals with all these aspects. This chapter presents three types of measurement systems: microscopy, compositional analysis/material identification, and mechanical observations. A table sums up an exhaustive list of the major observation devices and precise references of works that use these types of methods. Then, the main methods are detailed with macroscopic and mesoscopic observations, microscopic observation, physicochemical analysis, and mechanical observations.

43.1

Introduction

One can wonder why people are so interested on the failure mode of bonded assemblies. One reason is probably the need for manufacturers and users in various industrial fields to trust in this technology of multi-material assemblies. Users are very often interested in having cohesive failures in the adhesive, mainly for a reason that is not so objective. The idea is linked to the need to have predictable failures. So, cohesive failure means that physical phenomena happen in the “material” of the adhesive. It seems to be easier to predict it than with the contribution of the interfaces. However, sometimes these failures occur at the interface with interesting strength, etc. So, how can we understand these fractures?

43.1.1 Structural Bonding, a Plural Science By definition, a bonded structure is composed of two substrates connected together by a bonding agent called the adhesive. This bond defines two interfaces connecting each other the substrates to the adhesive. The assembly made of the adhesive and the

43

Techniques for Post-fracture Analysis

Substrate 1

1197

Interface 1

Adhesive joint

Adhesive Substrate 2

Interface 2

Fig. 1 Structure of an adhesive joint made of two interfaces and the adhesive Fig. 2 Structural bonding, a plural science

Mechanics [3] (A)

(B) Bonding Collage

Polymers Science [1]

(C)

Physical chemistry of surfaces and interfaces [2]

two interfaces constitutes an adhesive joint. Figure 1 presents a very schematic view of this structure. The analysis of this structure, therefore, requires understanding the behavior of each constituent: adhesive, interfaces, and substrates. Each of these elements, by its essence, contributes to influencing the response of such a structure subjected to mechanical loading.

43.1.2 The Need for a Plural Science to Solve the Equation of Bonding Structural bonding is a plural science. Representing it, in a schematic way, is not easy as the domains linked to structural bonding interpenetrate. Figure 2 presents the problem to solve for finding an acceptable solution for the behavior of a bonded structure. The solution is defined in the center (in white). This area is the overlap of three strong scientific domains: the science of polymers (1), the physico-chemistry of surfaces and interfaces (2), and mechanics (3). Understanding and being able to predict the behavior of a bonded structure mean being able to cover and interact with

1198

R. Créac’hcadec

these three areas. This plurality makes difficult, but nevertheless very interesting, the problematic of structural bonding.

Area of Polymers Science (1) Adhesives are for the most part polymers. Polymers constitute a class of materials in their own. Because these materials possess dissipative processes, they are very useful for bonded assemblies. They are generally characterized by high breakage energies. Area of Physico-chemistry of Surfaces and Interfaces (2) The notion of interface defines the boundaries of a domain. Indeed, the field of application of the “science of polymers” is limited because of the presence of the two substrates. These polymer/substrate boundaries provide an interface that needs to be understood. An adhesive is not only a polymer, but it is also a material with an adhesive capacity. Understanding of the phenomena of adhesion requires analysis on several scales. Several concepts from the molecular scale up to the macroscopic scale are thus studied. Total adhesion is generally analyzed by decomposition into mechanical adhesion, physical adhesion, and chemical adhesion. Mechanical adhesion incorporates the notions of wetting, mechanical anchoring, and diffusion/ interdiffusion. Physical adhesion is concerned with the different forces that interact at short and long distances (van der Waals, electrostatics, etc.). Finally, chemical adhesion is concerned with covalent bonds. Numerous references highlight the complexity of these phenomena and theories on adhesion. Area of Mechanics (3) The last domain proposed to solve the equation of bonding is mechanics. The problems proposed by the industry must meet a specification. This document specifies under which conditions and under which constraints the bonded assembly must resist: durability, environment, stress rates, temperatures, fracture loads, etc. The difficulty of this field is, therefore, to be able to propose means of characterization and prediction of the mechanical strength of the assembly.

43.1.3 Interactions Between These Areas The solution of the structural bonding problem requires a very wide range of skills which are difficult to dissociate from one to another. Thus, in order to show these interactions, Fig. 2 defines three subdomains marked “A, B, and C,” highlighting these interactions between the “science of polymers,” the “physico-chemistry of surfaces and interfaces,” and “mechanics.” The intersection of these interactions constitutes an approach for solving the problem of structural bonding. Initially, several decades ago, these interactions were weak. Today, these areas are increasingly interacting. The best marker of this evolution is the organization of congresses around adhesive bonding which often revolve around two parallel sessions, one physico-chemistry and the other rather mechanical.

43

Techniques for Post-fracture Analysis

Polymer (6)

1199

Contamination layer (H2O, CO2, …) (4)

Substrate-modified polymer layer (5) Layer created as a result of surface treatments (3) Surface layer of the substrate (2) Substrate (1)

Fig. 3 Representation of a polymer-metal interphase (Coulaud 2007)

43.1.4 What About the Role of Interfaces The analysis of the failure modes of bonded assemblies differs from the “classical” analyses of materials by the eternal question, what is the role of the interfaces?

Example of Structure: Polymer/Metal Interface The interphase is a geometric boundary that defines the transition between the adhesive and the polymer. This zone is characterized by numerous physical, chemical, and mechanical phenomena. Figure 3 presents a sectional view of this bonding zone (Coulaud 2007). The structure is characterized by numerous layers from the substrate (1) to the polymer (6): the surface layer of the substrate (2), a layer resulting from surface treatments (3), a contamination layer (H2O, CO2, etc.), and a layer of polymer modified by the substrate (5). This structure is specific to the nature of the materials used (substrate and polymer) as well as to the manufacturing process. It is therefore necessary to question its influence on the macroscopic behavior observed. Example of a Macroscopic Behavior Figure 4 shows the influence of interfaces on the macroscopic behavior of adhesives (Arnaud 2014). For Hysol ®-EA9309.3NA adhesive, modified Arcan shear tests were performed. Two treatments were compared. The first is an unclogged sulfur anodizing treatment. The second treatment comprises corundum abrasion 220 followed by an anti-stick treatment. The same curing cycle was applied to the adhesive. For treatment A, the maximum load reached is close to 16 kN, and the maximum displacement at break is approximately 0.3 mm. The observed behavior is nonlinear and identical for the three tests performed. For the specimen with antiadhesion, the nonlinear behavior followed is the same. For the second, load and displacement at failure are considerably lower. Moreover, for the three tests carried out, the displacements at break (visible with the green, blue, and brown dots) are dispersed. Given the plasticity of the adhesive, the fracture loads are similar. For all six tests, the failure is adhesive. This example clearly highlights the contribution of interfaces to the response of a bonded joint. Figure 4b shows the multilayer structure of an adhesive joint: the adhesive, the two interfaces, and the substrates. Vallat (2013) symbolizes the contribution of this structure as a chain composed of five

1200

R. Créac’hcadec

(a)

(b) Traitement A Treatment A

16

Load FT (kN)

14

TreatmentB B Traitement

12

Substrate 1 Interface Substrate 1 / Adhesive

10

Adhesive

8 6

Interface Substrate 2 / Adhesive

4 2 0 0.00

0.10

0.20

0.30

Substrate 2

Displacement DT (mm)

Fig. 4 Macroscopic role of interfaces on the behavior of an adhesive in an assembly: (a) demonstration of the influence of surface treatment on modified Arcan shear tests (Arnaud 2014) and (b) multilayer structure of a bonded joint (Vallat 2013)

links. Each link is an important element. It is the failure of one of these links, placed in series, which generates the failure of the whole assembly.

43.1.5 The Importance of Multi-scale and Multi-physics Observations As a consequence of the structure of bonded assemblies itself, to understand fracture, it is necessary to use multi-scale and multi-physics observations. In this chapter, most of these methods are presented.

43.1.6 The Importance of Observation on the Understanding of the Physics of Failure Structural Adhesives Are Sources of Defects The manufacturing process generates intrinsic defects during the production of the bonding. Adams characterizes these defects (see ▶ Chap. 42, “Nondestructive Testing”). Figure 5 presents the various defects observed. They may be bound to the surface preparation with areas having adhesion defects, with or without kiss bonding. These defects can be related to macroscopic porosities: absence of adhesive or to microscopic porosities. They can also be linked to the polymerization process: cross-linking defects, cracks generated during shrinkage, and cross-linking gradient. As a consequence, the modes of failure are a manifestation of different physics which can be observed in the fracture surfaces.

43

Techniques for Post-fracture Analysis

Fig. 5 Various defects which might be present in an adhesive bond line (Adams see ▶ Chap. 42, “Nondestructive Testing”)

1201

Cracks

Poor cure

Porosity Adherend Adhesive

Adherend

Disbond or zero-volume unbond

(a)

Voids

(b)

Stress (MPa)

40 30 20 10 0

0

2

4

6

8

Strain (%)

Fig. 6 Tensile tests on bulk samples: (a) Dog-bone samples, (b) Air bubbles in the section (Arnaud 2014)

Example of an Application: Dog-Bone Tensile Specimen Dog-bone tensile specimens are classical tests to characterize the mechanical properties: Young’s modulus and Poisson’s ratio. Arnaud et al. (2015) carried out these tests. Figure 6 presents the stress-strain curves. A linear behavior is firstly observed and then a nonlinear one. Tests were performed with a crosshead speed of 0.5 mm/min. Though special care was considered to manufacture the specimens, all the fractures occurred near a porosity observed in Fig. 6b. This a very simple example, but it highlights the need to well understand physics to explain the behavior observed. This analysis starts with the observation of the fracture surface. This chapter is dedicated to the different means to observe fracture in adhesive bonds. The main message that must be understood by the reader is the fact that observation setups are closely related to the comprehension of the physics observed.

1202

R. Créac’hcadec

a

b

Cohesive failure

c

Interfacial failure

d

Oscillatory failure

Alternating failure

Fig. 7 Different locus of failure and crack trajectories observed in mode I testing of adhesively bonded double cantilever beam (DCB) specimens. (a) Cohesive failure, (b) Interfacial failure, (c) Oscillatory failure, (d) Alternating failure. (Chen and Dillard 2001)

43.2

The Difficulty of Identifying an Adhesive or Cohesive Failure

43.2.1 Presentation of the Different Modes of Failure Chen et Dillard (2001) presented in their work different kinds of failures for adhesively bonded assemblies. Four types can be distinguished (Fig. 7): failure in the adhesive, also called cohesive failure or failure at the interface between the adhesive and the substrate. These modes are commonly observed and very easy to determine. Two types of failure called oscillatory and alternating failures also exist as described in Fig. 7c, d.

43.2.2 Subjectivity of the Analysis Figure 8 presents two specimens after failure. The test considered was a modified Thick-Adherend Shear Test that applied shear loadings to these interfaces (Cognard et al. 2011a). Two surface treatments were used: mechanical preparation with abrasion followed by acetone degrease and chemical treatment. With the first surface treatment, the failure is fully adhesive. With the second one, the failure is mainly cohesive (around 90% cohesive). Therefore, determining the role of the interface on the strength at failure seems to be quite easy. Figure 9 shows failure

43

Techniques for Post-fracture Analysis

1203

Fig. 8 Influence of the surface preparation on the failure mode for an epoxy adhesive: (a) results with a mechanical preparation and (b) results with a chemical one (Cognard et al. 2011a)

Fig. 9 Failures in single-lap joints with steel substrates: (a) fully adhesive failure, (b) 50% cohesive failure, (c) mixed mode adhesive-cohesive failure (Legendre et al. 2016)

surfaces for single-lap shear tests with hard steel sheet metal substrates with a zinc coating (Legendre et al. 2016). The failure is adhesive (Fig. 9a) and mixed-mode adhesive/cohesive in Fig. 9b. The main question that can be asked is about the surface described in Fig. 9c. How is the cohesive failure defined? If there are blue areas on both sides of the joint, it is easy to determine a cohesive failure. However, looking close to the interface, a “white” zone can be observed: a close adhesive failure zone. It seems that in this zone the interface transmitted the kinematic deformation of the substrate to the adhesive and after that broke. Research is still under way to determine whether this area can be defined as cohesive or not. It depends today on the country and on the industrial field where it is considered. One element in discussion is the fact that it is possible to predict or not a reliable strength at failure with this type of fracture surface.

1204

R. Créac’hcadec

43.2.3 A Multitude of Experimental Setups to Try to Understand the Failure of Bonded Assemblies The purpose of this chapter is not to give a detailed and exhaustive list of methods of analysis but to propose multi-scale and multi-physical analysis methods to analyze the failure of bonded assemblies. Thus, the following table gives a non-exhaustive list of characterization setups used in the three domains of bonded assemblies described above. Three types of analyses can be considered despite some recovery. – Microscopy techniques are used to mainly observe at different scales surfaces and adhesives from the millimeter (macroscopic observation) to the nanometer. These observations were in the past realized in two dimensions. Nowadays the third dimension can be more easily considered thanks to confocal microscopy or X-ray micro-tomography systems. – The second domain is used to analyze the composition and the identification of the material. These ones are mainly physicochemical analysis based on thermal and spectroscopy applications. – The third domain is based on the comprehension of the deformation of the bonded assemblies to understand fracture. It is mainly linked to kinematic observations in order to define the localization of the mechanical failure process. Table 1 presents these three domains. References of recent works are listed for each analysis. Only certain methods are described later in this chapter more precisely.

43.3

Macroscopic and Mesoscopic Optical Observations

43.3.1 Failure Surfaces Within the Adhesive The load path influences the failure mode within the adhesive. Créac’hcadec et al. (2014) characterized the behavior of a thick flexible adhesive (Sikaflex-265) under different monotonic tensile/compression-shear loads, using a modified Arcan fixture, which was designed to limit stress concentrations in order to define an experimental methodology enabling structural adhesives to be characterized up to failure. The results allow defining the load and displacement envelopes characterizing the adhesive to failure under multiaxial loading conditions. For the different tests, cohesive failures were observed around the median plane of the adhesive (Fig. 10). These fracture surfaces underline the good quality of the cohesion between Sikaflex ®-265 and ABS substrates using Sika ®Primer-206 G+P. It seems that a specific fracture pattern characterizes each test configuration. The load direction seems to determine the topography of the fracture pattern. For tensile tests, fracture surfaces are smooth and have a fault in the center of the surface as well

43

Techniques for Post-fracture Analysis

1205

Table 1 Different experimental devices to observe failures in bonded assemblies Acronym Name explanation Microscopy AFM Atomic force microscope Confocal microscopy FIB FFM OP SAM SEM

Focused ion beam Friction force microscopy Optical profilometer Scanning acoustic microscopy Scanning electron microscope

STM TEM/ STEM

Scanning tunneling microscopy Transmission electron microscopy/ scanning transmission electron microscopy X-ray micro-tomography

XRM

Compositional analysis and material identification Thermal analysis DSC Differential scanning calorimetry DTA Differential thermal analysis TGA Thermogravimetric analysis Spectroscopy AES Auger electron spectroscopy EBSD

Electron backscatter diffraction

EDS/ EDX FTIR

Energy-dispersive X-ray spectroscopy

GC-MS GDOES/MS ICPOES/MS

Gas chromatography mass spectrometry Glow discharge-optical emission spectrometry/mass spectrometry Inductively coupled plasma-optical emission spectrometry/mass spectroscopy Instrumental gas analysis Laser ablation-inductively coupled plasma-mass spectrometry Liquid chromatography-mass spectrometry Low-energy electron-induced X-ray emission spectrometry Nuclear magnetic resonance spectroscopy

IGA LA-ICPMS LC-MS LEXES NMR

Fourier transform infrared spectroscopy

References Critchlow 96 Aguiar et al. (2012), Beuguel et al. (2017), and Hase et al. (2014) Pecnik et al. (2015) – – Oehler et al. (2012) Aguiar et al. (2012) and Hase et al. (2014) – Mine et al. (2017) and Liu (2013)

Bele et al. (2017), Bouterf et al. (2017), and Liu (2013)

Devaux et al. (2015) Devaux et al. (2015) – Comyn (1992) and Critchlow et al. (2006) Haddadi et Tsivoulas (2016) and Koseki et al. (2016) Cardell et Guerra (2016) Gorassini et al. (2016), Zięba-Palus et al. (2016), and Zhao et al. (2016) Zięba-Palus et al. (2016) Heikkilä et al. (2016) Gazulla et al. (2017)

Vera et al. (2012) Uerlings et al. (2016) Cacho et al. (2016) Charbonnier et al. (1986) Böhm et al. (2016) (continued)

1206

R. Créac’hcadec

Table 1 (continued) Acronym Raman

Name explanation Raman spectroscopy

RBS TOFSIMS TXRF

Rutherford backscattering spectrometry Time-of-flight secondary ion mass spectrometry Total-reflection X-ray fluorescence spectroscopy X-ray photoelectron spectroscopy/ electron spectroscopy for chemical analysis X-ray diffraction X-ray fluorescence

XPS/ ESCA XRD XRF

XRR X-ray reflectivity Mechanical observations AT Acoustic emission testing 2D/3D –DIC

Two- or three-dimensional digital image correlation

DVC

Digital volume correlation Nanoindentation Ultrasonic measurement

References Cardell et Guerra (2016), Davies et al. (2009), and Zhao et al. (2016) – Kadiyala et al. (2016) Kanrar et al. (2014) Choudhury et Jones (2006) and Kadiyala et al. (2016) Tsunekawa et al. (2011) Aldrich-Smith et al. (2005) and Procaccini et al. (2011) Yeom et al. (2016) Croccolo et Cuppini (2009) and Magalhães et de Moura (2005) Créac’hcadec et al. (2008), Créac’hcadec et al. (2015), and Maurice (2012) Bele et al. (2017) and Bouterf et al. (2017) Davies et al. (2009) and Pecnik et al. (2015) Ghiassi et al. (2014)

Fig. 10 Fracture patterns of the adhesive in bonded assemblies for tensile/compression-shear tests: (a) tensile test, (b) shear test, (c) traction-shear test, (d) compression-shear test (Créac’hcadec et al. 2014)

as a V-shaped line at one of the ends of the sample (Fig. 10a). The analysis of the patterns represented in Fig. 10b shows clearly rhombus shapes. Small streaks are also observed associated with a failure at different levels in the adhesive thickness when loaded under shear. Under tensile-shear loads (Fig.10c), the failure mode is quite similar to the case of tensile loadings; the tensile part of the load seems to govern the failure mode. Moreover the tangential displacement at failure is lower

43

Techniques for Post-fracture Analysis

1207

under tensile-shear loads than under shear loads. Fracture patterns under compression-shear (Fig. 10d) show a rough surface where streaks are more pronounced and denser than during the shear loading situation. The fracture path is not straight and very irregular as the adhesive has been ripped out locally associated with the compression part of the load.

Adhesive/Cohesive Mixed-Mode Failure Surface The load path also affects the type of failure, i.e., adhesive failure close to the interface or cohesive failure in the adhesive. Arnaud et al. (2015) made similar modified Arcan tests for a bicomponent epoxy adhesive. The main part of the work was to identify the influence of the water aging of bonded assemblies. The fracture surfaces shown in Fig. 11 present the initial mode of failure without aging. They observed that in tension, the failure was mainly cohesive for the unaged state. Then, it became partially cohesive during aging. But at the last aging time, it was again fully cohesive. This evolution seems to indicate that the interface was slightly affected by humid aging in tension. However, it was not clear enough to conclude on a real interfacial degradation. In the other loading directions, the failure is mainly adhesive and does not reveal any evolution during aging (Fig. 11b–d).

43.3.2 The Importance of Local Geometry on Adhesive Joint Failure One way to understand fracture surface is to know where the fracture initiated. A lot of works were made to understand the contribution of the free edges of bonded assemblies. These zones are mainly the reasons of failure in adhesive joints. Karachalios et al. (2013) explained the failure process of single-lap shear tests with the localization of damage initiation at the extremity of the substrate in the middle of the test specimen. The failure process then develops on each side of bonded joint to finally reach to interface (Fig. 12a). This analysis was obtained thanks to the fracture surface observed (Fig. 12b). This study was carried out with ductile substrates and adhesives.

Fig. 11 Fracture profiles of the modified Arcan test during seawater aging; results are presented for the initial fracture surfaces (Arnaud et al. 2015): (a) tensile test, (b) shear test, (c) traction-shear test, (d) compression-shear test

1208

R. Créac’hcadec

a

Location of damage initiation

Top adherend

I

Bottom adherend

Damage propagation inside the overlap II

Direction of damage propagation

Bottom adherend

Top adherend’s edge ji

b

Adhesive stress whitening

Curved damage front

Width of joint

Loaded adherend end

Hackle region

Z

Y X

Overlap length

Fig. 12 Failure mechanism in single-lap shear test: (a) Description of failure mechanisms for ductile adhesives, (b) failure pattern of a 60 mm overlap with AV 119 adhesive (Karachalios et al. 2013)

Understanding the contribution of the free edge is very important. To analyze these geometries, finite element methods are very useful and used. Cognard et al. (2011b) analyzed the influence of some geometrical parameters on the stress concentrations in the case of the single-lap joints (Fig. 13). The stress concentrations can contribute to fracture initiation in the adhesive joints and thus can lead to an incorrect analysis of the adhesive behavior. Therefore, understanding the stress distribution in an adhesive joint can lead to improvements in adhesively bonded assemblies. This study was accomplished using finite element simulations under elastic assumption using refined meshes and using a pressure-dependent model, which accurately describe the elastic limit of the adhesive. It has been shown that geometries used

43

Techniques for Post-fracture Analysis

a

1209

b

y

y

Substrate

Substrate

d4

X

Adhesive d6

Substrate

x

B

A

Adhesive

d5

R2

Substrate

c

y Substrate

δ7

δ6

ρ

x

δ5

Adhesive

α

δ2

Substrate

Fig. 13 Influence of local geometry on the strength at failure (“o” zone of maximal equivalent stress in the adhesive under elastic assumption): (a) straight geometry, (b) geometry with a chamfer, (c) geometry with beaks (Cognard et al. 2011b)

in various studies in order to reduce the influence of edge effects can generate quite large stress gradients close to the adhesive-free edges. Specific geometries can significantly limit the influence of edge effects .The use of substrates with beaks and a cleaning of the free edges of the adhesive allow a large reduction in the stress state close to the free edges of the adhesive and thus allow an optimization of the maximum transmitted load in single-lap joints. The numerical results underline that the overlap length and the adhesive thickness also have an influence on the stress distribution in the joint. Moreover, it has been shown that significant evolution of the stress can exist throughout the thickness of the adhesive; therefore, simplified methods that roughly analyze the average stress state through the joint can overestimate the maximal transmitted load of single lap joints. Thus, refined analyses of the stress distribution in the adhesive are necessary to obtain precise design of adhesively bonded assemblies. To sum up, it is necessary to take special care to the local geometry: the principle proposed (Fig 14a) and how it is manufactured (Fig. 14b–d).

43.4

Microscopic Optical Observations

43.4.1 One-Dimension Roughness Mechanical adhesion is mainly explained by the surface topology. Thanks to a rugosimeter Arnaud (2014) determined the parameters of a one-dimension roughness, parameters which favor in particular the mechanical attachment, in order to be able to reuse them during future coating applications. Thus samples of aluminum

1210

R. Créac’hcadec

Fig. 14 Geometry of the free edge of modified Arcan test specimen to reduce the edge effects with beaks: (a) definition of the local geometry (Cognard 2008), (b) cross view for an adhesive thickness of 2e = 0.1 mm, (c) cross view for an adhesive thickness of 2e = 0.2 mm, (d) cross view for an adhesive thickness of 2e = 1 mm (Cognard et al. 2010)

Fig. 15 Analysis of a substrate with a rugosimeter: (a) measuring device and (b) roughness profile generated after passing the rugosimeter along a line (Arnaud 2014)

2017 were studied, in the form of bars of 10 mm thickness and surface 10
100 mm2. A roughness profile was obtained by passing the roughness tester on a straight line of the substrate (Fig. 15b). For better precision, it was necessary to measure several lines. The results are given in Table 2. The mean gap, defined as the arithmetic mean of the absolute values of the distances between peaks and troughs, noted Ra, and the total gap, defined as the most important distance between the highest peak and the deepest trough, noted Rt, measured were, respectively, 0.455 μm and 3.46 μm. These two parameters are the most influential in terms of roughness.

43.4.2 Scanning Electron Microscopy The scanning electron microscope is a tool able to observe surfaces from 10 to 0.1 μm. Figure 16a presents an example of a scanning electron microscope (SEM) study of an Arcan aluminum-adhesive-aluminum-modified assembly. An overall view of the adhesive shows the presence of fillers or porosities in the adhesive.

43

Techniques for Post-fracture Analysis

1211

Table 2 Roughness parameters for an untreated substrate (Arnaud 2014) Line Ra (μm) Rt (μm)

L1 0.489

L2 0.546

L3 0.491

L4 0.395

L5 0.427

L6 0.383

Mean value 0.455

Standard deviation 0.064

3.86

3.99

3.89

2.8

2.65

3.54

3.46

0.587

Fig. 16 Scanning electron microscopy of an aluminum-epoxy-bonded assembly: (a) overview, (b), (c) zooms of some details (Arnaud 2014)

Two different patterns are distinguished, surrounded in Fig. 16a. A zoom is made on these details in Fig. 16b, c. Thanks to these observations, we are able to confirm the presence of charges in the adhesive and to measure them. The diameter of these beads announced by the supplier is 0.13 mm, which seems consistent with the observations of microscopy. Scanning electron microscopy is generally completed with addition measurement systems like energy-dispersive X-ray spectroscopy; one would be able to distinguish a void from a sphere, for example. A list of these systems is detailed in Table 1.

1212

R. Créac’hcadec

Fig. 17 TEM micrograph of the interphase: evidence of structural defects (Beuguel et al. 2017)

More detailed information can be obtained. This is the case of the work of Beuguel et al. (2017). They observed decohesion between a polypropylene matrix and a polyamide nodule. Figure 17 was obtained with a higher resolution thanks to a transmission electron microscope.

43.4.3 Nanoindentation The discontinuity between two elements, the substrate and the adhesive or the matrix and the nodules, is a source of questions. Mechanicians are very interested in having local information on the mechanical properties. Is it an interface? Is it an interphase? What happens in this discontinuity? The use of nanoindentation provides local variations in the Young’s modulus. The principle is based on a load applied very locally thanks to a pyramid. The process of indentation is measured with a load-displacement curve. The posttreatment gives then the reaction of the material indented. Davies et al. (2009) used this technique trying to approach a variation of the Young’s modulus between an aluminum substrate and an epoxy adhesive along the geometric interface. In Fig. 18b zero defines the origin of this interface. For a positive abscissa, the values are defined for the adhesive. For a negative abscissa, this is the substrate that is measured. Several measurements were made for three adhesive thicknesses: 0.33 mm, 0.65 mm, and 1.23 mm. They conclude that if an interaction between the substrate and the adhesive exists for the materials considered, this zone is very thin around two microns of thickness.

43.4.4 Atomic Force Microscopy The atomic force microscopy is able to provide information up to 10 nm. The measurement principle is based on the flexion of a beam which is moved over a

43

Techniques for Post-fracture Analysis

1213

b 100

Joint thickness (mm) 0.33 0.65 1.23

Modulus (Mpa) 2.45 (0.11) 2.63 (0.19) 2.71 (0.15)

Modulus, GPa

0.33mm 1.23mm

a

10

1 −40 −30 −20 −10 0 10 Position, μm

20

30

40

Fig. 18 Nanoindentation of an epoxy adhesive: (a) measured Young’s modulus values, mean (standard deviation) of 10 values for each sample, measured at mid-thickness by nanoindentation; (b) across aluminum/adhesive interface, two joint thicknesses (Davies et al. 2009)

Fig. 19 AFM measurements (Critchlow et al. 2006): (a) AFM image of hand-polished nickel tooling (5
5 μm area). Vertical scale 60 nm; (b) AFM image of abraded nickel tooling, vertical scale approximately 0.3 μm

surface. The interaction between the tips of the beam induces a deflection of the beam which is measured. This interaction permits to define the topology of the measured surface. This is the first mean to use AFM. The second principle is based on the recording of the load necessary to deflect the beam. Thanks to this load measured, it is possible to obtain a map defining the local mechanical properties of the surface. The measurement process is quite long but gives a lot of information for small surfaces. Figure 19 gives results coming from the work of Critchlow et al. (2006). They analyzed the difference between hand-polished and abraded surfaces.

43.4.5 Confocal Microscopy The confocal microscope is an optical microscope able to measure information on a surface with a small depth of field. As a consequence, the volume studied has a maximal depth of 350–400 μm. To observe structures through this volume, the

1214

R. Créac’hcadec

a

b

0 4 8 12 Z (μm) 16 20

0 5 10 15 20 25 30 Y (μm) 35 40 45 50

c

0 5

30 35 20 25 X (μm) 10 15

40 45

50

Fig. 20 Confocal microscopy: (a) confocal micrograph of PP/PA6 blend: 3D observation of nodular morphology (Beuguel et al. 2017), (b) 3D measures of a zinc coating (Legendre et al. 2016), c roughness profile obtained along the white cross line (b)

material should be translucent. Beuguel et al. (2017) used this technique to observe the nodular morphology of a PP/PA6. Figure 20a defines the volume studied. The nodules are in red. It is possible to observe the variation of their dimensions, their distribution in a 3D space. Legendre et al. (2016) also used this technique to have a 3D topology of a steel sheet surface with a zinc coating. As previously shown in this chapter for the one-dimension roughness, it is possible to extract from this geometry the profile of the surface along a line (Fig. 20b, c).

43.4.6 X-Ray Microtomography In recent years, the development of measurement techniques has evolved toward three dimensions. The microtomography principle is based on the X-ray method. A specimen is lighted with an X-ray source. The image obtained is recorded thanks to a two-dimension screen placed behind the specimen. To obtain the third dimension, the specimen is placed on a rotary vertical axis. Between each measurement, a step of rotation is imposed. All these steps describe a complete turn. The whole images are then post-treated to build a three-dimension volume. Maurice (2012) used this technique to measure the voids in an aeronautical adhesive film. In this industrial field, the process imposes to use a pressure chamber, i.e., an autoclave, to reduce

43

Techniques for Post-fracture Analysis

1215

Fig. 21 Microtomography analysis of an aeronautical adhesive: (a) 3D tomography measurement, (b) SEM cross view, (c) measurement of porosity by microtomography (Maurice 2012)

porosity in the assembly. It is slightly dependent on the pressure and temperature cycles. Figure 21 presents these results. Large voids were observed. The size of these defects is relatively important: their height is similar in scale to the adhesive thickness, and their length can be two or three times larger. Microtomography observation, performed on a sample of about 10 mm height, width, and length, confirmed this observation, and numerical post-processing of the results using myVGL™ software enables the void ratio to be measured as, at least, 20% in volume.

43.5

Physicochemical Analysis

43.5.1 Scanning Electron Microscopy with Energy-Dispersive X-Ray Spectroscopy To understand X-ray spectroscopy, it is firstly necessary to understand the term spectroscopy. It can be defined as an experimental study to decompose in a succession of elementary data a complex and physical phenomenon. The set of these elementary data is called the spectrum. The data can be an energy, a wavelength, a frequency, etc. An energy-dispersive X-ray spectroscopy gives information about the chemical composition of the analyzed material. The impact of the electron beam on the sample produces X-rays which are characteristic of the elements present on the sample. Figure 22 shows an analysis of the particles described in red in Fig. 16. The lines of the spectrum are specific for an atom. Figure 22 defines lines for carbon (C), oxygen (O), aluminum (Al), silicon (Si), and gold (Au). Carbon (C) and oxygen (O) are detectable. They correspond to the polymer material detected by the EDX beam behind the bead. The fillers are therefore glass beads, and very little porosity is detectable. The gold comes from the coating made before the analysis to make the samples conductive (Arnaud 2014).

1216

R. Créac’hcadec C

2500

2000

1500

1000

Au Au

500 Si

O Al

Au Kev

0 0

9.830

Fig. 22 EDX analysis of two particles in an adhesive joint in red (Fig. 16) (Arnaud 2014)

The EDX spectroscopy is often coupled with a SEM in order to have the distribution of the atom in a representative surface. Arnaud (2014) studied with this method a coating based on molybdenum disulfide MoS2 (Fig. 23). Sulfur and molybdenum are present in good atomic mass ratio (Fig. 23c). In addition to the aluminum 2017, alloy made of aluminum and copper, there are elements of silicon, magnesium, oxygen, and carbon which must constitute the remainder of the coating. The EDX analysis in Fig. 23 confirms that sulfur (e) and molybdenum (f) are located in the same places and are very complementary to aluminum (d), constituting the majority of the coating. The position of the silicon appears to be an inclusion or pollution. The random position of the other species, carbon (b), oxygen (c), and magnesium (d) suggests their presence in the coating at lower amounts, which is confirmed by the percentages by mass. In this context, it is necessary to make an important remark about fracture analysis. Pollution of the broken specimen must be fought. Particular attention should be given to samples just after their failure. Thanks to the progress of the analysis methods and the increase of their sensitivity, it is now possible to observe and detect very little particles. These particles can come from the test specimen but also from a contamination of the environment. In order not to be orientated in a wrong way, no pollution of the test samples must be satisfied.

43.5.2 Scanning Electron Microscopy with Raman Spectroscopy When a well-known monochromatic ray light is applied to a material, the reflected signal is out of phase due to an exchange of energy between the ray light and the environment. This effect is called the Raman effect. The spectroscopy analysis of the reflected light gives information on the material itself. It is then possible to identify

43

Techniques for Post-fracture Analysis

1217

Fig. 23 SEM observations of an aluminum 2017 substrate treated with a molybdenum disulfide coating: (a) zoom
100, (b) zoom
1,000, (c) EDX study of the composition of the zone with a zoom
1,000, (d) aluminum, (e) sulfur, (f) molybdenum (Arnaud 2014)

molecules, chemical bonds, their proportions, and their structure. The method used by Raman analysis produces better spatial resolution than the attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectroscopy technique and makes it possible to analyze smaller samples. Davies et al. (2009) tried to identify modification of the interface between an epoxy adhesive and an aluminum substrate. Across the interface (Fig. 24a), they observed local spectra (Fig. 24b). These spectra allow local variations in molecular species to be detected. Analysis of these results in terms of molecular groups is

1218

a

−50 −40 −30 −20

Y (mm)

Fig. 24 (a) Optical view of the cross section analyzed by Raman scattering (100
objective lens in white light). Substrates (white) at top and bottom of photo. Each black circle corresponds to a Raman spectrum recorded, in the adhesive, (b) three Raman spectra along the cross section (Davies et al. 2009)

R. Créac’hcadec

−10 0 10 20 30 40 50 −60

−40

b 9000

−20

0 20 X (mm)

40

60

822 809

Intensity (a.u.)

1112

6000

629 638

1105

670

915

1180 12311297 1248

49.8 μm

3000

∼1450

1608

1514 1660 1670

26.2 μm

0

2.6 μm

800

1200 Raman Shift (cm-1)

1600

complicated, as this was a formulated commercial adhesive. The spectra (Fig. 24b) point out the chemical bonds: C=C, C–O, C–H, CH–OH, and C6H6–H. Each Raman shift is representative of a chemical bond. The analysis was completed in their work with other analysis methods. It was the crossing of all the method used that can propose scientific structured results. More information can be found in their work (Davies et al.).

43.5.3 X-Ray Photoelectron Spectroscopy Whereas X-ray Raman spectroscopy applies monochromatic ray light on a surface, photoelectron spectroscopy applies an X-ray. The depth impacted by the X-ray radiation is up to 5 nm. It enables to analyze the chemical composition and the chemical bonds of the surface. For example, Kadiyala et al. (2016) analyzed joint

43

Techniques for Post-fracture Analysis

1219

Fig. 25 (a) XPS survey of neat and fractured joint surfaces of PI; (b) The C1s core-level XPS spectra for joint surfaces fractured at 25  C (Kadiyala et al. 2016)

interface fractures of a thermoplastic semicrystalline polyimide for different temperatures: 25  C, 150  C, and 225  C. As Fig. 25b points out, the type of chemical bonds and their quantities can be identified.

43.5.4 Differential Scanning Calorimetry The curing process of an adhesive is analyzed by solving the energy equation (Devaux et al. 2015): ρCp

@T @α ¼ ∇ðλ∇T Þ þ ϕr , with ϕr ¼ ρΔH r @t @t

where ρ is the density of the adhesive, Cp the specific heat, λ the conductivity of the adhesive, T the temperature, Hr the total heat of the polymerization reaction, and α the degree of cure. ϕr is the heat flux produced by the polymerization reaction. Two main phenomena are involved during adhesive curing: polymerization and cross-linking. There is a competition between these two events during the curing process. For a bicomponent epoxy adhesive, the uncured resin is mixed with a hardener according to a ratio recommended by the manufacturer in order to initiate the curing process. The chemical reaction within the mixture proceeds, and polymer chains cross-link to each other. As a result, molecular weight increases, causing an increase in viscosity. The cross-linking is an exothermic reaction; thus, the formation of the polymer network generates heat which results in a temperature increase. The molecular weight continues to increase until the gel time. It is the time for which the adhesive transforms from a liquid phase to a solid phase. This transition is also called the gel point. From this point, the molecular weight increases until forming a fully cross-linked polymer network. As a consequence, it is necessary to have reliable

1220

R. Créac’hcadec

information on the curing process. The differential scanning calorimetry can provide such information.

Kinetic Analysis As an example, Devaux et al. (2015) studied Hysol EA-9321 adhesive from Henkel, a two-component thixotropic, solvent-free epoxy-based resin. The adhesive is prepared using a mixing ratio of 2:1 by weight of the respective constituents (resin and hardener). According to the manufacturer’s data sheets, a specimen cured during 1 h at 82  C is completely cured (curing degree α = 1), and the glass transition temperature of specimens cured during 5–7 days at 25  C is 88  C. The heat released during the cure reaction was detected by using a heat-flux differential scanning calorimeter connected to a thermal analyzer. This calorimeter is made of two steel pans. One of these is empty, and the other one contains sample weighting between 5 and 10 mg. At first, dynamic and isothermal scans were carried to determine kinetic parameters of each one. Then, both kinds of scans were conducted to simulate few steps curing cycles. Dynamic scans were carried out in the temperature range of 25–200  C at constant heating rates of 5, 10, 15, and 20  C/min. Isothermal scans were run at temperatures ranging from 25 to 100  C. Equilibrium at the target isothermal temperature was reached in the sample holder with a rate of 20  C/min. This rate was chosen sufficiently fast to prevent the reaction between resin and hardener before the start of isothermal scan. Coupled scans were made in the same way as previously, i.e., a dynamic scan was conducted until the required isothermal temperature, and then a second dynamic scan was conducted until the second required isothermal temperature. For each scan, a second heating run on the same sample under the same conditions was carried in order to define the baseline along which the curve heat flow versus time is integrated. The DSC curves in dynamic analysis are shown in Fig. 26. The shape of the exotherm was heating dependent. The value of the total heat released was determined by integrating heat flow versus time under the exotherm along a straight baseline. This heat of reaction, ΔHT, was independent of the heating rate (Table 3). Figure 27 shows the heat flow released versus time for isothermal scans. The maximum heat flow is reached faster for high curing temperatures, due to the acceleration of the reaction between resin and hardener. Table 4 presents the heat of reaction released during isothermal scanning. ΔHISO is calculated by integrating the area under the heat flow versus time curves along a straight baseline, defined as previously. The curing degree, α, is based on isothermal and dynamic scans and is calculated as ΔHt/ΔHT; ΔHt is the heat of reaction released during isothermal scan at time t and ΔHT the total enthalpy of reaction calculated during dynamic scans. In this way, intermediate curing states of the adhesive as a function of time over isothermal temperatures were determined.

43

Techniques for Post-fracture Analysis

1221

Fig. 26 Dynamic scanning: heat flow at different heating rates as function of (a) time and (b) temperature (Devaux et al. 2015)

Table 3 Total heat of reaction at different heating rates (Devaux et al. 2015) Parameter ΔHT (J/g)

dT/dt ( C/min) 5 334.1

Fig. 27 Heat flow at different isothermal temperatures (Devaux et al. 2015)

10 351.03

15 326.66

0.0018

35°C 45°C 60°C 70°C

0.0016 0.0014 Heat Flow (W/g)

20 353.1

82°C

0.0012 0.001 0.0008 0.0006 0.0004 0.0002 0 0.00

0.20

0.40

0.60

Curing degree, α (−)

0.80

1.00

1222

R. Créac’hcadec

Table 4 Isothermal scanning results (Devaux et al. 2015) Parameter ΔHISO (J/g) α () t

Tcure ( C) 35 130.73 0.38 4 h25

45 202.07 0.59 3 h33

60 240.03 0.70 2 h40

70 310.03 0.91 1 h30

82 341.18 1.0 46 min

Temperature (°C) −50

0

50

Heat flow (mW)

Sea water

100 Sea water + drying

to the woven

Reference

Fig. 28 DSC results: influence of the accelerated aging on an epoxy adhesive (Arnaud 2014)

Analysis of the Glass Transition Temperature (Tg) The evolution of Tg provides information on the physical phenomena affecting the polymer studied. Thus, Arnaud (2014) studied the evolution of an accelerating aging under temperature and seawater absorption. Figure 28 presents the evolution of the heat flow versus temperature for a rate of 10  C/min. Four types of specimens were tested. The reference submitted to a cold curing cycle. The Tg measured was 55  C defined by the inflection point on Fig. 28 (“reference curve”). A second specimen was cold cured and post-cured at 40  C during 1 month. Tg increased to 71  C (Fig. 28; “to the woven” curve). It was fully cured like the study presented by Devaux et al. (2015). The third curve defines a cold-cured adhesive and aged in seawater at a temperature of 40  C (Fig. 28; “seawater” curve). The Tg obtained increased to 63  C indicating the post-cured of the adhesive during aging. A fourth specimen was cold cured aged in seawater at a temperature of 40  C and dried in an oven at 40  C during 1 month (Fig. 28; “seawater + drying” curve). The Tg obtained was 70  C. This specimen points out the reversible phenomenon of the seawater aging. The evolution of the Tg measured by DSC gives information on the physical phenomena.

43

Techniques for Post-fracture Analysis

1223

Fig. 29 TGA analysis (Zhao et al. 2016)

43.5.5 Thermogravimetric Analysis The thermogravimetric analysis consists on measuring with precision the weight of a specimen submitted to a thermal loading under a controlled atmosphere with air, azote, oxygen, etc. The measured variations transcribe the chemical reactions that occurred. The principle is similar to a measured combustion. At the end, residues are generally analyzed to quantify nonorganic charges. To explain this method, it is possible to quote the work made by Zhao et al. (2016). They analyzed with thermogravimetric analysis the composition of an adhesive made of two components: epoxy and a flame retardant named bisphenol-A bridged penta(anilino) cyclotriphosphazene (BPA-BPP). The results are plotted in Fig. 29. The evolution of weight was described regarding the temperature under nitrogen atmosphere. Different compositions were studied, starting with the BPA-BPP and the epoxy themselves. The weight of BPA-BPP started to decrease around 250  C. At 700  C, the residual weight was 60%. It means that there were residual nonorganic components. For the epoxy analysis, two main variations occurred (Fig. 29b) around 400  C and 600  C. Each peak corresponds to a chemical reaction. Then they studied the mixture.

43.5.6 Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) Spectroscopy Technique As the previous spectroscopies presented before, the attenuated total reflectanceFourier transform infrared (ATR-FTIR) spectroscopy technique is based on the infrared light on the material. The returned signal is then measured. The difference is linked to the properties of the material. Each peak measured is specific to a chemical bond. It is also possible to define their quantities. Gorassini et al. (2016)

1224

R. Créac’hcadec 1022

Abs

894

1155 1367 1314

3336 2893

1641 b a Cellophane

4000

3600

3200

2800

2400

2000

1800 cm-1

1600

1400

1200

1000

800

550

Fig. 30 ATR-FTIR spectra of cellophane, backing side of unaged (a) and artificially aged (b) tape C (Gorassini et al. 2016)

studied the composition of an old pressure-sensitive adhesive, used to assemble pages in very old books, in order to recover and understand the initial chemical properties of the adhesive and their modifications due to aging. Three adhesive were tested: a cellophane available in literature and a commercial adhesive unaged and aged. They confirmed that the adhesive used was mainly made of cellulosze acetate (Fig. 30).

43.6

Mechanical Observations

The following points are not often defined as a “post-fracture analysis” because these measurements are generally carried out in the laboratory during an instrumented test. However, it is very important for mechanical engineers to establish the relation between the loads applied, the deformations of the bonded assemblies, and their fracture surfaces. As the experimental methods need posttreatment after the test, they can be defined as “post-fracture analysis.” Moreover, these methods can bring a lot of information to understand the fracture of adhesives.

43.6.1 3D Digital Image Correlation 3D digital image correlation systems are able to measure the displacement of a surface in three dimensions (Cognard et al. 2005; Créac’hcadec et al. 2008; Maurice 2012). It is made of a pair of cameras and a lighting system that take pictures at a defined time rate. The frames are then recording with different signals like the tensile

43

Techniques for Post-fracture Analysis

1225

Fig. 31 3D digital image correlation system: (a) Modified Arcan experimental device, (b) image recorded and post-analyzed with the relative displacement of both sides of the adhesive joint in the out of plane y-direction DN, (c) tensile load, relative displacement DN curve (Maurice 2012).

load of the test machine. Figure 31 presents an example of a modified Arcan test measured with this type of system. Figure 31b presents a post-treated frame with the adhesive joint area analyzed composted of the adhesive and two aluminum substrates. The correlation system provides a kinematic field for the entire picture. Here the relative displacement of both sides of the adhesive in the y-direction is plotted. The resolution of this system is around 1/100 pixels. It depends on the scale factor to convert it in millimeters. As a result, Fig. 31c plots a load-displacement curve in blue. It is then possible to develop law behavior to explain such observations (Créac’hcadec 2008). Most of all, Maurice (2012) improved the posttreatment

1226

R. Créac’hcadec

Fig. 32 Influence of strain rate and behavior under cyclic loads of bonded assembly under shear loads: (a) strain rate effect and (b) behavior under cyclic loads (Maurice 2012)

method to explain the cleavage along the x-direction that can be observed if the alignment of the Arcan device is not well done. This technique is very interesting to understand the type of fracture. It is then possible to measure different phenomenon behaviors of the adhesive joints, like strain rate effect or cyclic behaviors (Fig. 32). For fatigue test, an improvement is necessary to have real-time measurements (Créac’hcadec et al. 2015). This method is very interesting and can be used for local measurements. Davies et al. (2009) pointed out the linear evolution of the shear strain for a thick epoxy adhesive. No specific behavior close to the interface was observed (Fig. 33).

43.6.2 Digital Volume Correlation The digital image correlation gives information on the surface of an observed specimen. The next question is then what happens in the material. That’s the reason why digital volume correlation is developing a lot. The principle is similar to the microtomography principle described before. However, a 3D reconstruction is made for each step of loading like the 3D image correlation system. It is then possible to have in three dimensions the evolution of the strains versus loading path. Boufert et al. (2017) studied the microstructure of a core and observed with a resolution of 12 μm. The volume is defined in Fig. 34a. They applied a load thanks to the indentation sphere (Fig. 34b). They post-treated their recorded frame and were able to plot a stress field (Fig. 34c). This method is very useful to have information in the assembly. This is not very developed in the adhesive field, but in the coming years, it will probably bring to the adhesion community very important data.

FT (kN)

43

Techniques for Post-fracture Analysis 18 16 14 12 10 8 6 4 2 0

1227

(v) (iv) (iii)

(ii) (i)

0

0.2

0.4 0.6 DT (mm)

(iii)

0.8 (i)

(ii)

(iv)

(v)

Fig. 33 2D digital image correlation for a pure shear deformation of a thick epoxy adhesive joint (Davies et al. 2009)

Fig. 34 Digital volume correlation: (a) 3D rendering of the microstructure of the core material studied herein and observed at a resolution of 12 μm. The size of the volume is 4.8
4.8
3.2 mm3; (b) section of the microstructure with the indenter (in white) that applied the load, (c) stress field extracted from the DVC measurements (Bouterf et al. 2017)

1228

R. Créac’hcadec

Fig. 35 Ultrasonic images (C-scans) of a sample in immersion in Harrison solution: (a) without clear coat after 331 h (b) with clear coat after 315 h of immersion in Harrison solution. White scale bars: 5 mm. The arrows mark the initial defects (holes of 0.5 mm diameter) (Oehler et al. 2012)

43.6.3 Ultrasonic Testing Nondestructive tests are very interesting to bring information on adhesion. This is the case for ultrasonic tests. The principle is described in the chapter dedicated to NDT methods (see ▶ Chap. 42, “Nondestructive Testing”). An ultrasonic wave is generated to a surface. The return signal is measured. The difference is linked to the material itself. Oehler et al. (2012) worked on decohesion of an interface subjected to an immersion in Harrison solution. A hole with a diameter of 0.5 mm was made in the coating. A good coating and a bad one were tested. Decohesion was observed with ultrasonic scans (Fig. 35) for the clear coat. This method was used by Maurice (2012) for composite single-lap shear tests (Fig. 36). Tests were performed with a tensile testing machine. The test was stopped at different load levels. Ultrasonic measurements were performed at each step of loading. The overlap was observed. At the corner, a fracture initiated linked to the edge effects and then propagated to the center. This is particularly visible in Fig. 36b. After this propagation, the fracture of the whole specimen occurred.

43.7

Conclusions

Solving the equation of adhesion needs information coming from multi-physics and multi-scale analysis. 1. Fracture analyses are a huge source of information. This information comes from various supports: microscopy, compositional analysis/material identifications, and mechanical observations. 2. Each means of analysis should not be taken separately. It is the intersection of these different types of analysis that constructs the failure understanding.

43

Techniques for Post-fracture Analysis

1229

Fig. 36 US scan for a composite single-lap shear test sample: (a) geometry of the specimen, (b) example of US scans obtained at different loads for different specimens; the scale goes from 0.3 (red) to 4 dB (blue) of signal loss (Maurice 2012)

3. In this chapter, the mainly experimental devices are presented, but this list is not exhaustive. 4. The development of measurement methods is increasing with precision and sensitivity. The today’s means are certainly not those of tomorrow’s ones. 5. Having a good picture or a good spectrum needs good operators to prepare the sample, not to pollute them, to master the observation device, etc. 6. Moreover, good pictures are not synonymous of good interpretations. To have a good explanation, it is necessary to have a long experience, especially in these observation techniques. The failure comprehension is then the intersection of specialists of each domain area. 7. Scientists or industrials that work on adhesion should be able to interact with all these domains. This is not so simple due to the initial training of each one. Acknowledgements This work was realized using different results coming from PhD theses I had the pleasure to supervise. So I really want to thank N. Arnaud, O. Devaux, J. Legendre, J. Maurice, for the work they performed. I also want to thanks the different research collaborators (industrials and academic) for their help and their financial support. I also want to thank Philippe Elies and Gérard Sinquin (Université de Bretagne Occidentale) for the measurements they performed in the Pimm-DRX technology platform: Imaging, Microscopy and x-ray diffraction.

References Aguiar TR, Andre CB, Arrais CAG, Bedran-Russo AK, Giannini M (2012) Micromorphology of resin–dentin interfaces using self-adhesive and conventional resin cements: a confocal laser and scanning electron microscope analysis. Int J Adhes Adhes 38:69–74. https://doi.org/10.1016/j. ijadhadh.2012.05.009 Aldrich-Smith G, Jennett N, Housden J (2005) Adhesion of thin coatings – the VAMAS (TWA 22-2) interlaboratory exercise. Surf Coat Technol 197:336–344. https://doi.org/10.1016/j. surfcoat.2004.07.113

1230

R. Créac’hcadec

Arnaud N (2014) Analyse de l’effet du vieillissement en milieu humide sur le comportement mécanique d’adhésifs en assemblages sous sollicitations multiaxiales. Université de Bretagne Occidentale, Brest Arnaud N, Créac’hcadec R, Cognard JY, Davies P, Gac PYL (2015) Analysis of the moisture effect on the mechanical behaviour of an adhesively bonded joint under proportional multi-axial loads. J Adhes Sci Technol 29:2355–2380. https://doi.org/10.1080/01694243.2015.1067018 Bele E, Goel A, Pickering EG, Borstnar G, Katsamenis OL, Pierron F, Danas K, Deshpande VS (2017) Deformation mechanisms of idealised cermets under multi-axial loading. J Mech Phys Solids. https://doi.org/10.1016/j.jmps.2017.01.002 Beuguel Q, Ville J, Crepin-Leblond J, Mederic P, Aubry T (2017) Influence of clay mineral structure and polyamide polarity on the structural and morphological properties of clay polypropylene/polyamide nanocomposites. Appl Clay Sci 135:253–259. https://doi.org/10.1016/j. clay.2016.09.034 Böhm R, Hauptmann M, Pizzi A, Friedrich C, Laborie M-P (2016) The chemical, kinetic and mechanical characterization of tannin-based adhesives with different crosslinking systems. Int J Adhes Adhes 68:1–8. https://doi.org/10.1016/j.ijadhadh.2016.01.006 Bouterf A, Adrien J, Maire E, Brajer X, Hild F, Roux S (2017) Identification of the crushing behavior of brittle foam: from indentation to oedometric tests. J Mech Phys Solids 98:181–200. https://doi.org/10.1016/j.jmps.2016.09.011 Cacho JI, Campillo N, Viñas P, Hernández-Córdoba M (2016) Dispersive liquid-liquid microextraction for the determination of nitrophenols in soils by microvial insert large volume injection-gas chromatography–mass spectrometry. J Chromatogr A 1456:27–33. https://doi. org/10.1016/j.chroma.2016.05.098 Cardell C, Guerra I (2016) An overview of emerging hyphenated SEM-EDX and Raman spectroscopy systems: applications in life, environmental and materials sciences. TrAC Trends Anal Chem 77:156–166. https://doi.org/10.1016/j.trac.2015.12.001 Charbonnier M, Romand M, Roche A, Gaillard F (1986) Adhesive bonding of aerospace materials – surface characterization of metallic adherends. Int J Adhes Adhes 6:199–206. https://doi.org/ 10.1016/0143-7496(86)90006-0 Chen B, Dillard DA (2001) The effect of the T-stress on crack path selection in adhesively bonded joints. Int J Adhes Adhes 21:357–368. https://doi.org/10.1016/S0143-7496(01)00011-2 Choudhury T, Jones FR (2006) The interaction of silanes with amorphous silicon oxide surfaces. Int J Adhes Adhes Silane Coupling Agents Silane Coupling Agents 26:79–87. https://doi.org/ 10.1016/j.ijadhadh.2005.03.007 Cognard JY (2008) Numerical analysis of edge effects in adhesively-bonded assemblies application to the determination of the adhesive behaviour. Comput Struct 86:1704–1717. https://doi.org/ 10.1016/j.compstruc.2008.02.003 Cognard JY, Davies P, Gineste B, Sohier L (2005) Development of an improved adhesive test method for composite assembly design. Compos Sci Technol, JNC13-AMAC-Strasbourg 65:359–368. https://doi.org/10.1016/j.compscitech.2004.09.008 Cognard JY, Créac’hcadec R, Sohier L, Leguillon D (2010) Influence of adhesive thickness on the behaviour of bonded assemblies under shear loadings using a modified TAST fixture. Int J Adhes Adhes Spec Issue Joint Des 30:257–266. https://doi.org/10.1016/j.ijadhadh.2009.11.003 Cognard JY, Bourgeois M, Créac’hcadec R, Sohier L (2011a) Comparative study of the results of various experimental tests used for the analysis of the mechanical behaviour of an adhesive in a bonded joint. J Adhes Sci Technol 25:2857–2879. https://doi.org/10.1163/016942411X569345 Cognard JY, Créac’hcadec R, Maurice J (2011b) Numerical analysis of the stress distribution in single-lap shear tests under elastic assumption – application to the optimisation of the mechanical behaviour. Int J Adhes Adhes 31:715–724. https://doi.org/10.1016/j.ijadhadh.2011.07.001 Comyn J (1992) Practical surface analysis, 2nd edition: Volume 1: Auger and X-ray photoelectron spectroscopy Edited by: D. Briggs and M.P. Seah John Wiley & Sons, 1990, 657 + xiv pp. Int J Adhes Adhes 12:124. https://doi.org/10.1016/0143-7496(92)90036-U Coulaud M (2007) Rôle des interfaces et interphases dans les assemblages collés. PhD thesis, Institut National des Sciencs Appliquées de Lyon, Lyon

43

Techniques for Post-fracture Analysis

1231

Créac’hcadec R, Cognard JY, Heuzé T (2008) On modeling the non-linear behaviour of thin adhesive films in bonded assemblies with interface elements. J Adhes Sci Technol 22:1541–1563 Créac’hcadec R, Jamin G, Cognard JY, Jousset P (2014) Experimental analysis of the mechanical behaviour of a thick flexible adhesive under tensile/compression-shear loads. Int J Adhes Adhes 48:258–267. https://doi.org/10.1016/j.ijadhadh.2013.09.040 Créac’hcadec R, Sohier L, Cellard C, Gineste B (2015) A stress concentration-free bonded arcan tensile compression shear test specimen for the evaluation of adhesive mechanical response. Int J Adhes Adhes 61:81–92. https://doi.org/10.1016/j.ijadhadh.2015.05.003 Critchlow GW, Litchfield RE, Sutherland I, Grandy DB, Wilson S (2006) A review and comparative study of release coatings for optimised abhesion in resin transfer moulding applications. Int J Adhes Adhes 26:577–599. https://doi.org/10.1016/j.ijadhadh.2005.09.003 Croccolo D, Cuppini R (2009) Adhesive defect density estimation applying the acoustic emission technique. Int J Adhes Adhes 29:234–239. https://doi.org/10.1016/j.ijadhadh.2008.06.001 Davies P, Sohier L, Cognard J-Y, Bourmaud A, Choqueuse D, Rinnert E, Créac’hcadec R (2009) Influence of adhesive bond line thickness on joint strength. Int J Adhes Adhes 29:724–736. https://doi.org/10.1016/j.ijadhadh.2009.03.002 Devaux O, Créac’hcadec R, Cognard JY, Mathis K, Lavelle F (2015) FE simulation of the curing behavior of the epoxy adhesive Hysol EA-9321. Int J Adhes Adhes 60:31–46. https://doi.org/ 10.1016/j.ijadhadh.2015.03.005 Gazulla MF, Rodrigo M, Orduña M, Ventura MJ, Andreu C (2017) High precision measurement of silicon in naphthas by ICP-OES using isooctane as diluent. Talanta 164:563–569. https://doi. org/10.1016/j.talanta.2016.12.023 Ghiassi B, Verstrynge E, Lourenço PB, Oliveira DV (2014) Characterization of debonding in FRP-strengthened masonry using the acoustic emission technique. Eng Struct 66:24–34. https://doi.org/10.1016/j.engstruct.2014.01.050 Gorassini A, Adami G, Calvini P, Giacomello A (2016) ATR-FTIR characterization of old pressure sensitive adhesive tapes in historic papers. J Cult Herit 21:775–785. https://doi.org/10.1016/j. culher.2016.03.005 Haddadi F, Tsivoulas D (2016) Grain structure, texture and mechanical property evolution of automotive aluminium sheet during high power ultrasonic welding. Mater Charact 118:340–351. https://doi.org/10.1016/j.matchar.2016.06.004 Hase A, Wada M, Mishina H (2014) Scanning electron microscope observation study for identification of wear mechanism using acoustic emission technique. Tribol Int 72:51–57. https://doi. org/10.1016/j.triboint.2013.12.006 Heikkilä I, Eggertson C, Randelius M, Caddeo-Johansson S, Chasoglou D (2016) First experiences on characterization of surface oxide films in powder particles by Glow Discharge Optical Emission Spectroscopy (GD-OES). Metal Powder Report 71:261–264. https://doi.org/ 10.1016/j.mprp.2016.03.005 Kadiyala AK, Sharma M, Bijwe J (2016) Exploration of thermoplastic polyimide as high temperature adhesive and understanding the interfacial chemistry using XPS, ToF-SIMS and Raman spectroscopy. Mater Des 109:622–633. https://doi.org/10.1016/j.matdes.2016.07.108 Kanrar B, Sanyal K, Misra NL, Aggarwal SK (2014) Improvements in energy dispersive X-ray fluorescence detection limits with thin specimens deposited on thin transparent adhesive tape supports. Spectrochim Acta B At Spectrosc 101:130–133. https://doi.org/10.1016/j.sab.2014.07.018 Karachalios EF, Adams RD, da Silva LFM (2013) Single lap joints loaded in tension with high strength steel adherends. Int J Adhes Adhes 43:81–95. https://doi.org/10.1016/j.ijadhadh.2013.01.016 Koseki S, Inoue K, Usuki H (2016) Damage of physical vapor deposition coatings of cutting tools during alloy 718 turning. Precis Eng 44:41–54. https://doi.org/10.1016/j.precisioneng.2015.09.012 Legendre J, Jacquet D, Créac’hcadec R (2016) The role of steel plasticity on the failure pattern of single lap joints. Presented at the Euradh 2016 and Adhesion ‘16, 11th European Adhesion Conference and 13th International Triennial Conference on the science and Technology of Adhesion and Adhesives, Glasgow Liu Y (2013) “Tri-3D” electron microscopy tomography by FIB, SEM and TEM: application to polymer nanocomposites (PhD thesis). INSA de Lyon

1232

R. Créac’hcadec

Magalhães AG, de Moura MFSF (2005) Application of acoustic emission to study creep behaviour of composite bonded lap shear joints. NDT & E International 38:45–52. https://doi.org/10.1016/ j.ndteint.2004.06.005 Maurice J (2012) Characterization and modeling of the 3D elastic plastic behavior of structural adhesives films for aeronautical applications. PhD thesis, Université de Bretagne Occidentale, Brest Mine A, De Munck J, Van Ende A, Poitevin A, Matsumoto M, Yoshida Y, Kuboki T, Van Landuyt KL, Yatani H, Van Meerbeek B (2017) Limited interaction of a self-adhesive flowable composite with dentin/enamel characterized by TEM. Dent Mater 33:209–217. https://doi.org/ 10.1016/j.dental.2016.11.010 Oehler H, Alig I, Lellinger D, Bargmann M (2012) Failure modes in organic coatings studied by scanning acoustic microscopy. Prog Org Coat Coat Sci Int 2011 74:719–725. https://doi.org/ 10.1016/j.porgcoat.2011.09.017 Pecnik CM, Courty D, Muff D, Spolenak R (2015) Fracture toughness of esthetic dental coating systems by nanoindentation and FIB sectional analysis. J Mech Behav Biomed Mater 47:1–11. https://doi.org/10.1016/j.jmbbm.2015.03.006 Procaccini R, Ceré S, Pellice S (2011) Development and thermal evolution of silver clusters in hybrid organic–inorganic sol–gel coatings. Surf Coat Technol 205:5464–5469. https://doi.org/ 10.1016/j.surfcoat.2011.06.018 Tsunekawa S, Aoki Y, Habazaki H (2011) Two-step plasma electrolytic oxidation of Ti–15V–3Al–3Cr–3Sn for wear-resistant and adhesive coating. Surf Coat Technol 205:4732–4740. https://doi.org/10.1016/j.surfcoat.2011.04.060 Uerlings R, Matusch A, Weiskirchen R (2016) Reconstruction of laser ablation inductively coupled plasma mass spectrometry (LA-ICP-MS) spatial distribution images in Microsoft excel 2007. Int J Mass Spectrom 395:27–35. https://doi.org/10.1016/j.ijms.2015.11.010 Vallat MF (2013) Tests d’adhérence et pertinence des mesures, 17ème Journées d’Etudes sur l’Adhésion. JADH13, Aussois Vera P, Uliaque B, Canellas E, Escudero A, Nerín C (2012) Identification and quantification of odorous compounds from adhesives used in food packaging materials by headspace solid phase extraction and headspace solid phase microextraction coupled to gas chromatography–olfactometry–mass spectrometry. Anal Chim Acta 745:53–63. https://doi.org/10.1016/j.aca.2012.07.045 Yeom B, Jeong A, Lee J, Char K (2016) Enhancement of fracture toughness in organic/inorganic hybrid nanolaminates with ultrathin adhesive layers. Polymer 91:187–193. https://doi.org/ 10.1016/j.polymer.2016.03.078 Zhao B, Liang W-J, Wang J-S, Li F, Liu Y-Q (2016) Synthesis of a novel bridgedcyclotriphosphazene flame retardant and its application in epoxy resin. Polym Degrad Stab 133:162–173. https://doi.org/10.1016/j.polymdegradstab.2016.08.013 Zięba-Palus J, Nowińska S, Kowalski R (2016) Application of infrared spectroscopy and pyrolysis gas chromatography for characterisation of adhesive tapes. J Mol Struct Mol Mol Mater Biol Mol Syst Nanostruct A Collect Contrib Presented XIIIth Int Conf Mol Spectrosc 1126:232–239. https://doi.org/10.1016/j.molstruc.2015.11.050

Part IX Applications

Adhesively Bonded Joints in Aircraft Structures

44

L. John Hart-Smith

Contents 44.1 44.2 44.3 44.4 44.5

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesive Characteristics Required for Design and Analysis . . . . . . . . . . . . . . . . . . . . . . . . . Surface Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Design of Adhesively Bonded Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Additional Considerations for the Design and Analysis of Adhesively Bonded Joints in Fiber–Polymer Composite Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.6 Design Features Ensuring Durability of Bonded Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.7 Load Redistribution Around Flaws and Porosity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.8 Effects of Thermal Mismatch Between Adherends on Strength of Bonded Joints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.9 Inspection, Testing, and Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.10 Concerns About the Misplaced Emphasis of the Quality-Control Inspections for Bonded and Composite Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.11 Bonded Repairs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.12 Other Industry-Specific Factors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44.13 Examples of Use of Adhesive Bonding in Aircraft Structures . . . . . . . . . . . . . . . . . . . . . . . 44.14 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1236 1238 1243 1248 1253 1256 1260 1264 1265 1271 1276 1277 1278 1281 1282

Abstract

This chapter focuses on experience with adhesively bonded joints in aircraft structures. It covers their analysis (emphasizing the inherent nonuniformity of adhesive bond stresses), their design, some common pitfalls in past and current practices (the worst of which is the continued myth that adhesive bond load transfer is distributed uniformly over the bond area, so that all that is needed to

L. J. Hart-Smith (*) Boeing Company, Long Beach, CA, USA e-mail: [email protected]; [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_44

1235

1236

L. J. Hart-Smith

make stronger joints is more bond area), and the critical need for ensuring that the glue is stuck (not by after-the-fact monitoring, which is all but impossible and therefore unreliable, but by proper specifications and quality control during manufacture). It covers the load redistribution around defects and damage, showing that not all damage causes an immediate loss of strength (even though repairs may be necessary to protect against subsequent environmental degradation). It emphasizes a little-appreciated point that the adhesive bond, itself, must never be designed to be the weak link in the chain of strength, so that it cannot act as an instantly catastrophic fuse; the weak link must be in an adherend outside the joint for bonded designs to be reliable. This chapter also includes a discussion of the choice of appropriate bonding tool concept and the great benefits from sometimes using nontraditional tools. The key issue is shown to be that the governing tolerance is not that for the parts, or the tool, but for the thickness of the bond line. The real critical tolerance is on the order of 0.025 mm (0.001 in.), not
0.4–0.8 mm (0.016–0.032 in.). This requires a far more thorough coordination between design of the parts and the bond tool than has customarily been applied in the past. It is made abundantly clear that tighter and tighter tolerances on the parts and tools are not the way to go. One must rely on a drape-to-fit manufacturing philosophy, the benefits from which are shown to be enormous. The issue of inspection is discussed because a failure to understand what can work and what needs to be checked has led to far reaching adverse consequences. Misplaced faith in after-the-fact inspections is shown to have replaced the proper emphasis on proper process control during manufacture. It is shown to be impossible to restore 100% of bonded-joint strength after it has been manufactured incorrectly (or manufactured “correctly” in accordance with process specifications that simply do not work). And the futile search for the Holy Grail of being able to monitor degradation of bond strength during service has led to a ban on using this technology, in some instances, even when there is no other alternative approach that works. This chapter concludes with examples of the successful widespread application of adhesively bonded joints in aircraft primary structure.

44.1

Introduction

There are two basic classes of adhesive bonding in aeronautical structures. One is structural bonding, with epoxy, phenolic, or acrylic adhesives, that transfers load between members. The second is sealants that protect against corrosion at interfaces. The stiffnesses of these classes of polymers differ greatly, but the two basic needs are remarkably similar. The first is that the adhesive or sealant will stay stuck for the life of the structure, in all service and storage environments, while the second is that the adhesive will not fail even when the surrounding structure has been broken. Given that the mechanical properties of these polymers are orders of magnitude lower than those of the adherends they bond together, this second requirement may seem difficult to achieve. Nevertheless, it has been possible by distributing the load

44

Adhesively Bonded Joints in Aircraft Structures

1237

transfer as a shear load over a sufficiently large area and the use of sufficient steps for bonded joints between thick highly loaded adherends. The key is to protect the adhesive from direct or induced peel loads while using the shear capability of the adhesive to transfer the loads. This requires efficient joint designs, which are easily established for thin structures but which become progressively more complicated for thicker structures. The details of the joints are described here and in other chapters in the volume. Ensuring that the adhesive or sealant is stuck properly in the first place, and that it continues to adhere for the life of the structure, is actually the more challenging of the two needs because it is so misunderstood by far too many of those involved in this industry and others. The science of adhesion is vital to the success of bonded joints, but its importance is largely unrecognized today. The greatest difficulty in this regard is that success is completely dependent on correct surface preparation and processing, which cannot be verified by any of the standard nondestructive postinspections. Worse, while the correct techniques are known, many of the process specifications of past, and even current, applications of adhesive bonding violate these requirements. They continue to be used because the inappropriate methods do not always result in instantaneous separations between the components, so not everyone involved acknowledges that there is a problem. This doubt is reinforced because most of the inspection methods can detect only physical gaps between the members, not whether or not they are actually bonded together. The procurement system makes it difficult for military services to allocate costs that would be incurred to update repair manuals for products that are long out of production but still in service. There is also a concern, in a litigious society, that any change in specifications would cast doubt on the integrity of every part made under the older procedures, regardless of whether or not such concerns could be justified. Sadly, the consequence of not implementing improvements promptly has been a reluctance to apply the technology as widely as it ought to be because of the bad reputation created by every group of in-service failures. The bonding processes themselves are actually remarkably tolerant of quite wide deviations from nominal requirements, but when the specifications are inappropriate, or simply not followed, the consequences are very widespread. Worse, defects in mechanically fastened structures tend to be local, and 100% of the design strength can be restored by drilling out and replacing defective fasteners or even a few discrete parts. But processing errors in bonded and composite structures, once made, create situations in which no repair short of complete disassembly and remanufacture can restore 100% of the design strength and durability. It is not possible to inspect quality into bonded and composite structures afterward; they need to be made correctly in the first place. The cultural changes needed to ensure this are actually the greatest of all the obstacles to the more widespread use of bonded structures. Conversely, the durability of those bonded structures that were made appropriately is legendary, in regard to both fatigue and corrosion. The incentives for widespread use of bonded aerospace structures remain tremendous. This chapter is presented from the perspective of what actually happens, rather than what should happen, in the hope that it will encourage improvements to be

1238

L. J. Hart-Smith

made throughout the industry and regulatory authorities. Without these, it is unlikely that the full benefits of this technology will ever be attained because of the bad image created by the occasional lapses in quality and highly publicized consequences. Even so, it should be noted that, with the only exception of the well-known Aloha 737 incident in 1988, the kind of surface preparation and processing problems described here have resulted only in additional costs, which have inhibited the application of this technology, rather than created circumstances that made safety a concern. The contents of this chapter focus on what over 30 years of experience has been identified as the key issues that need to be understood in creating reliable bonded structures – issues that are rarely even mentioned in standard treatises on the subject, such as why it is imperative to not design to nominal loads but to ensure that the bonded joints will never be the weak link in the structure. This chapter addresses the following topics: Adhesive Characteristics Needed for Design and Analysis; Surface Preparation; Design of Adhesively Bonded Joints; Additional Considerations for the Design and Analysis of Adhesively Bonded Joints in Fiber–Polymer Composite Structures; Design Features Ensuring Durability of Bonded Joints; Load Redistribution Around Flaws and Porosity; Effects of Thermal Mismatch Between Adherends on Strength of Bonded Joints; Inspection, Testing, and Quality Control; Concerns About the Misplaced Emphasis of the QualityControl Inspections for Bonded and Composite Structures; Bonded Repairs; Other Industry-Specific Factors; and Examples of Use of Adhesive Bonding in Aircraft Structures.

44.2

Adhesive Characteristics Required for Design and Analysis

At the macro level, at which most bonded-joint analyses are made, the mechanical data needed are complete stress–strain curves in shear for a range in temperatures covering the operating environment. Typical structural adhesives are stronger and more brittle at low temperatures and weaker and more ductile at high temperatures than at room temperature, as shown in Fig. 1. Despite these great differences in bulk individual mechanical properties, the strength of structural joints (in which the adhesive strains are far from uniform) is far less sensitive to the test temperature than that of short-overlap test coupons (in which the adhesive strains are close to uniform, which is why they are used to generate the stress–strain curves). (This is explained in more detail below, in terms of elastic–plastic adhesive models (Hart-Smith 1973a, 1974).) The nonlinear behavior shown in Fig. 1 is not the result of plastic deformation akin to that for ductile metals. It can best be understood in the context of the strain invariants that eventually resulted in the development of the strain-invariant failure theory (SIFT) failure model for fiber–polymer composites. This had its foundation in earlier such analyses developed by Gosse (see Gosse and Christensen 2001) for the failure of homogeneous polymers, including adhesives constrained between much stiffer adherends. There are only two possible failure mechanisms available for any solid homogeneous material. One is dilatation which, in the context of adhesively

44

Adhesively Bonded Joints in Aircraft Structures

70

1239

10 (ksi)

60

−55°C, (−67°F)

Shear stress (MPa)

8 50 Room temperature 6

40

80°C, (180°F) 30

4

20 2 10 0

Nylon-epoxy adhesive (120°C, 250°F Cure) 0

1 2 Shear Strain (mm/mm, in./in.)

3

Fig. 1 Adhesive stress–strain curves in shear, as a function of temperature

bonded joints, is associated with the induced peel stresses at the ends of the bonded overlap. The other is distortion, which is the dominant behavior of the adhesive layer away from the ends. According to the SIFT model, there is absolutely no interaction between these two mechanisms, which is why it is permissible to prepare separate analysis methods for shear-load transfer in bonded joints and for premature failures by peel. Peel failures represent a failure of the adhesive layer to utilize its full sheartransfer capability. The value of the first strain invariant, the sum of three orthogonal strain components, J 1 ¼ ϵ1 þ ϵ2 þ ϵ3 ,

(1)

cannot be measured on any pure (neat) adhesive test coupon because failure by distortion according to the other (von Mises shear strain) invariant, γ VMcrit

ffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 2  0 2  0 2 i 1 h 0 0 0 0 ϵ  ϵ2 þ ϵ2  ϵ3 þ ϵ3  ϵ1 , ¼ 2 1

(2)

(where the prime denotes principal values) occurs before the dilatational limit had been reached. Failure of polymers by dilatation (increase in volume) occurs only in a constrained environment, as between two circular rods bonded at their ends and pulled apart, as shown in Fig. 2 (or between the adherends in bonded joints). Transverse constraints, perpendicular to the load, are needed to prevent the natural Poisson contractions which, if they occurred, would decrease the dilatational strains in the adhesive. The strain-invariant polymer failure model (SIFT) permits the superposition of separate analyses for shear and induced peel loads. There is no interaction between

1240

L. J. Hart-Smith

Fig. 2 Butt-jointed test coupon to measure J1 strain invariant for adhesives

Adhesive layer

Metallic rods

the two failure mechanisms. While the need for this has been minimized through sound design practice (gentle tapering of the ends of the adherends to minimize the induced peel stresses, as explained later), the new model puts this technique on a secure scientific foundation and also accommodates any applied transverse shear loads. At the microlevel, initial damage in bonded joints is defined by whichever of the two strain invariants reaches its critical value first. At the ends of the overlap in a badly designed bonded joint or test coupon, the dilatational J1 failure mechanism would always occur first. There are no significant peel stresses in the interior of the joint, so the distortional invariant will be exceeded first. The visual consequences of this failure may seem to be one of dilatation because the first indication is the formation of a series of hackles formed at roughly 45 from the adherends being bonded together, as shown in Fig. 3. (Exactly the same model applies for in-planeshear failures in the matrix between parallel fibers in composite laminates.) These are, in reality, failures by distortion, which result in tensile fractures perpendicular to the highest principal stress. Once hackles start to form, the adhesive layer is no longer a continuum but a series of parallel ligaments being bent elastically. There is no precise model for progressive damage to adhesive layers as more and more shear strain is applied, but a constant-stress model has served as a very useful model. This is explained in Fig. 4, based on the notion that damage starts at the knee in the stress–strain curve and that there should be some limit on how much can be tolerated, like the 2% offset definition of design ultimate load for mechanically fastened joints in ductile members. The tail end of the stress–strain curve is reserved for load redistribution around bond flaws and damage, as discussed later. It is recommended here that the knee in

44

Adhesively Bonded Joints in Aircraft Structures

1241

Load

Adhesive Tension Adherend, or fiber

Compression

Loads below elastic capability of adhesive; no hackles

Loads beyond elastic capability of adhesive; hackles begin to form

Loads well beyond elastic capability of adhesive; hackles spread and eventually link up

Fig. 3 Formation of hackles in adhesively bonded joints Design limit load must not strain adhesive beyond knee in stress-strain curve

Adhesive shear stress, t

Maximum adhesive stress and strain developed by design ultimate loads. No failure of adhesive. Corresponding elasticplastic model for design ultimate loads Model matches true modulus ge 0

1.625 ge

Final failure of adhesive

Maximum adhesive stress and strain associated with failure of adherends outside the joint. Adhesive should withstand a 50% higher load before failure. This reserve is needed for damage tolerance; an adhesive bond weaker than the adherends could act like a weak-link fuse. Load ∝ √Area under curve

Adhesive Shear Strain, g

Note that design process must account for nonlinear adhesive behavior, but a precise stress-strain curve is not mandatory. An approximation, based on a similar adhesive, will usually suffice.

Fig. 4 Relations between knee in stress–strain curve and design limit and ultimate loads

the stress–strain curve be defined to be the design limit load, for each environment (but it should be acknowledged that the bonded structures and composite repair facility for the Royal Australian Air Force, in Amberley, Queensland, have successfully set the design limit strain at twice that value). What is now clear is that the

1242

L. J. Hart-Smith

original notion (Hart-Smith 1973a, 1974) of designing to ultimate design load at the very end of the stress–strain curve is no longer appropriate. In the event that the adhesive is so brittle, in a particular environment, that there is insufficient “nonlinear” behavior to satisfy the conditions described in Fig. 4, design ultimate strength is established at the end of the short stress–strain curve in shear, and design limit load is then set at two thirds of that load. These conditions are more frequently encountered in deep space than in atmospheric flight conditions. Brittle adhesives are used on aircraft primarily in high-temperature applications as in the vicinity of engines. In such environments, even brittle adhesives are quite ductile. The stress–strain curves of adhesive layers under a close-to-uniform state of deformation, as in Fig. 1, are customarily measured on a thick-adherend test coupon using very sensitive displacement extensometers, developed by Krieger (1988), as shown in Fig. 5. Each aluminum adherend is typically 9.5 mm (0.375 in.) or 12.7 mm (0.5 in.) thick to enforce a close-to-uniform shear strain in the adhesive. Most adhesives are still characterized by the average shear stress measured on standard lap-shear test coupons (ASTM D-1002) in a variety of environments. This can be misleading in the context of joint design since this so-called adhesive shear strength varies with the adherend material, their thickness usually 1.3 mm (0.063 in.), and with the length of the overlap, as described by Hart-Smith (1993). Such coupons should never be used for anything other than quality-control tests, for which their complex mixture of variable shear and peel stresses makes them very useful.

Adhesive layer

Fig. 5 Thick-adherend adhesive test coupon and instrumentation (Printed with permission of American Cyanamid Corporation)

44

Adhesively Bonded Joints in Aircraft Structures

1243

Some researchers have also measured what they perceive to be fracture mechanics properties under up to three modes, the first being crack opening and the second shear, thinking that these properties could form the basis of bonded-joint design. The author has yet to see any successful joint designed using these properties and will not discuss them further. No such value can be related to the long overlap needed to ensure the minimum shear strain in the middle of the overlap, where the adhesive is not critical and is low enough to prevent the accumulation of creep damage to the bond line. Worse, if the bonded joint is critical under peel-stress (Mode I) loads at the ends of the overlap, it should be redesigned to make the adherends thinner there, so that the crack-opening mode is no longer critical. It makes little sense to accurately establish the low strength of a badly designed bonded joint when it is so simple to modify the design to achieve a far higher strength, for which the adhesive is not critical in peel – one in which the merits of the adhesive’s shear strength are allowed to dominate.

44.3

Surface Preparation

Proper preparation of surfaces to be bonded is the most critical step in creating durable bonded joints. This is best known for bonded aluminum structures because of the publicity given to the USAF-funded Primary Adhesively Bonded Structures Technology (PABST)-bonded fuselage demonstrator, performed at the Douglas Aircraft Company in Long Beach, California, in response to almost 10 years of widespread in-service disbonds associated with the first generation of 180  C (250  F)-cured epoxy adhesives in combination with etched, rather than anodized, surfaces. All earlier bonded aircraft structures (done mainly by de Havillands in England and Fokker in Holland) had been trouble-free because they used phenolic adhesives over chromic acid-anodized surfaces. In the 1940s and 1950s, the pioneers in this field, such as de Bruyne (1957) and Schliekelmann (1983), had studied potential durability issues very thoroughly before they undertook production bonding. Unfortunately, their diligence created the illusion that no bonded structures would ever have durability problems, so when others later changed materials and processes for bonding to simplify the manufacturing of bonded structures, they omitted to conduct the same kind of thorough durability testing and relied on short-term static tests alone, with disastrous consequences. Whole fleets of aircraft built to US bonding specifications during the late 1960s and early 1970s had all their bonded structures remanufactured. Some tried to do only local repairs, to no avail, as indicated by Hart-Smith and Davis (1996), in which it was shown that nearly half of all bonded repairs were to the same structures that had already been repaired for the same global processing errors. The disbonds would not occur until water was absorbed into the adhesive layer, which took time, but was inevitable. Local repairs are useful for impact damage to a properly bonded structure, in which none of the surrounding structure is likely to disbond, but they are unsuitable for global processing errors, as shown in Fig. 6.

1244

L. J. Hart-Smith

Exterior and internal views of the same bonded honeycomb panel, after a second failure (Large Dark gray Area), showing innumerable small holes drilled to inject resin after the first failure, to make it impossible to detect those disbonds, and the shiny resin-free inner skin surface to which the injected adhesive (two irregular light gray areas) had clearly never bonded.

Fig. 6 The futility of local bonded repairs to global interfacial failures Aluminum oxide, Al2o3

Depth of pores depends on time, temperature, concentration of etchant, and voltage potential during anodizing

Porous columnar structure, into which primer penetrates

Dense non-porous structure

Aluminum

Fig. 7 Representation of pores in anodized aluminum surface prepared for bonding (Thrall Jr. and Shannon (1985), Chap. 4; Marceau, Boeing)

All of this should have come to an end in the middle of the 1970s when phosphoric acid anodizing and optimized etch processes were established, particularly since Bethune at Boeing developed the simple wedge-crack test, ASTM D-3762 (Marceau et al. 1977), that made it easy to distinguish between properly and improperly prepared bonding surfaces. The properly prepared surfaces had a stable oxide coating, with many pores that the primer could penetrate, as shown in Fig. 7.

44

Adhesively Bonded Joints in Aircraft Structures

1245

It should be noted that, while the test procedure is very reliable, it is rendered ineffective if the “permissible” extent of interfacial disbond growth during the 1 h-long test is set so high that it fails to reject unacceptable surface treatments. The tests reported by Thrall Jr. and Shannon (1985) show that the limit should have been set at no greater than 3.2 mm (0.125 in.), which is consistent with the very large number of PABST tests that never showed more than 1.6 mm (0.063 in.) of growth. Yet the limit set for the ASTM standard was 19 mm (0.75 in.) which, according to the tests reported by Thrall Jr. and Shannon (1985), would result in more than 95% of all of their bonded structures having been “acceptable,” no matter how bad they actually were. It is also worth noting that a key element of the PABST bonding processes was the use of a phenolic-based primer, BR-127. And it is particularly significant that the first generation of 180  C (250  F)-cured epoxy adhesives that absorbed water so fast, and were associated with the many earlier in-service bond failures, were later found to perform just as reliably on the improved surface treatments as the second generation of such adhesives that were far more resistant to water absorption. Sadly, even 20 years later, the author was to see local repairs being made to large engine cowls using the same obsolete surface preparations as had caused the failures that were being repaired. When he pointed out the inevitability of further failures to the rest of the panel, in the not too distant future, it was explained to him that the use of the better newer materials and processes in the repairs would violate the aircraft’s type certificate. It was illegal to not repair it in any way other than to make it conform to the original structure when it was built! This attitude was not confined to commercial aircraft. Repairs to helicopter blades continued to be made using paste adhesives over sandpaper-abraded aluminum surfaces. In this case, the explanation provided was that the model concerned had been out of production for over 20 years, so there was no mechanism to update the technical orders governing its maintenance and repair. In these cases, and many others, the futility of the obsolete repairs was well understood, even by those performing them. Is it not time to put an end to such practices? Surely the interests of aviation are better served by reliable surface treatments for bonding than by perpetuating other techniques that are known to be unreliable. Interestingly, the same organization that provided the service data cited by Hart-Smith and Davis (1996) has since found that they suffered not a single failure on 3,000 bonded composite patches they applied over grit-blasted/silane aluminum surfaces. The financial incentives to mandating the use of only reliable surface treatments are well established. Other metal alloys, such as steel and titanium, also need appropriate surface treatments for bonded applications. The surface treatments for various metals are described in other chapters. For bonded and co-bonded fibrous composite structures, the need for proper surface treatment is far less known. The absence of reliable after-the-fact inspection techniques that besets metal bonding is exacerbated by the failure to include any durability tests as part of the quality assurance program for the production of bonded composite structures. Just as with the adhesive bonding of metallic structures, the lap-shear test coupon alone has been found to be insufficient to ensure the durability

1246

L. J. Hart-Smith

of any bonded joints. At best it can ensure only that the adhesive was cured properly (adequately cross-linked) – not that the adhesive is actually stuck. The reason for this inexcusable omission would seem to be that corrosion has been seen in conjunction with the failed bonded aluminum structures and that fiber–polymer composites cannot corrode. Whatever the cause, the need for the surfaces being bonded to have a higher energy level (be more active) than the surface of the adhesive is just as great for bonding polymers as for bonding metal alloys. Most of the surface treatments for composite surfaces to be bonded ensure precisely the opposite – that the surface will be totally inert. Some peel plies have even been coated with a silicone release agent to ensure that the peel plies can be stripped off without the slightest damage to the underlying substrate (Hart-Smith et al. 1996). This is not to imply that the problems with bonded composite aeronautical structures are as widespread as they were with bonded metallic structures. Rather, the problem is one of not knowing because there is no way of finding out before the parts separate in service. Worse, since there is no mandated durability test for bonded composite structures equivalent to the wedge-crack test, at the time of manufacture, such structures continue to be produced with no means of ensuring that they will not fall apart at some later date. Of all the specified surface treatments for bonded composite joints, only low-pressure grit blasting has an unblemished service record. The reason why it has not been standardized on is inexplicable since none of the other techniques can be applied for bonded repairs, so there is no way of any manufacturer or operator avoiding having to master this technique. One cannot put a peel ply back on a structure and peel it off a second time. Another critical issue with bonding of composite structures is pre-bond moisture, which can exist in a multiplicity of forms, each with its own distinct adverse consequences. Moisture absorbed in cured laminates and driven to the surface by the heat of the bond cycle prevents any adhesion from occurring without creating detectable symptoms. Such water lowers the energy level of the substrate, as described by Mahoney (1988), and can lead to the condition illustrated in Fig. 8, in which a complete area failed to bond but was not detected. Figure 8 shows a perfect replica of the original peel ply, known to be free of silicone, on both sides of the adhesive layer. There was no sign of adhesion at all. Yet this and many similar panels had passed all quality-control lap-shear test coupons and 100% ultrasonic inspection at the time of manufacture. It is known that these parts had been cured many months before they were bonded in another part of the country; they were not dried in an oven before bonding. Clearly, they should have been. Moisture absorbed by the adhesive before it is cured has the effect of creating a weak porous bond. Moisture condensed on the surface of uncured adhesive films taken out of the freezer and unwrapped before it had thawed out also collects at the interface and prevents the glue from sticking. Sometimes, even when the need to properly dry laminates before bonding has been recognized and specified, it has not always been enforced at the facilities of remote suppliers. Yet the need to do so continues to be demonstrated conclusively, along with the benefits from doing so, as reported by Myhre et al. (1982).

44

Adhesively Bonded Joints in Aircraft Structures

1247

10X

Note clear imprint at left of other peel ply in skin underneath adhesive layer, to which the adhesive also failed to bond.

Fig. 8 Peel ply imprint left by failure of adhesive to bond to a composite surface

Peel plies continue to be the most frequently specified surface preparations for bonding fiber–polymer composites. But they are not always reliable. Worse, there have been cases in which silicone was transferred from bagging film used in consolidation cycles. There is a tendency to underrate the importance of the so-called consumable products, like bagging film, breather, peel plies, etc., which are stripped off cured or staged parts and discarded after the cure. Some specifications even leave the selection of these materials entirely up to suppliers, with no restrictions. That is wrong! All surfaces that come in contact with the final composite part in both its uncured and cured states need to be controlled just as tightly as the prepreg itself. After all, most of them are cured together in an intimate contact. It is a standard practice to isolate tool-preparation rooms in which release agents are applied from any bonding or composite clean room. Workers there wear clean white gloves. Every article that comes in contact with uncured adhesives or composite materials, or surfaces to be bonded, needs to be treated as a potential source of contamination or a nonbond, no matter how insignificant the article seems to be. Not all procedures are as strict as this, but the loopholes in the paperwork need to be closed. There is one specific polyester peel ply that several organizations within Boeing prefer that is made on dedicated machines to eliminate contamination. This is the only peel ply permitted to be used for the 777 horizontal and vertical tails and on all subsequent production. Even so, the most stringent of tests are required to ensure the absence of silicone on each roll before it is allowed into the factory. Their prior experiences and current research sponsored by the Federal Aviation Administration (FAA) (see Bardis 2001) both confirmed the need for such care. The author would still prefer to also see the universal acceptance of something equivalent to the wedgecrack test for metal bonding – a sustained peel load in a hostile (hot/wet) environment to confirm the absence of all other mechanisms for creating inert bonding surfaces, as well. Any such lapse in quality will affect the entire structure. It is not difficult to detect the conditions that lead to global processing errors before they

1248

L. J. Hart-Smith

occur, provided that one looks, and nothing can be done about them if they are not detected until after the part is built.

44.4

Design of Adhesively Bonded Joints

The design and analysis procedures cited in this chapter were developed by the author under three government-sponsored R&D contracts over a period of years. The first was for NASA Langley during the period 1970–1973 (Hart-Smith 1973b, c, d, e), in which the elastic–plastic adhesive model was introduced and the first of the A4E series of Fortran computer codes was derived. The second increment of work was for the USAF at WPAFB, during and following the PABST bonded fuselage contract, from 1976 to 1983. Three new computer codes, A4EI, A4EJ, and A4EK, were produced and the effects of flaws and variable bond layers were covered (Hart-Smith 1982a, b, 1984a, b). There were many publications from the first two contracts. The third, also for the USAF, concerned bonded composite patches over cracks and corrosion damage in metallic structures, from 1997 to 2003 (Hart-Smith 2001). Not all of the analyses from the most recent contract have yet appeared in print. The design of double-lap and single-lap bonded joints between nominally uniformly thick adherends is straightforward. The design of the 100% full-load transfer bonded joints with no fail-safe rivets for the pressurized PABST bonded fuselage, Thrall Jr. and Shannon (1985), was reduced to a single table of overlap versus skin thickness, supplemented by a requirement to gently taper the ends of the overlaps locally for the thicker skins to prevent premature-induced peel failures. It is crucial to note that the overlaps are universal in the sense that they are independent of the magnitude of any applied loads. This enables the design of the bonded joints to be completed before the internal loads in a structure have been established. The key to this design method is explained in Fig. 9, for double-lap joints, in which the overlap is established as the sum of the elastic trough needed to ensure that the minimum adhesive shear stress is so low as to prevent creep from occurring there under even sustained loads and the “plastic” load-transfer zones needed to transmit the full strength of the adherends outside the joint. As first presented, this method is conservative in the sense that no credit is taken for the small increment of load transferred through the elastic trough. When designing patches in confined areas, the width of the plastic zones could be reduced by a length 2/λ to compensate for this effect. All of the tests during the PABST program to validate the method were at the slightly longer overlaps. Figure 10 presents the actual sizes used for the PABST fuselage, as a function of basic skin thickness. The splice plates should be made half as thick as the skins, for maximum theoretical efficiency, to make each end of the bonded overlap equally efficient. However, early tests of such joints showed a propensity for fatigue failures in the splice plate, where the skins butted together, rather than in the skins, which had a nominally equal stress. Consequently, the splice plates were made one gauge thicker. The overlaps in this table can be approximated by a simple design rule, that the

44

Adhesively Bonded Joints in Aircraft Structures h

i = inner

G

Ei

Eo

o = outer

tisult tosult (p1 + p2) = Lesser of t and t p p

2ti

p1

to

6 l

“Plastic” zones, unequal because of adherend stiffness imbalance

p2

l2 = G 1 + 1 h Eiti Eoto

Adhesive shear stress distribution

• • • •

1249

tp

tmin =

Elastic trough

tp 10

Plastic zones long enough to transfer entire adherend strength Elastic trough wide enough to prevent creep at its middle Design overlap = p1 + p2 + (6/l) Check on adequate strength by not exceeding maximum adhesive shear strain

Fig. 9 Design procedure for double-lap bonded joints

Central sheet thickness ti (in.)

0.040

0.050

0.063

0.071

0.080

0.090

0.100

0.125

Splice sheet thickness to (in.)

0.025

0.032

0.040

0.040

0.050

0.050

0.063

0.071

Recommended overlap1  (in.)

1.21

1.42

1.68

1.84

2.01

2.20

2.39

2.84

Strength of 2024-T3 aluminum (lb/in.)

2600

3250

4095

4615

5200

5850

6500

8125

Potential ultimate bond strength (lb/in.)2,3

7699

8562

9628

10,504 10,888 11,865 12,151 13,910

1Based on 160°F or 140°F/100% RH properties needing longest overlap. Values apply for tensile or compressive in-plane loading. For in-plane shear loading, slightly different lengths apply. 2Based on −50°F properties giving lowest joint strength and assuming taper of outer splice straps thicker than 0.050 in. Strength values corrected for adherend stiffness imbalance. 3For nominal adhesive thickness h = 0.005 in. For other thicknesses, modify strengths in ratio √h/ 0.005.

h

ti

P

P 

to

(Unit Conversion Factors: 1.0 in. = 25.4 mm, 1,000 lb/in. = 175 kN/mm)

Fig. 10 Standard design overlaps for double-lap bonded joints between aluminum adherends

1250

L. J. Hart-Smith

overlap be 30 times the thickness of the central adherend (skin). If there is a substantial stiffness imbalance between the adherends, the load transfer through the bond will be intensified at one end, with respect to the other, resulting in decreased shear strengths. Figure 10 includes a comparison between the theoretical bond strengths if the adhesive was strained to the very end of the stress–strain curve in shear and the strength of the 2024-T3 skins. For the thinnest skins, the bond has three times the strength of the skins, while for the thickest skins in the table, the ratio has fallen to about 1.5:1. This is why, for still thicker skins, it would be necessary to resort to splice designs with one or more steps in the overlap, to restore the margin in the bond strength. The need for this excess strength is described in a later section of load redistribution around flaws and defects. It should be noted, however, that the bond will fail by creep if an applied load sufficient to yield the aluminum adherends is maintained indefinitely. The higher bond lengths can be achieved only for loads that are quickly applied and removed. The information in Fig. 10 needed to be supplemented only by the design modifications in Fig. 11, whereby the ends of the overlap of the thicker adherends were tapered down to a maximum tip thickness of 0.75 mm (0.030 in.), to minimize induced peel stresses in the adhesive. (The corresponding tip thickness for fiber–polymer composite adherends is only 0.50 mm [0.020 in.] because of interlaminar weaknesses.) If this local tapering was omitted for the thicker adherends, the adhesive would fail under induced peel stresses long before its shear strength could be attained. The design of single-lap bonded joints is even simpler because the most critical location for any long-overlap bonded joint is in the adherends at one or both ends of ≈t 2

t

0.25 to 0.50 mm

0.127 mm (Nominal)

Adhesive layer

2.5 to 5.0 mm

Slope of 1 in 10 0.5 to 1.0 mm for Aluminum, 0.25 to 0.75 mm for fiber-polymer composites

Fig. 11 Tapering at ends of bonded overlap, to restrict induced peel stresses

44

Adhesively Bonded Joints in Aircraft Structures

1251

the adherends, as the result of combined membrane and bending stresses. These are minimized by increasing the overlap, as explained by Hart-Smith (1973c, 1978). Analyses for the PABST bonded fuselage established a design overlap-to-thickness ratio of 80:1. (This can be reduced to 60:1 for single-lap joints that are stabilized against bending of the adherends, as by being part of a sandwich panel.) Tapering of the ends of the overlap to alleviate induced peel stresses is still necessary using the proportions shown in Fig. 11. The bending moment at one end of the overlap is intensified greatly whenever one adherend is stiffer (thicker) than the other. It is not possible to make single-lap or single-strap (flush) bonded joints with a strength greater than the adherends outside the bonded overlap, so they tend to be used only for thinner more lightly loaded structures. The analyses referred above are essentially exact but can be applied only to bonded joints between adherends of uniform thickness. A new approximate method was developed during the composite repair of aircraft structure (CRAS) R&D contract that covers tapered adherends as well. The method, explained by HartSmith (2001/2002), relies on the knowledge that the adhesive stresses will be extremely low everywhere except in the immediate vicinity of changes to or interruptions in the thickness of the adherends. It starts with a closed-form analysis in which it is assumed that all the members are fused rigidly together over the entire overlap. The load sharing between the overlapped members can also account for residual stresses caused by thermal dissimilarities between adherends, as between composite patches and cracked metallic structures and between bonded titanium stepped plates at the ends and sides and the composite skins to which they are bonded. This level of analysis will predict local spikes of instantaneous load transfer at every discontinuity in load path. The second step in the analysis, also by closed form, distributes those spikes as plastic load-transfer zones of finite width, assuming locally uniform adherend thicknesses that match the real local thicknesses. If the plastic zone needed to transfer each load spike is less than the characteristic length 1/ λ defining the elastic load-transfer length, where λ is the exponent of the elastic stress distribution (see Fig. 9), the load spike is replaced by an elastic load distribution with an integral matching the calculated spike. This simple method was shown to be exactly equivalent to the precise solutions in the event that the adherends really were of uniform thickness. Bonded joints between thicker adherends than covered by Fig. 10 need steppedlap joint designs, as in Fig. 12, for the higher loads associated with thicker members. Instead of an increment of load transferred at each end of the bonded overlap, there is an increment of load transferred at each end of each step in such a joint. The addition of more steps requires an increase in the bonded area, but it is important to note that an increase in bond area without any increase in the number of steps is ineffective, as explained in Fig. 13. For bonded composite joints, the strength is predicted to keep increasing all the way to a single ply per step. Each step in a stepped-lap bonded joint is characterized by exactly the same governing differential equations as govern simple overlaps between uniformly thick adherends.

1252

L. J. Hart-Smith

Adhesive shear stress

21 cm, (8.35 in.) Titanium

Carbon-epoxy

Fig. 12 Stepped-lap bonded joints for thicker adherends (AIAA Astronautics and Aeronautics)

gmax = 0.2

Adhesive shear stress, (MPa)

50 40

8 6

30 20

4 (ksi)

10

2

0 0 Overlap = 2.5 cm, (1.0 in.)

a 50 40 MPa

b

MPa

c

50 40 30

Pelastic-plastic = 1,955 kN/m (11,172 lb/in.)

8

Pelastic-plastic = 2,184 kN/m (12,479 lb/in.)

6

30 20

4

(ksi)

gmax = 0.2

2

10 0

0 Overlap = 5.1 cm, (2.0 in.) 8

gmax = 0.2

Pelastic-plastic = 2,222 kN/m

(12,696 lb/in.)

6 4 (ksi)

20 10

2

0

0 Overlap = 10.2 cm, (4.0 in.) Note: All cross sections idential and all step lengths proportional

Fig. 13 Insensitivity of joint strength to bond area without an increase in the number of steps

44

Adhesively Bonded Joints in Aircraft Structures

1253

There have, for decades, now, been reliable analytical tools available for the design and analysis of adhesively bonded joints. Even so, there is still far too much reliance on the oversimplified model whereby the bond strength is assumed to be the product of some fictitious uniform adhesive “allowable” shear stress and the bond area. If more joint strength was needed, all one had to do, according to this procedure, was to increase the bonded area. Bonded joints do not obey such rules. Such a formula is valid today only in the context of short-overlap test coupons in which the goal is to create as closely as possible a uniform state of stress and strain in the adhesive. In the context of structural bonded joints, such a model ceased to be applicable by the end of the First World War, when airframes stopped being made from wood and fabric. That was the last time that the glue was stronger than the materials being bonded together, so that almost any design would work if the scarf angle was low enough. The most reliable of the mechanics-based bonded-joint models are of closed form because of the locally very high stress and strain gradients, and the need for iteration to cover material nonlinearities makes finite-element solutions difficult. The finiteelement models need to be converged with respect to grid size because accuracy of such analyses is not guaranteed. This problem has been exacerbated in the realm of fiber–polymer composites by the patently false simplifying assumption that the fiber and resin constituents can be homogenized into a single “equivalent” anisotropic solid. A by-product of this error is the myth of singularities at the edges of composite panels, at every change in fiber direction, which has spawned a whole field of study. The singularities vanish the instant that the notion of zero-thickness interfaces is discarded. Yet some researchers have devoted their entire career to the study of such mathematically created conditions that do not exist in physical reality. One needs to be very careful in interpreting finite-element analyses of bonded joints. Conversely, it was only through the use of properly converged finite-element models of discrete fibers in a block of resin that Gosse (1999, private communication) was able to finally validate his concept of dilatation as the primary failure mechanism for constrained polymers in bonded joints and composite laminates. There were no singularities, and the correct answer was unchanged when the mesh size was doubled or halved.

44.5

Additional Considerations for the Design and Analysis of Adhesively Bonded Joints in Fiber–Polymer Composite Structures

When boron–epoxy and carbon–epoxy composites were first introduced into aerospace structures, the only form of prepreg available was unidirectional tape. This imposed on the design of stepped-lap bonded joints some restrictions that were covered by empirical design rules. Those people involved in designing such joints, at that time, all knew not to put a layer of 90 fibers adjacent to the metal step plate and designed their laminate lay-up sequence and step heights accordingly. Consequently,

1254

L. J. Hart-Smith

interlaminar failures in the composite could be avoided and the failure forced into through-the-thickness breaking of either the composite or metallic adherends. This was meant to occur outside the bonded overlap, but it was rapidly learned from test that the thin end step of the (usually) titanium step plate would fatigue off if that step was too long. This meant that, as far as the adhesive was concerned, it was necessary only to show that it had an integrated shear strength greater than the strength of the weaker adherend. Peel stresses were eliminated automatically because the end step in the composite was always too thin to allow them to develop to a significant level. So the A4EG and later A4EI stepped-lap bonded-joint analysis codes made no provision for interlaminar failures in the composite adherends. Later, when carbon–fiber reinforcement was made available in the form of woven fabrics, it slowly became apparent that laminates could fail interlaminarly adjacent to, but not in, the adhesive layer. In retrospect, it probably took too long for the causes to sink in. Half of the surface of a woven fabric is covered by tows of 90 fibers. The application of a transverse shear force would cause cracks to run up through those bundles until they were arrested by the first 0 fibers they encountered, as explained in Fig. 14. If these initial cracks progressed any further, it would be in the form of delaminations parallel to the adhesive layer, as shown. These two codes could not provide such a check since there were no places to enter the interlaminar strengths. Initially, that would not have mattered since there was no way to measure the interlaminar shear strength of an all-90 laminate, either, so there was no reference data to compare against. The standard shear stress test coupon was for an all-0 laminate, although tests were also run on 0/90 cloth laminates and those containing
45 layers. But these were not standard properties and were not catalogued at the time. When the need for these data did become apparent, the McDonnell Aircraft Company engineers performed a very large 0° Fibers

90° Fibers

Load

Cracks

Substrate Adhesive

Load

Progression of crack growth in laminate with increasing load (Cracks do not propagate into 0° plies)

Fig. 14 Interlaminar cracking through bundles of 90 . Fibers adjacent to a bond line acted on by a 0 load

44

Adhesively Bonded Joints in Aircraft Structures

1255

number of tests on uniform adherends, not stepped-lap joints, to characterize the interlaminar failures as a function of laminate fiber pattern and adherend thickness. They then proceeded to use the A4EI code in reverse to establish fake adhesive properties which would match the measured interlaminar failure strengths using tables to document the adhesive “properties” to be used. This worked, but it was complicated badly by the presence of induced peel stresses at the ends of the overlap that caused their own set of interlaminar failures. These were governed by a different power law in regard to the adherend thickness from that governing the shear-load transfer, as explained earlier. Eventually, the author realized that what was needed was a revision to the A4EI codes to allow separate checks to be made for adhesive shear failures and interlaminar shear failures using two sets of properties that did not need to be fiddled. There also needed to be a peel-stress cutoff because the codes were being improperly used to analyze uniformly thick adherends as well as stepped-lap joints. The peel failures, if they occurred, would do so in the manner of Fig. 15; they would always precede the shear failures, undercutting it, unless they were excluded by the use of thin adherend tip thicknesses. The coding for the interlaminar shear effect has been completed, but not yet distributed. However, that for premature peel failures has not been. The author still feels that the correct response to induced peel failures is to reduce the adherend tip thickness, rather than accepting an unnecessary decrease in shear-load transfer capability. Also, until recently, the theory with which to analyze the shear and peel failures in the resin was not available. Some people believed that there had to be an interaction between the shear and peel stresses. With the advent of the SIFT failure model (Hart-Smith 2010) for fiber–polymer composites, it is now possible to analyze the stresses and strains in the resin, including those from residual thermal stresses resulting from curing the laminate at a temperature much higher than its operating temperature. It is now clear that the distortion and dilatation failures in the resin are separate and unrelated, so there is no interaction between shear and peel. The SIFT model characterizes strength in terms of two characteristic invariants for each

t sc Stresses acting on outer adherend

a

b

c a, b, and c indicate failure sequence

Fig. 15 Induced peel-stress failures in composite laminates with excessive tip thicknesses

1256

L. J. Hart-Smith

constituent (fibers and resin), so it also overcomes the problem of measuring the reference strengths to be inserted into the code for checking against. These invariants do not vary with the nature of the loads, so a test under one set of loads can be converted into a strength under all other sets of loads (unless there is a change in failure mode or in the constituent that is failing). There is now hope for a greatly improved analysis capability of the strength of both stepped and uniform bonded joints. Unfortunately, more work remains to be done. The original codes were written in Fortran IV; the author is not fluent in the later versions, and there has been no funding available to pay someone who is. Even so, the author knows of one organization where his revised listing of the code enabled them to get an operational version accounting for the shear effects. But they have yet to master the SIFT model, so they must still rely on the McDonnell empirical test data for the reference properties. Finite-element modeling techniques, using ply-by-ply models for the laminates, have been able to characterize the effect of the fiber direction adjacent to the bond on the load-transfer distribution; very little goes through the softer 90 layers. Unfortunately, they are even weaker than they are softer, so the problem of premature failure remains. Accounting for nonlinear adhesive behavior is not easy with finite elements, either. Even though the analysis codes associated with these extensions of the A4EI codes are not widely available yet, it is considered important to explain these additional phenomena, for which the understanding is complete. Without this explanation, there will continue to be confusion over measured test failures that the old theory cannot explain which brings with it a level of conservatism in design so severe as to unnecessarily restrict the use of bonded composite structures.

44.6

Design Features Ensuring Durability of Bonded Joints

Durability in bonded joints requires both that the bonded interfaces are stable (the glue stays stuck) and that the adhesive is not failed by the combination of mechanical loads and residual thermal stresses caused by dissimilar adherends. The first issue has nothing to do with the geometry of any bonded joint, although joints fail faster under peel-dominated loads than under shear loads. There are two limits associated with durability that are influenced by the geometry of the joint. The first of these is the peak shear and peel stresses at the ends of each overlap. This is obvious and well understood. The other is to limit the strain level near the middle of the overlap, even in the most severe environment. This is the little understood requirement represented by setting the minimum stress level at or less than 10% of the maximum, which defines the overlap needed to prevent any creep from accumulating. The need for such a requirement was exposed by some of the early fatigue tests on the PABST program. Quite misleading conclusions, both positive and negative, could be drawn from durability tests on short-overlap coupons, as explained by Thrall Jr. and Shannon (1985).

44

Adhesively Bonded Joints in Aircraft Structures

1257

The key to the success of these designs was the acknowledgment that the adhesives’ shear stresses were, and should be, highly nonuniform. There are enormous differences between the way adhesives behave, under what appear to be the same external loads, in short-overlap test coupons and long-overlap structural joints, as explained in Fig. 16. Tests on short-overlap coupons cannot be relied upon to differentiate between adhesives (and surface treatments) that will endure in service and those that will (or have) not. At best, they can make comparisons between slightly different adhesives in the same class. It is almost impossible to prove in the short term that a bonded joint will last 30 years or more in service. Any test under representative load conditions would have to last at least 30 years to indicate a satisfactory result. If the load intensities were increased or the test environment aggravated, to ensure a test result in a short time, there would be no way of knowing what the corresponding service life would be under realistic conditions. And the best adhesive systems would last for 30 years under realistic loads even when tested in artificially severe environments, unless the structural elements wore out first. The only saving grace is that the surface treatments associated with premature in-service interfacial failures could be accelerated and made to fail rapidly under adverse conditions. In other words, the inferior systems could be identified rapidly, but the best systems could be identified only by not appearing on the list of unacceptable systems.

a

C

B Adhesive shear strain

ge

tp Adhesive shear stress

A G

E

F

D Realistic overlap for structural joint

b Adhesive shear strain

Adhesive shear stress

Short overlap test coupon

Fig. 16 Differences between short-overlap test coupons and long-overlap bonded joints. (a) Realistic overlap for structural joint. (b) Short-overlap test coupon

1258

L. J. Hart-Smith

The reason why properly designed bonded joints do not suffer from mechanical fatigue failures is that the most critical conditions are not developed where the adhesive is protected by the adherends. This can be understood by characterizing the minimum and maximum adhesive shear strains firstly as a function of bonded overlap and secondly as a function of adherend thickness, accounting for the effect of the environment in each case. This is illustrated in Fig. 17, for room temperature. The reason why short overlaps should not be used for structural joints is that even the smallest loads can result in the critical conditions being developed at the ends of the overlap, under sustained load, or by an accumulation of incremental loads, because there is no restraint on the minimum shear strain developed in the middle of the overlap. Once the overlap has exceeded a critical value, proportional to the thickness of the adherends, there is such a constraint on the minimum shear strain, at the middle of the overlap, that simultaneously imposes a limit on the peak shear strain at the ends of the overlap, through compatibility of deformations. No matter how long the loads are applied, the peak shear strain at the ends cannot grow indefinitely. Testing during the PABST program showed that creep did accumulate steadily at the ends of the overlap as long as the load remained. However, it recovered during periods of unloading, every time, with exactly the same shear strain at the end of 8 h into the 5th load cycle as the 14th, for example. There is no counterpart to this behavior possible for short-overlap test coupons, in which such creep occurs but accumulates cycle by cycle instead of recovering. For still longer

gmax

0.5

Elastic-plastic shear strain

Adhesive 0.4 gmin shear 0.3 strain, g 0.2

0

gmin

gmax gmin at middle of overlap

gmax Elastic Shear Strain gmin

0.1

gmax at edge

gmax

1  (cm)

0

2

ge 0

gmin

1 2 Overlap  (in.)

0

1

 (cm)

2

1,500 Load limited by adhesive strength

200 Maximum joint strength, P (kN/mm)

Adherend strength cut-off

1,000

100

0 0

Fully plastic adhesive. (Load limited by bond area)

Adherend strength 500 cut-off

2.5

5

0

 (cm) Adherend thickness

a

1.6 mm (0.063 in.)

b

5 2.5 Overlap  (cm) 3.2 mm (0.125 in.)

0

2.5

5

 (cm)

c

6.4 mm (0. 25 in.)

Fig. 17 Effects of adherend overlap and thickness on maximum and minimum shear strains at room temperature

44

Adhesively Bonded Joints in Aircraft Structures

1259

overlaps, the minimum adhesive shear strain decreases asymptotically toward zero, but the peak values remain constant, since the load that can be applied to the bond is limited by the strength of the adherends. So, provided that the design overlap is long enough to extend to the far side of the transition between short- and long-overlap behavior, for the most critical of the environments, any further increases are of no benefit (apart from the provision of a reasonable assembly tolerance). The influence of the environments on the design overlaps is shown in Figs. 18 and 19 for the hot/wet and cold environments. It is usually found that the upper service temperature limit sets the design overlap because then the adhesive is at its softest and weakest, while the lowest temperature establishes the limiting joint strength because that is the condition of least strain energy, but not by much, as noted in Fig. 1. Figures 17, 18, and 19 also show that, for the thinnest adherends, there is a deep precipice on the peak shear strain when crossing the critical overlap and that the gmax

gmax

1.0 0.9

gmax

Elastic-plastic shear strain

gmin

gmin

0.8

gmin

0.7 Adhesive shear strain, g

gmax

0.6 0.5 0.4

gmax

0.3 0.2

Elastic shear strain gmin

0.1 0

1  (cm)

0

2

gmin 0

1 Overlap  (in.)

2

0

1  (cm)

2

1,500 (lb/in.)

200 Maximum joint strength, P (kN/mm)

100 500

0

Adherend thickness

0

a

Fully plastic adhesive load limited by bond area

Adherend strength cut-Off

1,000

Adherend strength cut-Off

2.5  (cm)

5

1.6 mm (0.063 in.)

0

b

2.5 Overlap  (cm) 3.2 mm (0.125 in.)

5

0

c

2.5  (cm)

5

6.4 mm (0. 25 in.)

Fig. 18 Effects of adherend overlap and thickness on maximum and minimum shear strains, at maximum service temperature

1260

L. J. Hart-Smith

0.5 g Just less than (ge + gp)

0.4

Adhesive shear strain, g

0.3

Elastic-plastic

0.2

gmax shear strain Elastic shear gmin strain

0.1 0

0

1

2

gmax

gmax gmin

gmin 0

1 2 Overlap  (in.)

0

Maximum joint strength, P (kN/mm)

(lb/in.)

Adhesive shear strength limit

1,000

0

Adherend thickness

0

2.5  (cm)

5

a 1.6 mm (0.063 in.)

Fully plastic adhesive. load limited by bond area

Adherend strength cut-off

Adherend strength cut-off 500

100

2

Load limited by adhesive shear strength

1,500

200

1

0

2.5 5 Overlap  (cm)

b 3.2 mm (0.125 in.)

0

c

2.5  (cm)

5

6.4 mm (0. 25 in.)

Fig. 19 Effects of adherend overlap and thickness on maximum and minimum shear strains, at minimum service temperature

precipice decreases with increasing adherend thickness. There was absolutely no precipice left for 6.35 mm (0.25 in.) thick adherends, indicating that the whole concept of preventing bond failures by limiting the peak adhesive shear strain would no longer prevail, even though the minimum adhesive shear strain could be made low enough. The key to success is that both limits on adhesive strain have to be present. This is why stepped-lap joints are necessary for thicker adherends. But with that design feature to limit the strains in the adhesive, the joints have proved to be just as durable as for the simpler joints used for thinner adherends. The stepped-lap titanium-to-carbon-/epoxy-bonded joints at the wing roots on the F/A-18 aircraft are a testament to this, at load intensities of almost 5.25 kN/mm (30,000 lb/in.).

44.7

Load Redistribution Around Flaws and Porosity

One of the most remarkable characteristics of well-designed bonded joints, as defined earlier, is their tremendous tolerance for quite large local defects. Provided that the surface treatment and processing were correct, the damage would not spread. (On the other hand, quite the opposite was true in the case of global processing errors, for which it was only a matter of time for absorbed moisture to attack the interfaces on poorly bonded metallic structures. The mechanism for spreading initial failures on

44

Adhesively Bonded Joints in Aircraft Structures

1261 3

20

15 Adhesive shear stress (MPa) (At room temperature)

2 (ksi)

10 1 5

0

0

0.64 mm (0.025 in.)

50.8 mm (2.0 in.) lb/in.) 175 kN/m (1,000 lb/in.) 0.127 mm (0.005 in.)

1.27 mm (0.05 in.) −6

Flexibility = 12.7 mm/kN (2.2 ⫻ 10

in./lb)

Fig. 20 Adhesive shear stress distribution for bonded joint with no defects

improperly bonded composite structures is not clear, but the result was just as inevitable.) The same state of nonuniform stress and strain that ensures the durability of properly designed bonded joints is responsible for the ability to tolerate local flaws with no loss of overall strength. It is self-evident that, if the adhesive layer ever was uniformly critically strained, there could be no tolerance to even the slightest flaw. Figure 20 shows the calculated adhesive shear stress distribution, at room temperature, for a bonded splice on the side of the pressurized PABST fuselage. What is remarkable is that, even at the 1.3-P proof pressure condition of 175.13 N/mm (1,000 lb/in.), the adhesive was not even strained beyond its elastic capability at the ends of the overlap. (The lightly loaded elastic trough appears to be unnecessarily long, but it is actually sized by the hot/wet environment, not by any room-temperature event.) If it is now supposed that there is a half inch disbond one quarter of an inch from the edge of the 2 in. overlap per side, the adhesive stresses would be modified slightly, as shown in Fig. 21. The increment of load that used to be transferred through the now defective area is now shared by extra loads in the immediate vicinity of each side of the defect – without affecting the peak stress at the ends of the overlap. This is indicative of a robust capacity of properly designed bonded joints to tolerate large local defects, provided that one can rely on the remainder of the bonded area remaining stuck. (This capacity is lost whenever the adjacent bonded areas are also on the point of failing.) If it is further supposed that the same defect, or damage, had occurred right at the edge of the overlap, the redistributed adhesive stresses would be as shown in Fig. 22. Again, remarkably, the value of the peak adhesive stress would not be affected. It would simply be moved to the edge of the defect, where load transfer again becomes possible. This kind of defect would need to be sealed to prevent the intrusion of water, which would freeze and expand at high altitude, thereby spreading the initial damage or defect, under what is known as the freeze/thaw cycle. However, the same size defect in Fig. 21 would best be recorded but otherwise left alone. Any attempt to repair it

1262

L. J. Hart-Smith 3

20

15

Adhesive shear stress (MPa) (At room temperature)

2

10

Stresses with disbond (Small increase in peak value)

5

Stresses without disbond

(ksi) 1

0

0

1.27 mm (0.05 in.)

0.64 mm (0.025 in.)

50.8 mm (2.0 in.)

175 kN/m (1,000 lb/in.) 0.13 mm (0.005 in.) 12.7 mm (0.5 in.) Bond flaw Flexibility = 15.6 mm/kN (2.73 ⫻ 10−6 in./lb)

Fig. 21 Load redistribution due to local bond flaw near the edge of the overlap 3

20 Stresses with disbond 15

Adhesive shear stress (MPa) (At room temperature)

Shift in stresses due to disbond at edge. No increase in peak

10

2 (ksi) 1

5

0

0

1.27 mm (0.05 in.)

50.8 mm (2.0 in.)

175 kN/m (1,000 lb/in.) 0.13 mm (0.005 in.)

0.64 mm (0.025 in.)

12.7 mm (0.5 in.) Bond flaw Flexibility = 15.6 mm/Kn (2.73 ⫻ 10−6 in./lb)

Fig. 22 Load redistribution due to local bond flaw at the edge of the overlap

would break the environmental protection, by cutting through the primer and exposing bare untreated metal on the edge of any hole that might be drilled to enable resin to be injected to fill the gap and make the discrepancy undetectable in future. All that would be accomplished is to decrease the remaining life, without any increase in joint strength. If the surfaces have been prepared properly, most local damage will not spread. If they were not prepared properly, local repairs are pointless, since the adjacent bonded areas will soon need to be repaired themselves, as noted earlier.

44

Adhesively Bonded Joints in Aircraft Structures

1263 3

20

15

Adhesive shear stress (MPa) (At room temperature)

No significant change in stress at edge due to large central disbond

2 (ksi)

10 1 5

Zero stress over disbond Area

0

0

0.64 mm (0.025 in.)

50.8 mm (2.0 in.) lb/in.) 175 kN/m (1,000 lb/in.) 1.27 mm (0.05 in.)

25.4 mm (1.0 in.) Bond flaw

0.127 mm (0.005 in.) Adhesive layer

Flexibility = 15.8 mm/kN (2.72 ⫻ 10−6 in./lb)

Fig. 23 Load redistribution due to local porosity in a bonded overlap

If the defect was created in the form of a trapped bubble in the middle of the overlap, as in Fig. 23, when the edges of the overlap were pinched off and there were too few small vent holes in the splice plates (this omission is really usually only a problem with large-area doublers), there would be a large area of porosity where the gap between the adherends was too great for the adhesive layer to fill. The natural occurrence of porosity is discussed by Hart-Smith (1982b). Obviously, since there was no load being transferred there anyway, the presence of such occasional areas of porosity may be considered unimportant, unless it were to cause a misfit with adjacent stiffeners (which is a problem with larger-area bonded doublers with no vent holes). It should also be noted that it is impractical to fill every little bubble in the area of porosity and that, even if this were possible, the gap between adherends could not be reduced and the locally thick bond layer could never pick up its designated increment of load anyway. The thicker bond layer necessarily associated with porosity, unless it occurred everywhere as the result of pre-bond moisture in the uncured adhesive film, ensures that the porous area will not fail, even if it occurred in an area of high nominal shear stress. What most porosity does is to transfer extra load to any adjacent thinner regions in the adhesive bond layer. Figures 20, 21, 22, and 23 are typical of local flaws in and damage to thin bonded structures. They are usually innocuous and should not be repaired, except when it is necessary to seal exposed edges to prevent the ingress of moisture. This is because of the large excess of strength of the bond over the adherends. For thicker structures, this excess strength is diminished, and flaws and damages can become more significant, as explained by Hart-Smith (1982a). The preceding examples refer to one-dimensional situations. When bond flaws are assessed in two dimensions, the need for the bond to always be stronger than the adherends becomes abundantly clear. If the adherends are stronger than a properly

1264

L. J. Hart-Smith

Fatigue cracks initiate in skin

Load in bond limited by adherend strength regardless of disbond length

Disbond

a

Adhesive Stronger than the Adherends

Disbond growth arrested Once fasteners accept Load in disbond area

Life limited by crack Initiation at most Highly loaded fastener

Disbond

b

Adhesive bond

Adhesive Weaker than the Adherends (Fasteners Needed to Provide Fail Safety and to Prevent Catastrophic Unzippping of Bond)

Fig. 24 Two-dimensional load redistribution around a large flaw in a bonded overlap: (a) adhesive stronger than the adherends; (b) adhesive weaker than the adherends

processed bonded joint with no defects, any large defect or damage acquires the characteristics of a through crack in metallic skins, as explained in Fig. 24. Provided that the bond outside the defective or damaged area is stronger than the adherends, the initial damage cannot possibly spread. Instead, the diverted load will either initiate fatigue cracks in the skin, just outside the damage, or delaminations in composite laminates, at the same locations. In either event, there will be a long interval in which the damage can be detected before it becomes critical. Without such protection from what appears to be merely excess strength in a one-dimensional assessment, large bond flaws would behave like cracks in metallic structures, even if all of the surface treatment and processing had been the best in the world. This is why it is always necessary to design bonded joints so that they can never become the weak link in the load path.

44.8

Effects of Thermal Mismatch Between Adherends on Strength of Bonded Joints

When thermally dissimilar materials are bonded together, residual thermal stresses are developed that usually detract the remaining strength available for transmitting mechanical loads. These phenomena occur whenever titanium edge members are bonded around the edges of composite panels to permit the use of mechanical fasteners during final assembly of the structure or to permit disassembly in service

44

Adhesively Bonded Joints in Aircraft Structures

1265

Composite Joint geometry

Metal

t Composite Unloaded joint (Showing residual deformations due to Thermal mismatch)

Residual stresses

g Residual strains

Tensile shear load (Left end critical)

t 1

3

2

Increasing load

Adhesive shear stress Critical shear-strain failure

g Compressive shear load (Right end critical)

1

2

3

Increasing load

Adhesive shear strain

Fig. 25 Effects of adherend thermal mismatch on adhesively bonded joints

for inspections and repairs. These thermal stresses are roughly proportional to the difference in temperature between the curing and operational temperatures. Their analysis is explained by Hart-Smith (1978) and illustrated in Fig. 25. The key issue is that, whereas the shear stresses and strains in the adhesive that are caused by mechanical loads have the same sign at each end of the overlap, the shear stresses and strains caused by adherend thermal mismatch have opposite signs from end to end. Consequently, the thermal effects weaken bonded structures below the strength that would be expected if such considerations had been omitted from analyses. Also, the critical end of the joint can change between tensile and compressive lap-shear loads. The issue is complicated by the fact that some of the thermal strains can creep out of short-overlap test coupons, but they cannot be ignored in long-overlap joints. These effects are more pronounced on thicker structure than on thinner adherends and, in extreme cases, can cause bonded joints between thermally dissimilar materials to actually self-destruct during cooldown after curing at elevated temperatures. The problems are significant for aircraft structures and can be the critical load cases for most space structures. For this reason, acrylic adhesives are used more on space structures than on aircraft structures, for which epoxy adhesives dominate.

44.9

Inspection, Testing, and Quality Control

Inspection of bonded and composite structures is one of the most contentious issues associated with these two technologies. A few years ago, the author was told by USAF colleagues of a PhD thesis at WPAFB in which it was estimated that

1266

L. J. Hart-Smith

inspection accounted for over one third of the life-cycle cost of composite structures in their fleet, while the material cost was less than 2% of that total. In spite of the willingness to spend so much money on inspections, mostly in service, the inability to guarantee the quality of such structures after they have been built is a great impediment to their more widespread use. One is forced to conclude that the inspection programs in place have not provided good value for money, as is explained in the next section. In fact, the service record of these structures is far better than this image suggests. The only problems of any significance have been those associated with failure to detect the use of inappropriate processing at the time of initial manufacture. One post-manufacture inspection method can reliably assess whether or not the glue is still stuck at any time during the service life of the structure. This is the bonded pull tabs described by Hart-Smith (1995) and illustrated in Fig. 26. If the surface has deteriorated, the tab will easily be pulled off the surface if a drop of water is applied in conjunction with the peel load. This is known to work on metal-bond structures, per the tests run during the PABST program. It has yet to be applied widely on composite structures, possibly out of concern to damage the underlying structure if the bond were not defective. However, the one manufacturer who did so as part of the process of evaluating and solving a problem reported that it worked just fine. The inherent inadequacies of some of the expensive standard inspections for bonded and composite structures are described later after describing the various kinds of defects. It is first necessary to differentiate between mechanically induced damage, which is relatively easy to detect, and weaknesses created by inappropriate processing, which can be found reliably at the time of manufacture only by a combination of shear and peel test coupons. This condition cannot be detected by standard ultrasonic inspections because, initially, there is no gap to be found. However, if the defects are not detected until much later, it is necessary to examine the fracture surface to confirm that the failure was interfacial (for a processing error) and, therefore, that it could potentially extend over the entire bonded area. Impact Do not pry tab off with a screwdriver. Doing so would scratch the surface and break the environmental protection

Water in eyedropper Pliers

Inspection tab

Adhesively bonded component

Fig. 26 Use of bonded tabs to assess bond strength at any stage in the life of bonded structures

44

Adhesively Bonded Joints in Aircraft Structures

1267

damage tends to be far more localized, and it is unlikely to spread because it takes far more force to break bonds cohesively. Such failures are characterized by a rough fractured adhesive surface on both adherends. Discrepancies found by ultrasonic inspections at the time of initial manufacture fall into a different category. These are misfits that cause voids and porosity in bonded joints. That is the one valuable contribution that these expensive and time-consuming inspections can make. However, there is an even quicker and far more reliable technique for securing the same information. At SAAB, instead of discarding the bleeder pack after the parts are made, it is part of the inspection process for the fuselage panels of the SAAB 340 and SAAB 2000 aircrafts (Hart-Smith and Strindberg 1995, 1997). The local absence of any squeeze out, or any local excessive squeeze out, is an unambiguous indication of where the detail parts did not fit together properly, as explained in Fig. 27. (The need for such bleeder packs, which were not needed for the bonded wing panels, arose because of the use of standard rigid external bonding tools.) Such local defects are usually structurally insignificant, as explained earlier, but, once detected, cause inordinate inspection costs during service to prove that they have not grown, even though extensive service history indicates that they will not if the processing had been reliable. However, if such discrepancies are repetitive, considerable future costs can be eliminated by modifying the bonding tool or the individual parts so that such misfits are eliminated from future production, as explained by Hart-Smith and Strindberg (1997). While it may seem contrary to intuitive thinking, by anyone who has not read and understood Deming’s books on Total Quality Management, defect-free bonded

Recessed bond between skin and doubler Bonded interfaces

Indication of good fit and flow Cured adhesive in breather/bleeder pack

Excessive squeeze-out

Fig. 27 Use of bleeder pack to indicate the location of any misfitting detail parts

1268

L. J. Hart-Smith

and composite structures really are the least expensive to make. The absence of defects is the path to even more reductions in inspection costs since, after ten consecutive defect-free assemblies have been made, it is permitted to switch to a sampling inspection plan instead of having to inspect 100% of every assembly. In the same vein, there are two positions on whether or not it is necessary to attach traveler coupons to every single part being anodized or if it is sufficient to validate the temperatures, chemical concentrations, voltages, and the like, in the tank farm only at the beginning and end of each shift. The only additional information that traveler coupons on every detail part can provide is assurance that the electrical connections needed for anodizing have been attached correctly. But that detail can also be verified by visual inspection and is sometimes nullified by stringing all coupons in a single patch together to save time in testing them, instead of attaching them to parts one at a time in such a way that they could be anodized only by current that first flowed through the part it was validating. The issue is clouded by the need to measure both voltage and current flow to ensure complete anodizing and not merely the use of polarized light inspections to detect anodizing. This method has failed repeatedly to detect under-anodizing, as noted during the PABST program. This condition has been created by undetected corrosion in the electrical circuit that reduced current flow, even though the correct voltage was maintained, and by too low temperature in the tank farm. In other words, all that the use of 100% traveler coupons ensures, beyond what validating the tank farm twice a shift can do, could have been ensured by diligent visual inspections during the processing. It would seem that, if these 100% inspections were detecting discrepancies not found by the twice-a-shift inspections, the processing specifications were not being followed carefully enough. The author would suggest that the extra tests should be superfluous if there is a stable fully trained workforce in the bond shop but that they serve as a useful insurance policy if there is so much labor turnover that additional training ought to be occurring. These shear and peel coupon tests are neither expensive nor time-consuming. However, it should always be remembered that increments of the inspection budget once dissipated on unnecessary tests will not be available at some future date to resolve some unanticipated real problem. Inspection dollars are most valuable when they are solving problems or confirming their absence, rather than in buying off discrepant parts as is – without causing the discrepancies to be eliminated from subsequent parts. Once reliable process specifications have been established, their application for metal bonding is customarily validated by two tests on coupons referred to as traveler coupons that are processed with the part. One of these tests is the lap-shear coupon (ASTM D-1002), tested at room temperature, and the other some form of peel test in a hot/wet (hostile) environment. Common peel tests are the wedge-crack test (ASTM D-3762 with a far more stringent requirement on the absence of interfacial failures than in this specification) and the climbing-drum peel test (ASTM D-1781). As noted earlier in this chapter, the first test ensures only that the resin in the adhesive has been exposed to the correct thermal profile, while the second ensures durability in service. Both tests are needed. The parts processed with the coupons are primed within the strict limits on exposure after the

44

Adhesively Bonded Joints in Aircraft Structures

1269

etching, anodizing, and rinsing have been completed. (In the auto-industry, the priming is often an electro-dip process at the end of the other surface treatments, but the lower volume of production and greater size of the parts have not favored this approach for aerospace. It would help if it had, since the need for such critical control of the primer thickness being sprayed on might then be avoided. But that appears to be a development for the future.) The bare surfaces will deteriorate with time if not primed promptly but can safely be stored after priming. They are then left on hold until the traveler coupon tests have been completed, typically in an hour, and not released to the assembly area until the coupon tests have been satisfied. The surfaces of the actual parts have to be reprocessed if the coupons fail the test. This delay is avoided when the tank farm is validated only twice per shift, rather than for each tank load of parts. But the saving must be weighed against the risk of far greater recovery costs if parts have been bonded together before some discrepancy is discovered. In the case of bonded composite parts, the standard process inspections are incomplete, in the sense that only lap-shear coupons are mandated. There is rarely any use of a peel test to ensure that the adhesive is stuck properly. In the author’s opinion, based on years of observation of manufacturing and in-service experience, there ought to be because there is no other way of discontinuing the use of surface treatments that have been found to be inadequate. This issue is discussed by HartSmith (2003). It is vital to understand that each kind of test evaluates only one factor. The shear test cannot assess durability, and the peel test cannot ensure complete crosslinking of the resin. Perhaps the worst violation of this principle is the use of ultrasonic inspections to overrule failure to pass the coupon tests for parts that are deemed too expensive to scrap. There are never any written specifications allowing this, but there are also no written instructions prohibiting material review boards from making such decisions. Fortunately, since the introduction of phosphoric acid bonding and phenolic-based adhesive primers, the processes have become more robust than the testing techniques, and such decisions have not created a safety problem. There is a similar confusion with bonded composite structures, related to a widespread failure to understand that process-verification coupons cannot do so if the requirements are set so low that even badly processed parts can exceed the requirements. It is therefore often necessary that the coupons not have the same fiber pattern as any individual part. Only an all-0 lap-shear coupon can impart enough load to fail a properly processed adhesive layer. Failure outside the bonded overlap in a far weaker quasi-isotropic laminate tells nothing whatever about the strength of the adhesive layer or whether or not it had been fully cured. So the use of such fiber patterns as bonded-joint test coupons is always inappropriate, no matter what the design of the structure is. Similarly, all-0 peel test coupons are needed to evaluate durability because it is so easy to divert any interlaminar crack through any layer of 90 fibers. Sadly, not only is there yet any agreement on the need for durability tests of bonded composite joints, there is not even a universal recognition of the need for the composite coupons that are used to be strong enough to pass only when a fullstrength bond has been created.

1270

L. J. Hart-Smith

Ultrasonic inspections can, at best, tell only that no impact damage has occurred that was of sufficient magnitude to create internal delaminations and that the parts did or did not fit together properly when they were first bonded together. However, there are some far more reliable and far less expensive visual inspections that can ensure proper fit – and reveal a lot about the processing too! This is explained in Fig. 28, taken from Hart-Smith (1995). The fillet cannot achieve the preferred shape if the adhesive was not heated up correctly to make it flow; it cannot wet a contaminated surface; and it cannot even form if the parts are too far apart. The shape of the fillet is an invaluable reliable indicator of a good bonded joint, and such an inspection is both rapid and inexpensive. The total absence of any fillet indicates a gap just as reliably as any ultrasonic inspection. Indeed, it is necessary to seal any edges where there is no fillet before attempting to use ultrasonic inspections because liquid that could ingress through an open edge could mask the extent of any such gap from any ultrasonic inspection, which relies on gaps to create signals of discrepancies. A porous spew for an epoxy adhesive would indicate the presence of pre-bond moisture but would be expected for a phenolic adhesive. (Despite their excellent service record, phenolic adhesives tend not to be used much in aeronautical structures today because of the higher pressures needed than for curing epoxy adhesives.) To summarize the salient points about inspections and quality control for bonded structures, it is vital that appropriate processes be specified – and followed. Verifying this requires both shear and peel tests in a hot/wet environment for metal bonding and bonded composite structures. Ultrasonic inspections can be relied upon to find only in-service impact damage, not progressive degradation. However, ultrasonic inspections at the time of manufacture can identify gaps and misfits and point the way to major cost savings by eliminating such discrepancies from future parts, even if they were structurally insignificant. Any defect, once detected, incurs enormous subsequent inspection costs to prove that it has not grown. It is far less expensive to make parts with no defects than to buy them off as is. Although it is usually not

Adhesive

Good flow and wetting

Recessed fillet, probably due To poor fit or lack of pressure

Poor flow, probably due To slow heat-up rate

Porous spew, probably due to moisture in adhesive or excessive heat-up rate

Fig. 28 Importance of visual inspections of bonded structures

44

Adhesively Bonded Joints in Aircraft Structures

1271

mentioned in any specifications, visual inspections can be incredibly valuable, even if they do not eliminate the need for all other tests.

44.10 Concerns About the Misplaced Emphasis of the QualityControl Inspections for Bonded and Composite Structures Just as the primary purpose for the structural analysis of aeronautical structures should be to make the structures better than if the analyses had not been performed, the primary purpose of inspections should be to make the structures better than if the inspections had not been performed. Unfortunately, because analysis and inspection seem to be done too late in the construction process, both fail to achieve their best potential. Indeed, the primary emphasis seems to be to respond to regulations that the structures must merely be shown to be safe. In both cases, the products would be even safer if those tasks were performed much earlier. The current illogical emphasis on after-the-fact ultrasonic inspections would appear to have its origin in practices developed for the assembly of rivetted metallic aircraft many decades ago. Instead of step-by-step inspections during manufacture, reliance was placed on thorough after-the-fact inspections. Most mistakes could be corrected by drilling out and replacing a few badly driven rivets. This would restore the structure to 100% of its design strength, with no loss of safety or reduction in fatigue life. It may or may not have been the least expensive approach to inspection, but at least, in that context, it was reliable. Fixing any problems as they were created would have been far less expensive (in the author’s opinion, based on his experiences and observations), but if there were no defects to find, the intermediate inspections were perceived as a waste of money and repeated interruptions to the assembly schedule. The problem with adhesively bonded and fiber–polymer composite structures is that after-the-fact repairs are usually incapable of restoring 100% of the loss of strength caused by the defects, particularly in components with large loads in them. In short, it is not possible to inspect quality into such parts after they have been made and assembled. The only way to ensure high quality in bonded and composite structures is by appropriate process control, before and during the manufacture and assembly of the parts. Sadly, many of the approved process standards and inspections fail to do so reliably. Worse, many of the approved repairs merely conceal the existence of the damage from further inspections without any increase in strength. The author must conclude that many of the inspections that are performed are done merely because it is possible to do so, rather than as part of a rational quality assurance program. This problem is exacerbated by the absence of what are perceived as critical after-the-fact inspection capabilities. But that excuse misses the fundamental point – belated repairs to bonded structures are inherently incapable of restoring 100% of the design strength. And, with bonded metal structures, these same after-the-fact repairs break the environmental protection that was created during bonding and thereby decrease the service life of the structure. The purpose of this section is to draw attention to some of these issues, in the hope that their exposure will encourage improvements. It should be noted at the

1272

L. J. Hart-Smith

outset that this is not a matter of simply spending more money doing the same things more thoroughly. There is widespread irrefutable evidence that improved quality during manufacture greatly decreases the cost of making and inspecting the parts. If there are no defects to be found, there are no costs incurred by fixing them and no interruptions to the flow of manufacturing through Material Review Board (MRB) activities. The focus here will be on bonded structures, but many of the same considerations apply to composite structures, too. It is first necessary to distinguish between the two possible mechanisms of failure of bonded joints. The first, which has proved to be by far the most troublesome, is adhesion failures in which the adhesive was never stuck to the adherends, even if there was no gap between the parts to be detected by ultrasonic inspection. This is an interfacial failure that has not degraded the adhesive itself but totally destroyed its capability to transmit load. The second is cohesive failure within the glue line, which results from impact damage. Such damage tends to spread extremely slowly, if at all, because in a properly designed bonded joint, the undamaged joint itself is never the weak link, as explained earlier. The adherends will fail before such damage can be spread. With interfacial failures, however, there is no need for a load to cause the nonbond to spread. There is a third kind of defect, porosity, that occurs during assembly when the parts do not fit together properly or when the bonding tool prevents them from fitting together. The importance of interface control is explained by Hart-Smith (1997). Whereas there is a standard tolerance on the parts on the order of
0.75 mm (0.03 in.), the tolerance on a glue line that is nominally
0.125 mm (0.005 in.) thick is only
0.025 mm (0.001 in.). This is impossible to satisfy unless the parts being bonded are able to drape to fit, as with skin/doubler combinations. If the gap between the parts is more than about 0.25–0.50 mm (0.010–0.020 in.), the adhesive will flow out of the gap if it is cured by high temperature. Only paste adhesives can be used to fill gaps between parts, and they tend to have lower environmental durability than high-temperature-cured film adhesives. More critically, they take far longer to cure, requiring that the parts be left in assembly jigs for far longer. Worse, since paste adhesives start to cure as soon as they are mixed, this imposes severe limitations on the permissible lay-up time. But the problem with misfitting parts does not end with the creation of porosity or voids. Gaps between parts to be bonded together are not uniform over the entire bond area. Consequently, there are adjacent areas of ultrathin bonds, which are very stiff and attract all the load transfer, and excessively thick bonds, which shed the load they should have been transferring to adjacent thinner bonds, as explained in Fig. 29. This nonuniformity in load transfer is in addition to the nonuniformities discussed earlier whereby most of the load is transferred immediately adjacent to the ends of each bonded overlap. What the variable bond-line thickness causes that is totally overlooked when one thinks of a uniform adhesive shear stress throughout the entire bond area (which is still a popular model in spite of the evidence) is that, once the thicker bond areas have been established, whether by porosity or voids, even if it were possible to fill the cavity with adhesive, let alone ensure that it adhered, the thick bond regions would not transfer load anyway. This false perception has even worse consequences. The believers in the concept of a uniform adhesive shear stress

44

Adhesively Bonded Joints in Aircraft Structures

1273

Adhesive shear stress

Zero bond thickness

Common displacement between adherends Very thin adhesive layer

Very thick adhesive layer Infinite bond thickness

Load transfer attracted to regions of thin bondlines

Load transfer diverted away from regions of thick bondlines

(Bondline thickness exaggerated. Not to scale)

0

Fig. 29 Load-transfer redistribution caused by variable bond-line thickness

throughout the bonded area consequently believe that increasing the bond area increases the joint strength. The reality, as explained earlier, is that the strength increases with bond area only until a plateau in strength is attained. Further increases in bond area are futile; only the addition of more steps can increase the shear strength of the joint. The only way to deal with this potential problem of unanticipated variation in shear transfer is to ensure that it never occurs by properly coordinating the design of the structure and the bonding tool(s) before the parts are bonded together. When this has not been done, it is standard practice to identify and locate any thicker gaps and to insert additional adhesive, so that no voids or porosity are created. It is then quite wrongly believed that this must be a “good” bond because ultrasonic inspections cannot identify any defects. In a similar vein, it should be obvious that, if the adherend surfaces are not processed appropriately, there is no way to make the glue stick other than to disassemble the parts and start all over again. But, if there are no gaps between the parts, the inability of the standard ultrasonic inspections to detect any of the defects that they are assumed to be capable of detecting results in the bonds being accepted when they should not have been. There has, for decades, been a lot of effort devoted to detecting the so-called kissing bonds. But even if such a method were established, discovering the problem at the end of manufacturing leaves absolutely no possibility of restoring the strength of the bonds to anywhere near their design capacity or, sometimes, even to achieving any increase in strength. At most, such “repairs” would simply conceal the existence of the zero-strength bonds. Worse, such defects would be global, not local the way porosity tends to be.

1274

L. J. Hart-Smith

The only bonds that are capable of being restored are cohesive failures from impact damage, which is usually found visually, provided that the repair is performed before contaminants have entered the cavity. With very few exceptions, adhesively bonded structures have demonstrated excellent durability in service. The problem is that processing errors, whether they were caused by inappropriate approved processing specifications whereby the process was being diligently but unknowingly done wrong or by undetected lapses in process control with processes that would have worked if they had been done right, have tended to involve hundreds or thousands of parts that needed to be fixed. Even so, the author knows of only one fatality to which a poor bond contributed – and, even then, that contribution was the result of an inappropriate choice of adhesive, not to a lapse in quality control or inspections at the time of manufacture. (Even then, there were many missed opportunities to prevent the final failure long before it happened; the nonbonds were merely the trigger that started the process, not the ultimate cause.) The problem with the interfacial failures has been one of cost to rebuild those structures, not safety. Most bonded structures have been in secondary structures. But, because of the few fleet-wide problems that have occurred, there has been a reluctance by most aircraft manufacturers to apply adhesive bonding as widely as it should have been, based on its technical merits. This reluctance has been compounded by the erroneous belief that doing so required better inspection techniques rather than better process control. It has been intensified by the observation that the most reliable manufacturers of bonded aircraft structures had a stable workforce, so that the skills built up could be maintained. This is contrary to the hireand-fire mentality in other countries in which the existence of specials skills is denied to support the prevailing business plans. Much of the porosity generated in large-area bonds could have been prevented by proper venting during cure. This can be achieved by a grid of small holes in doublers to overcome the tendency to pinch off the edges and trap any air inside. This was standard practice for the bonded fuselage panels on the SAAB 340 and SAAB 2000 commuter aircraft. The wings of the same aircraft were bonded in a special nonrigid bonding tool that totally eliminated all bond defects, resulting in far better bonded structures at a far lower price. This is described by Hart-Smith and Strindberg (1997). In this case, the vents were achieved by having loose fasteners in the determinate assembly holes used to keep the skin, doublers, and stiffeners located properly. This technique permitted a complete visual inspection of the parts inside the vacuum bag to ensure that they all fitted together properly before the parts were bonded together because the creation of short escape paths for any trapped air or volatiles eliminated the need for bulky bleeder packs which, in many other applications, did more harm than good because the associated bridging prevented the application of pressure in some areas. Because pressures are uniform inside autoclaves, there is no mechanism to displace the air in voids, only to compress it. This reluctance to apply bonded structures more widely has had one beneficial result, in the author’s opinion. Apart from the titanium-to-composite stepped-lap bonded joints used on the wings of the F/A-18 aircraft and on the tails of many military aircraft, there have been relatively few applications of bonded structures

44

Adhesively Bonded Joints in Aircraft Structures

1275

where the local loads have been of high intensity. This is a good thing, in the sense that bonded joints should never be allowed to become the weak link in the structure. Unfortunately, this need usually is not even acknowledged, let alone understood. This is easy to achieve when only thin or laminated components are bonded together. Rivetted joints are more appropriate for higher load intensities since the multiplicity of discrete fasteners retards the spread of any damage. A bonded design in which the adhesive is the weak link, even if the processes are all performed correctly, is tantamount to a weak-link fuse over the whole area, with the potential for damage to spread rapidly once started, as was explained earlier. In addition, there is not yet enough acceptance of what is needed to manufacture thick bonded structures. (Fokker was smart enough to use bonded laminated wing skins on the F-28 aircraft so that no single bond line was heavily loaded, even though the total load intensity in the skin was substantial.) This is most evident in the widespread failure to understand how to deal with induced peel stresses when the ends of the bonded overlaps are too thick. This problem does not arise with stepped-lap joints or laminated structures with the ends of the overlaps staggered, because the outermost end steps where the peel stresses would develop are always thin, which is consistent with Fig. 12. But these high-load bonded joints achieved their reliability because of their complexity, which comes at a higher cost. In the absence of a proper understanding of induced peel stresses, it has become fashionable to believe that a fracture mechanics analysis of the peel stresses is what is needed. The author believes that this approach has arisen because the standard shear-transfer analyses can be applied to adherends of any thickness and contains no interaction with peel stresses to cause a reduction in the predicted shear strength of the bonded joint. This, in turn, has arisen because, prior to the recent development of the SIFT model for predicting the strength of fiber–polymer composite laminates (Hart-Smith 2010), it had not been appreciated that there was no interaction. Also, the SIFT model does not require the assumed existence of a preexisting crack at the end of the bond line. The two cohesive failure mechanisms for the adhesive really are governed by independent criteria, both of which must be satisfied. The two mechanisms are governed by different power laws related to the adherend thickness, as noted earlier. It is not clear to the author how fracture mechanics methods developed for cracks in homogeneous materials can ever be related to the peel failure of adhesively bonded metallic structures or to delaminations in distinctly heterogeneous fiber–polymer composites; the calculations seem to him to have all the feel of fitting an analysis to test data by use of adjustable parameters that cannot be measured directly. Nevertheless, it is now common for such analyses to be carried out as part of the certification process. But, the author is unaware of any use of such output to improve the design of bonded structures. They seem to be accepted as an unavoidable burden. Indeed, in some cases, these analyses have caused the rejection of what experience has taught the author to bonded joints that were not critical in peel. Regardless of their technical merits, or otherwise, a valid peel-stress analysis which identifies that a bonded joint is critical under induced peel stresses is an unambiguous warning that the design is unsatisfactory because the peel-stress failures will undercut the potential shear strength of the joint. The most appropriate

1276

L. J. Hart-Smith

response to the results of such an analysis is to reduce the thickness of the adherends where the peel-stress failures are predicted to occur, at the edges of the overlap, until further analysis shows that peel stresses are no longer critical. This can always be done in such a way as to maintain the joint shear strength. Indeed, it can actually enhance the shear strength by eliminating the pinch-off that can occur at the ends of uniformly thick adherends. At this stage, the peel stresses would no longer govern, so even a crude analysis would suffice. The problem with now being capable of performing an analysis for peel stresses has led to a failure to employ the good design practices that were developed during the time it was not possible to perform any such analysis. The author’s concerns about the use of fracture mechanics as an analysis tool for bonded joints is exacerbated by its inherent incapability of establishing what the overlap should be because this is governed by the minimum shear stress in the middle of the overlap where there is absolutely no possibility of a preexisting crack that would allow fracture mechanics to be employed and where, even if there were, it would never show up as being critical in a fracture mechanics analysis because the bond stresses are almost zero there. In the same vein, overly complicated co-cured composite parts with embedded wrinkles in the fibers are often “passed” by ultrasonic inspections that detect no voids in the structure. Wrinkles in the fibers, which cause a far greater loss of measured strength than typical levels of porosity in parts that are rejected, do not count as defects because of the widespread acceptance of the myth that fiber–polymer composites are homogeneous. Sometimes, the complexity of co-cured parts is so great that detectable voids are created. But, no matter what is done after that point, it is impossible to straighten out the fibers to restore the strength of the laminates. In other words, the inspection comes far too late to permit the structures to be restored to their design strength, and, in some cases, there is not even an acknowledgment that there is a problem with the parts. The author would recommend that it is time to totally rethink the quality assurance programs for bonded and composite structures. It is known, from experience, which currently approved processes should be banned, and it is known what needs to be done to ensure low-cost defect-free parts by proper coordination of the parts and the bonding tools and by adherence to reliable processes. It is necessary to stop relying, instead, on the approach of pretending to inspect quality into the parts and assemblies after they have been bonded together because, by that stage, any defects that have occurred will ensure that the parts can never be restored to 100% of their design strength.

44.11 Bonded Repairs Bonded repairs should be looked upon as joints that are made at some time after initial manufacture. They need to be made under the same rules and procedures, but it is obviously not possible to repeat the very same surface treatments for bonded metal structures or even for composite structures if a peel ply had been used. In both cases, low-pressure grit blasting has proved to be the most reliable surface treatment

44

Adhesively Bonded Joints in Aircraft Structures

1277

for repairs. This is often followed by the application of a silane coupling agent when composite patches are bonded over cracks in metallic structures. Such patches have been found to be both very reliable and effective (Baker and Jones 1987); they have most often been used when there is no other alternative but to scrap the part because rivetted repairs are sometimes impossible or ineffective (not stiff enough to restrain a crack from any further growth). As noted earlier, local bonded repairs to bonded metallic structures that were made according to inappropriate specifications are an exercise in futility since the remainder of the structure is also about to disbond. Such structures need to be totally remanufactured in accordance with better processes even when the specific repair manual states otherwise. The most critical issue about the repair of damaged or disbonded composite structures is the difficulty of thoroughly drying the laminates before the repair is executed. Absorbed water takes a very long time to remove from both face sheets and honeycomb cores. This causes such parts to be out of service for a long time.

44.12 Other Industry-Specific Factors A little publicized fact about adhesive layers is that they act as electrical insulators. Today’s jet transports made from aluminum protect the passengers and crew because they are contained in a complete Faraday cage. There are a number of adverse consequences if that cage is interrupted by insulators between the individual metallic skin panels. The most obvious is that lightning strikes will cause far more damage if there is no continuous conductive path between the strike zone and discharge point. Indeed, simulated lightning-strike tests during the PABST program showed extensive local burning of the adhesive layers at the bonded joint between adjacent skin panels. This is why composite aircraft need special conductive coatings to compensate for the poor conductivity of even carbon fibers and for the periodic total interruptions through bonded joints. A lesser-known issue is that the establishment of even small potential differences between the panels on aluminum aircraft caused by the overzealous application of sealant to prevent corrosion can interfere with some of the small voltages involved in transistorized communication systems. The possible interferences with these systems that are responsible for the powering off of electronic devices during takeoff and landing are of concern. It is therefore important that the individual panels that make up the exterior skin of aircraft, and presumably rockets and missiles also, are grounded together with sufficient connections. The notion of an all-bonded structure is really a myth. Some minimum number of tightfitting dry rivets is needed to provide electrical connections. These are easily located in the many low-stress areas in bonded joints without causing any fatigue problems. A rivet hole in the middle of a bonded overlap, where the stress level is only half that of the surrounding skins, has a fatigue life some 20 times longer than in a rivetted joint between the skins. Once this need for some holes is acknowledged, it makes sense to use such holes for determinate assembly, to minimize the need for most of the traditional costly assembly tools, and to simplify bagging by eliminating the need

1278

L. J. Hart-Smith

for holding fixtures. This concept was used on the bonding tools to both simplify and improve the manufacture of the bonded stiffened wing skins used on the SAAB 340, as discussed by Hart-Smith and Strindberg (1997). There needs to be a balance between coating every fastener and every faying surface with sealant to prevent corrosion and the need for electrical continuity. Perhaps the cost savings from determinate assembly may provide the encouragement to include the necessary connections to create a true Faraday cage on aerospace structures. Another issue about bonded aircraft structures is that the surfaces created by anodizing and etching to enhance the adhesion of adhesive primers are more prone to corrosion than normal rolled or machined aluminum surfaces. It is therefore necessary to be very careful, particularly in bilge areas and around galleys, to create and maintain reliable corrosion protection. The polymers used in adhesives tend not to degrade with time and environmental exposure, but some do absorb water and other chemicals. Even so, the primary concern has been for the durability of the interfaces.

44.13 Examples of Use of Adhesive Bonding in Aircraft Structures Some aircraft manufacturers have made far more extensive use of adhesive bonding than others. De Havillands (now absorbed into BAe) and Fokker were the pioneers in using adhesive bonding in primary aircraft structures. The Fokker F-27 Friendship and F-28 Fellowship aircraft are renowned for their successful widespread use of adhesive bonding. SAAB and Cessna used primary structural bonding widely after the PABST program had validated the second successful generation of materials and processes. However, other major airframers have restricted the use of adhesive bonding to mainly secondary structures, primarily because of the failures created by the pre-PABST bonding processes and surface treatments developed in the USA. The most significant difference between the two levels of application would appear to be the stability in the labor force, for a variety of reasons. It is not the result of technical ignorance of what is needed to make bonding work, or of the benefits it can create. Figure 30 shows the extensive application of adhesive bonding to the aluminum airframe of the SAAB 340 aircraft. This aircraft has an incredible structural efficiency and durability, which could not have been achieved with conventional rivetted structures. Cessna made even more extensive application of bonding to the fuselage of the Citation III jet aircraft and used the same technology to make wings on other aircraft with far fewer fuel leaks than on conventional rivetted wing boxes. Figure 31 shows a typical frame/longeron intersection point on the Citation III fuselage skin, showing how not only the waffle doubler and longerons are bonded to the skin but the outer half of the frame is, too. What is significant about this design is that both stiffener flanges are continuous where they contact the skin; there is no weakness associated with the traditional mousehole in the frames to allow longerons to pass through.

44

Adhesively Bonded Joints in Aircraft Structures

1279

Fig. 30 Application of adhesive bonding to SAAB 340 fuselage, wings, and tail Frame inner member Rivetted secondarily To outer member

Frame inner member

Longeron joggled over continuous frame outer cap

Frame outer member Waffle doubler

Skin

Fig. 31 Cessna Citation III bonded frame/longeron intersection

The secondary structures, control surfaces, and fixed panels on the Boeing 747 made extensive use of metal bonding, mainly with honeycomb. Many of these components have been replaced by composite structures on later models, but most of these components should still be classified as bonded structures. For example, the 777 composite tails (see Fig. 32) are made by co-bonding pre-cured stiffeners to green skins. These are classified as primary structures, surpassing in both size and load intensities the earlier NASA-funded flight demonstration programs on the 737 and DC-10. All current-production Airbus aircraft have primary composite structures in their horizontal and vertical tails. Today, virtually all modern transport aircraft have composite control surfaces, wing-root fillets, and various other secondary structures.

Forward torque box panel bond assy (lower)

Jack screw fitting

Main torque box skin panel bond assy (Upper)

Aluminum

Titanium

Fiberglass

Tip assembly

Trailing edge ribs

Leading edge assembly

Forward torque box ribs

Main torque box ribs

Forward torque box panel bond assy (Upper)

Trailing edge panel bond assembly (Upper)

Trailing edge panel bond assy (lower)

Auxiliary spar

Pivot fittings

Seal panels

Front spar assembly

Rear spar

Trailing edge “T” beams

Fig. 32 Co-bonded composite primary structure on Boeing 777 tail

Seal panels

Centreline rib

Strakelet assembly

Main torque box Skin panel assys (Lower)

Tip rib

Tip assembly

Legend Carbon-fiber reinforced plastic (CFRP)

1280 L. J. Hart-Smith

44

Adhesively Bonded Joints in Aircraft Structures

1281

Military aircraft have made much more use of primary composite structures than on commercial transport aircraft, but most new general aviation aircraft since the resurgence of this industry a decade ago have made even more extensive use of composites, co-cured and bonded, in their primary structures. Their smaller size favors this method of manufacture. Adhesive bonding has a tremendous advantage over co-curing of composites for high production rates. Co-cured structures require the use of the largest tools for a greater time than bonding together of simple details, many of which can be cured in a single autoclave cycle. This was recognized by the team that developed the Lear Fan all-composite executive transport aircraft over 20 years ago. It was planned for a production rate of one aircraft a day. The tooling costs would have been unaffordable if the design had been based on large integrally stiffened assemblies. This issue is still important today. The minimum-cost prototype development program is often a co-cured design because the effects of production rate are not considered. Once the structural tests have been completed, the option of less expensive more dispersed production is lost, and high production costs are often locked in place. It would make more sense not to minimize the cost of the prototypes in isolation but to build prototypes of the lowest-production-cost design instead.

44.14 Conclusion This chapter has focused on adhesively bonded joints in aircraft structures, emphasizing why and how they work, when they are superior to mechanically fastened joints, some examples of great successes, and an explanation (with cures) for less than successful applications. There is more coverage of experience with bonded joints than the science that enabled them to work. Those details are found in much greater depth in other chapters in this book. However, there is also some discussion of devoutly held beliefs about bonded joints that are so wrong that they have inhibited the more widespread use of bonding because of unwarranted concerns about durability and monitoring degradation. Despite having been able to characterize the nonuniform transfer of load through bonded joints for decades, the most widely taught method simply, and quite erroneously, assumes that the load distribution is uniform to “simplify” design and analysis. Inspections have not been used wisely since so many of them are applied at the wrong time in the life of the product. Sadly, the most valuable inspections, visual, are downgraded as being too primitive to use because they rely on skill, a variable human factor, ignoring the fact that most of the instruments used can never find some of the most critical information that is sought. There is a tendency to conduct an inspection for reasons more related to the existence of an instrument that is supposed to be able to detect some defect rather than that there is any real need, or likely benefit, from the inspection. Worse, the most critical source of bond failure – the failure to adhere in the first place – is often not acknowledged as the most critical (because any such error will have global impact). Instead, there is too much focus on inspections for local damage, which the adhesive

1282

L. J. Hart-Smith

is often able to tolerate with no loss in strength by rearranging the load-transfer distribution. The reason for the misplaced emphasis is merely that it is possible to detect that kind of discrepancy. The hope is that, by exposing these issues, and how to address them, there will be a greater willingness in the future to use bonding where it really is the best approach – and to not knowingly take shortcuts (even if they are “approved”) that give it a bad name because the impact is so widespread. The confirmed success stories about adhesive bonding in aircraft structures are substantial and have been for decades; they just need better publicizing to neutralize the bad press from the failure cases, none of which can reasonably be blamed on ignorance. The author concludes with the following suggestion to encourage the more widespread use of bonded aeronautical structures. Waiting until someone devises a reliable after-the-fact inspection method for bonded structures before accepting the advantages of bonded structures, where they make sense, is going to take an extremely long time if it is ever accomplished. And, in any case, as explained earlier, such a technique could not enable the design joint strengths to ever be restored; it is too late to try to inspect quality into a part after it has been built. Some aircraft manufacturers have already demonstrated that carefully following appropriate processes is both feasible and reliable. The service record of properly bonded aircraft structures is exemplary. There is no need to wait for an additional safety net, for which no need has yet been found, to justify less diligence during the manufacturing processes before committing this proven technology to production. Acknowledgments Parts of this chapter have earlier been published as “Aerospace Industry Applications of Adhesive Bonding” (Adhesive Bonding, edited by R. D. Adams, Woodhead Publishing Ltd., Cambridge, 2005, pp. 489–527) (Reproduced with permission from Woodhead).

References Baker AA, Jones R (eds) (1987) Bonded repairs of aircraft structures. Martinus Nijhoff, Dordrecht, pp 77–106 Bardis J (2001) Effects of surface preparation on long-term durability of composite adhesive bonds. In: Proceedings of MIL-HDBK-17 meeting, Santa Barbara, 16 Oct 2001 de Bruyne NA (1957) Fundamentals of adhesion. In: Bonded aircraft structures. A collection of papers given in 1957 at a conference in Cambridge, England, Bonded Structures, Duxford, pp 1–9 Gosse JH, Christensen S (2001) Strain invariant failure criteria for polymers in composite materials. In: AIAA paper AIAA-2001-1184, presented to 42nd AIAA/ASME/ASCE/AHS/ASC structures, structural dynamics, and materials conference, Seattle, 16–19 Apr 2001 Hart-Smith LJ (1973a, 1974) Analysis and design of advanced composite bonded joints. NASA Langley contract report NASA CR-2218, January 1973; reprinted, complete, Aug 1974 Hart-Smith LJ (1973b) Adhesive-bonded double-lap joints. NASA Langley contract report NASA CR-112235, Jan 1973 Hart-Smith LJ (1973c) Adhesive-bonded single-lap joints. NASA Langley contract report NASA CR-112236, Jan 1973 Hart-Smith LJ (1973d) Adhesive-bonded scarf and stepped-lap joints. NASA Langley contract report NASA CR-112237, Jan 1973

44

Adhesively Bonded Joints in Aircraft Structures

1283

Hart-Smith LJ (1973e) Non-classical adhesive-bonded joints in practical aerospace construction. NASA Langley contract report NASA CR-112238, Jan 1973 Hart-Smith LJ (1978) Adhesive-bonded joints for composites – phenomenological considerations. Douglas Paper 6707, presented to Technology conferences, associates conference on advanced composites technology, El Segundo, 14–16 March 1978; in Proceedings, pp 163–180; reprinted as designing adhesive bonds. Adhes Age 21:32–37, Oct 1978 Hart-Smith LJ (1982a) Design methodology for bonded-bolted composite joints. USAF contract report AFWAL-TR-81-3154, 2 vols, Feb 1982 Hart-Smith LJ (1982b) Adhesive layer thickness and porosity criteria for bonded joints. USAF contract report AFWAL-TR-82-4172, Dec 1982 Hart-Smith LJ (1984a) Effects of flaws and porosity on strength of adhesive-bonded joints. Douglas paper 7388, presented to 29th National SAMPE symposium and exhibition, Reno, 3–5 Apr 1984; in Proceedings, pp 840–852 Hart-Smith, LJ (1984b) Bonded-bolted composite joints. Douglas paper 7398, presented to AIAA/ ASME/ASCE/AHS 25th structures, structural dynamics and materials conference, Palm Springs, 14–16 May 1984; published in J Aircraft 22, 993–1000 (1985) Hart-Smith LJ (1993) The bonded lap-shear test coupon – useful for quality assurance, but dangerously misleading for design data. McDonnell Douglas paper MDC 92K0922, presented to 38th international SAMPE symposium & exhibition, Anaheim, 10–13 May 1993; in Proceedings, pp 239–246 Hart-Smith LJ (1995) Reliable nondestructive inspection of adhesively bonded metallic structures without using any instruments. McDonnell Douglas paper MDC 94K0091, presented to 40th international SAMPE symposium and exhibition, Anaheim, 8–11 May 1995; in Proceedings, pp 1124–1133 Hart-Smith LJ (1997) Interface control – the secret to making DFMA ® succeed. McDonnell Douglas paper No. MDC 96K0132, presented at SAE aerospace manufacturing technology conference & exposition, Seattle, 2–5 June 1997, and published in Proceedings, pp 1–10, SAE Paper No. 972191 Hart-Smith LJ (2001) Explanation of delamination of bonded patches under compressive loads, using new simple bonded joint analyses. Presented to 3rd quarterly CRAS review, in conjunction with Fifth Joint DoD/FAA/NASA conference on aging aircraft, Kissimmee, 10–13 Sept 2001 Hart-Smith LJ (2001/2002) A demonstration of the versatility of rose’s closed-form analyses for bonded crack patching. Boeing paper MDC 00K0104, presented to 46th international SAMPE symposium and exhibition, Long Beach, 6–10 May 2001; in Proceedings, 2002: a materials and processes Odyssey, pp 1118–1134 Hart-Smith LJ (2003) Is it really more important that paint stays stuck on the outside of an aircraft than that glue stays stuck on the inside? Boeing paper PWMD02-0209, presented to 26th Annual meeting of the adhesion society, Myrtle Beach, 23–26 Feb 2003 Hart-Smith LJ (2010) Application of the strain invariant failure theory (SIFT) to metals and fibrepolymer composites. Philos Mag 90(31):4263–4331 Hart-Smith LJ (n.d.) The Goland and Reissner bonded lap joint analysis revisited yet again – but this time essentially validated. Boeing paper MDC 00K0036, to be published Hart-Smith LJ, Davis MJ (1996) An object lesson in false economies – the consequences of not updating repair procedures for older adhesively bonded panels. McDonnell Douglas paper MDC 95K0074, presented to 41st International SAMPE symposium and exhibition, Anaheim, 22–25 March 1996; in Proceedings, pp 279–290 Hart-Smith LJ, Strindberg G (1995) Developments in adhesively bonding the wings of the SAAB 340 and 2000 aircraft. McDonnell Douglas paper MDC 94K0098, presented to 2nd PICAST & 6th Australian aeronautical conference, Melbourne, 20–23 March 1995; abridged version in Proceedings 2:545–550 Hart-Smith LJ, Strindberg G (1997) Developments in adhesively bonding the wings of the SAAB 340 and 2000 aircraft. McDonnell Douglas Paper MDC 94K0098, full paper published in Proc Inst Mech Eng, Part G J Aerosp Eng 211:133–156

1284

L. J. Hart-Smith

Hart-Smith LJ, Redmond G, Davis MJ (1996) The curse of the nylon peel ply. McDonnell Douglas paper MDC 95K0072, presented to 41st International SAMPE symposium and exhibition, Anaheim, 25–28 March 1996; in Proceedings, pp 303–317 Krieger RB Jr (1988) Stress analysis concepts for adhesive bonding of aircraft primary structure. In: Johnson WS (ed) Adhesively bonded joints: testing, analysis and design, ASTM STP 981. American Society for Testing and Materials, Philadelphia, pp 264–275 Mahoney CL (1988) Fundamental factors influencing the performance of structural adhesives. Internal Report, Dexter adhesives & structural materials division. The Dexter Corporation, 1988 Marceau JA, Moji Y, McMillan JC (1977) A wedge test for evaluating adhesive bonded surface durability. Adhes Age 1977:28–34 Myhre SH, Labor JD, Aker SC (1982) Moisture problems in advanced composite structural repair. Composites 13(3):289–297 Schliekelmann RJ (1983) Adhesive bonding and composites. In: Hayashi T, Kawata K, Umekawa S (eds) Progress in science and engineering of composites, vol 1. Fourth international conference on composite materials, North-Holland, pp 63–78 Thrall EW Jr, Shannon RW (eds) (1985) Adhesive bonding of aluminum alloys. Marcel Dekker, New York, pp 241–321

Aerospace Industry

45

Christian Désagulier, Patrick Pérés, and Guy Larnac

Contents 45.1 45.2

45.3

45.4

45.5 45.6

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Specific Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2.1 Launchers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2.2 Spacecrafts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2.3 Where Are the Adhesive Bonds? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2.4 Which Adhesives for What? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2.5 What Are the Adhesive Selection Drivers? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.2.6 And the Adhesive Tapes? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Material and Process Characterization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.3.1 Structural Bonding or Not? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.3.2 What Specific Properties Do Space Applications Call for? . . . . . . . . . . . . . . . . . . Surface Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.4.1 The Case of Composites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.4.2 The Two-Part Epoxy Adhesive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.4.3 The Metallic Parts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.4.4 Aluminum Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.4.5 Titanium Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Nondestructive Inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Strength, Failure, and Durability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.6.1 Sandwich Structure Bonding Construction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.6.2 Monolithic Room Temperature Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.6.3 A Commercial Bonding Case . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1286 1288 1288 1289 1291 1294 1301 1301 1302 1302 1302 1314 1314 1315 1316 1317 1318 1318 1325 1325 1326 1327

C. Désagulier (*) ArianeGroup, Les Mureaux, France e-mail: [email protected]; CHRISTIAN.DESAGULIER@astrium. eads.net P. Pérés · G. Larnac ArianeGroup, Saint Médard en Jalles, France e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_45

1285

1286

C. Désagulier et al.

45.7

1329 1329 1330 1330

Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.7.1 By Way of Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45.7.2 And Finally . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Abstract

This chapter aims to introduce some of the specific bonding-related issues that face those responsible for bonding operations in space applications. We have to consider two cases: structures for launchers and structures for satellite where the environment in terms of load is different. In the case of launchers, extremely high mechanical loads over a short time occur, and for satellites, there is the gravity magnified by about 20, followed by exposure to extremely high temperatures over a very long period, with the observation and telecom satellites under zero gravity. To design for such broad specifications and to select materials, such as adhesives, it is just the first in a series of challenges. To perform tests on Earth reproducing space-like environments while subject to terrestrial gravity is the second and not the least significant. This chapter is intended to be an introduction to the issues which have to be considered together in the case of spatial bonding from the bottom, on Earth, upward, to space.

45.1

Introduction

At the beginning of the space conquest in the 1960s, the structural parts of launchers and satellites were mainly made of metal. Here and there, lighter aluminum foil or fiber glass-reinforced plastic composite parts replaced metal as the satellites grew in mass and volume in parallel to the launchers. The word “composite” began to be used to designate structures built up using several different types of materials from various origins (metallic or organic). One found, on one hand, the monolithic composite parts, which refers to the organic stacking of woven glass fabric at various angles, pre-impregnated with resin and cured under pressure. On the other hand, one found mainly 100% metallic, light plane structures made with aluminum foil skins bonded with an organic film to honeycomb core material. At the very beginning, based on conventional design reflexes, the metallic and organic parts were mechanically fitted together, and those first composite parts were reserved for no structural parts. This was also due to a lack of experimental feedback, i.e., of confidence, in the behavior of such polymeric parts in a spatial environment. For assembly, the mature aeronautical industry used nut/bolt and rivet systems after drilling, which came from the general mechanical industry. If saving mass was obviously an issue in aircraft, it was even more critical in spacecraft. Before escaping the grips of gravity, the energy required for launching is directly dictated by the weight, and the amount of energy is borne by the launcher. It was the space

45

Aerospace Industry

1287

industry that applied the aeronautical principles, as far as the arrival in the scene of composite materials was concerned. At the beginning of the 1970s, the epoxy resin associated with glass and boron fibers brought improved specific strength (strength divided by density), enabling the introduction of composite monolithic structural parts instead of metallic ones. Finally, carbon fibers arrived in the 1980s and with them, their assembly at the stage of primary structural parts, in monolithic truss structures or sandwich panels, as parts of the bodywork of satellites. At present, aircraft manufacturers are developing processes capable of one-shot 5-m-diameterwide, 20-m-long main structural parts. They use several processes like automatic tow of carbon fiber placement, 3D mixed carbon/aramid woven fabric, out-ofautoclave curing, and unidirectional or fabric, pre-impregnated with resin. Dry fibers preformed and consolidated after resin injection or with infusion process are also in production line. In conclusion, priority is given to innovation to make as much headway as possible to reduce the weight of planes, to switch the more numerous parts in gray aluminum to black ones. An Airbus A300–600 was 5% composite material in 1970; this proportion is ten times higher in Airbus A350 XWB today. The same trend can be seen in the launcher industry. The composite rate increased from 15% in Ariane 5 launcher to up to 63% in the new launcher Ariane 6 and even more in A6 with four booster versions where the level of composite reaches 63%. Thanks to the adhesive bond technology, 40% cost savings in production were achieved in aeronautics in comparison with riveting where the weight has been divided by 2.5 for some parts. Bonding in spacecraft dates back to the same era as composite materials, with composites including, on one hand, the previously mentioned fibers reinforced with plastic, like fiber glass, carbon, aramid reinforced with epoxy, phenolic, and thermoplastic resin and, on the other hand, lighter metal alloys, mainly such as aluminum and titanium. Even more so than aircrafts, spacecrafts were the ideal candidates for composite material application and especially carbon fiber-reinforced plastic (CFRP), characterized by high specific stiffness and specific strength, i.e., the ratio values of the Young’s modulus and strength to failure divided by the density. These specific mechanical properties associated with a low coefficient of thermal expansion (CTE) result in a light, highly loaded, dimensionally stable structure. For a designer, composite design generates potentially sizeable weight savings, keeping in mind that for space injection by a launcher, “mass is money.” Since the nut/bolt and rivet systems after drilling are dedicated to joining metal, bonding is mandatory for assembling composite parts, and the rules of designing and justifying holes are well known with metal. Because of its anisotropic behavior, designing and justifying a hole in a composite part is much more complex. In principle, local reinforcement is generally performed. The lay-up sequence is also optimized. Drilling composites is not a safe operation as delamination can occur. Since adhesives with thermo-structural resistance became available, bonding was implemented in aircrafts, more largely in helicopters than in airplanes, such as filling agents for rooms and load-wide spreading agents between the bolted or riveted metal parts.

1288

C. Désagulier et al.

Let us be reminded that in the 1960s, two phenomena appeared with metallic structures assembled with adhesive bonds: – Adhesive bond corrosion. It was the degradation of the oxide layer on the metal junction due to hydrolysis of the polymer of the adhesive. – Sandwich structure corrosion. This phenomenon was generated by the confined atmosphere in the cell with the presence of humidity and corrosive gas coming from the polymerization residues. Solutions have been founded with: – Hydrophobic adhesives – Deposition of anticorrosion primer based on chromate and suppression of plating aluminum – Improvement of the stability of the oxide layer with chromate or phosphorus anodic conversion (protection of honeycomb cells) The protection against atmospheric corrosion was also one of the first applications of the bonding agents between metallic parts, as wet barriers. With the arrival of composites, the issue of assembling composite and metal parts together was brought sharply into focus. If the resin formulators went a significant way to meeting requirements (mechanical, thermal), dedicated metal surface preparation prior to bonding had to be developed, most of the metals being subjected to surface corrosion with the production of brittle oxides, especially aluminum alloy. Prior to functional application, to bond, particular attention is paid to the behavior of adhesives in a space environment, that is to say, the vacuum, the thermal cycling, and the radiations. Most of the time, adhesives exhibit very different behavior according to the materials with which they have been brought into contact in the space environment against temperature; here, one is referring to the possible damage under thermomechanical fatigue. There is now a choice of composite material made from various compositions, fibers, resins, and adhesives, capable of dealing with the space environment provided that the preliminary selection fills such stringent specifications.

45.2

Specific Requirements

Adhesive bonds in the launcher and the spacecraft have an eminent place in the aerospace industry.

45.2.1 Launchers Launchers are built in several stages and assembled together. Each stage is separated from the launcher at different altitudes in order to release weight when propellant is

45

Aerospace Industry

1289

fully used. Depending on the architecture, solid or liquid propulsion or a mixture of both systems, the launcher has to be protected against the high temperature met on the surface of the structures. After 141 s of flight, the launcher Ariane 5 reaches an altitude of around 67 km at about 2,000 m/s! Ariane 5 is considered as a “powered launcher” due to its two solid boosters. The initial thrust delivered by the two boosters and the cryogenic engine Vulcain 2 is around 13,000 kN (90% coming from the boosters). The most demanding mechanical load occurs in the launch phase mainly during booster combustion with a very high acceleration (more than 20 g). Even if the propulsion phase with solid propulsion lasts around 2 min, a lot of vibration over a broad range of frequencies and heavy acoustic environment (150 db) are generated. Therefore, a damping solution is searched and adhesive bonds with damping materials are used. When cryogenic propellants are used for liquid propulsion, cryogenic insulation is needed to limit propellant evaporation ( 253  C for LH2 and 183  C for LOX) and ice formation.

45.2.2 Spacecrafts To illustrate the challenge when using bonding in spacecraft, here are some figures that will need to be dealt with. One must distinguish three phases which is the specificity of this task. In the first phase, lasting several months, one needs to deal with the integration of the satellite at the launch pad. Care has to be taken to avoid any risk of contamination and damage. As soon as the satellites are covered and protected by fairing, the atmosphere inside the fairing (temperature and humidity) is permanently controlled until the launch. The second phase lasts few minutes, starting from the ignition of the launcher up to the orbital injection of the payload (telecom, Earth, and deep space observation satellites, transportation vehicle for International Space). The most demanding mechanical load occurs in this phase during booster combustion with 20–30 g acceleration, 30 Hz/8,000 Hz vibration, and 150 dB, 5 Hz/120 Hz acoustical environment. The payload must be qualified to sustain these severe conditions. The separation of the stage from pyrotechnic system generates an energetic shock which could damage some sensitive equipment of the satellite (especially scientific satellites). This is followed by the orbital phase, of a very long duration, 15 years or more of orbital life for telecom satellites. The main environmental conditions with which compliance is required are temperature, thermal cycling, outgassing vacuum, radiation, and electrostatic drain (see Fig. 1). Here, one has to distinguish the spacecrafts dedicated to work in low Earth orbit (LEO) from those in geosynchronous Earth orbit (GEO).

1290

C. Désagulier et al.

Liberation SSO 800 km

LEO

GEO 36000 km /36000 km

500 km /1000 km

GTO 200 km /36000 km

MEO 20000 km /20000 km

LEO Environmental conditions (400km-800Km) *Temperature variations -90°C to 90°C; *Number of cycle/day = 15,4 *Life time : >>15 years *Vacuum: 10-4 torr to 10-9 torr. Main radiations : VUV (Vacuum Ultraviolet) Some radiative particles (98% electrons, 2% protons, Van Allen belt). ATOX (atomic oxygen) Micro meteorites and debris

GEO Environmental conditions (36000Km) * Temperature variations : -150°C to 120°C; * Number of cycle/day =1 *Life time :>15 years *Vacuum: 10-9 torr to 10-10 torr. Radiation (Van Allen belt, galaxy, solar, low VUV) Micro meteorites and debris

Environmental conditions in deep space *Temperature variations: can reach -180°C to +260°C * Vacuum: 10-4 torr to 10-14 torr. *Radiation *Micro meteorites *Atmosphere conditions of the planet (temperature, pressure and reactive gas)

Fig. 1 Overview of the environment conditions met by satellites in space

For the LEO applications, further issues regarding the external surface exposed to space are to be taken into consideration like atomic oxygen flux (ATOX) combined with solar ultraviolet radiation (UV), ions’ radiation, and various high-energy photonic exposure like gamma rays and micrometeoroids and dust collisions. The GEO satellites are mostly telecom orbits. In this case, the external equipment is subjected to a thermal environment depending on the Sun exposure, under a secondary vacuum for 15 years. The main source of radiation to consider is from the ultraviolet. Thermal cycling and radiation exposure met during long missions are the main mechanisms responsible of the aging of the structures. The space telescopes, due to their high thermal stability requirements, are confined in chambers. The Herschel and Planck telescopes built under the responsibility of Thales Alenia Space operate between 1.2 and 1.8 billion kilometers (L2 Lagrange point). The Herschel SiC mirror, the highest ever placed into orbit, and the composite Planck mirror, designed and manufactured by Airbus Satellites, operate at 250  C. As one can see, there are two families of constraints with which the bonding will have to deal successively: the constraints due to the launch, mainly static and dynamic, over a short period of time, and the constraints due to the space environment over a long period of time, mainly thermomechanical. Aging due to the thermal–mechanical stresses, mismatches between adhesive and adherents, metallic or composite, especially in extreme temperature variations, is the prime concern compared to mechanical fatigue.

45

Aerospace Industry

1291

If the spacecraft is manned, flammability in microgravity conditions, offgassing, and toxicity are added requirements as for simple aircrafts.

45.2.3 Where Are the Adhesive Bonds? Launchers Structural adhesive bonds are defined as any rupture of the adhesive bond will cause significant damage of the mechanical part and hence the loss of the mission. Thermal protections and thermal insulations are bonded to the structural parts which are made of metal or composite materials (see Fig. 2). External thermal protections of the stages are used depending on the design of the launcher and the aerothermal conditions. Because of their low thermal conductivity, structures made of composite materials must be protected against high thermal flux met during the flight. Thermal protection is required when the launcher is powered with solid boosters. Compared to liquid propulsion, solid propellant cannot be used to cool down the hottest part of the nozzle. The thermal protection is achieved with ablative materials Propulsion

Structure Thermal protection of the fairing (composite part) Thermal protection of the inter tank, Inter stage, Equipement bay (composite part)

Thermal insulation of the cryogenic tank of the upper stage Flexible bearing Thermal insulation of the cryogenic tank of the lower stage

Thermal protection of the booster (metallic part)

Internal protection of the motor case of the booster

Thermal protection of the cryogenic engine due to proximity of the boosters Thermal protection of the throat of the nozzle Thermal protection of the nozzle

Fig. 2 Ariane 5 launcher artist’s view – main thermal protection and structural bonding locations (ASL)

1292

C. Désagulier et al.

which are bonded onto the structural parts. For all these cases, thermal protection systems do not sustain general mechanical loads but only thermal loads. The design of adhesive bonds must take into account thermal gradients and thermal expansion mismatches with the structural part which is in general metallic. A detailed description of the nozzle design allows to see the complexity of this part due to various material properties mixing composite (carbon/carbon, carbon/ phenolic resin, glass/phenolic), elastomers, and metals (see Fig. 3).

Cut section of Ariane 5 booster showing propellant design Equipment displayed at the Paris Air Show in 2001 Throat inlet housing Throat ring

Flexible bearing

Throat inlet ring inlet ring. aft Nose ring forward Nose inlet housing Nose cap Forward end ring

Example of nozzle with its divergent (grey color) and its internal thermal protection (brown color) and the throat on top (black color) ASL

Cowl

Snubber assembly Forward exit cone liner Forward exit cone housing

Fixed Aft exit cone liner housing liner Fixed Aft end housing ring Flexible boot Structural support boot ring Bearing protector Compliance ring DC 93-104 insulation

Exit cone severance system

Cowl housing

Example of nozzle design with the flexible bearing Skirt Inter-stage skirt Vessel Bondline

Example of adhesive bond between vessel and inter-satge skirt

Fig. 3 Examples of adhesive joint solutions for solid rocket motor

45

Aerospace Industry

1293

Pushed by the constant need of reducing the upper stage weight in order to launch more payloads, engineers use composite materials and, as much they can, adhesive bonds. The first example is the sandwich structure used for the upper part of Ariane 5 launcher. The sandwich is made of two composite skins glued on an aluminum honeycomb core. This structure provides the best compromise lightweight/rigidity, and it is one of the best solutions to cope with buckling requirement which is very critical for slim structures. Adhesive bond solutions can be chosen to assemble two composite panels and to make a junction between an aluminum ring and a composite structure. An alternative solution of designing composite structures to be buckling resistant is to reinforce the structural part with bonded stiffeners (Collier et al. 2010) as used in aircraft fuselage (see Fig. 4). A good example of an optimized structure which uses adhesive bonds is the composite structure used in Ariane 5 to launch two heavy satellites (dual launch); named SYLDA (see Fig. 5). SYLDA is manufactured by Airbus Safran Launchers (ASL).

Satellite Satellite design is mainly governed by minimization of the weight of its skeleton structure for the benefice of telecom equipment or scientific devices. Two other stringent requirements are taken into account by satellite designers, the management of thermal exchange and the electrical “exchange.” The satellite is built with a central tube from which platforms for equipment, antenna reflectors, and solar panels are attached. The propulsion requires a high pressure tank located inside the central tube.

Composite panel

Doubler in composite Doubler in composite

Composite panel

Composite panel

Junction of two composite panels

Junction of sandwich or monolithic composite part with metallic ring (in blue)

Corrugated sandwich core (NASA)

Stiffened composite panel with Omega shape composite stiffeners (Collier 2010)

Fig. 4 Examples of adhesive bonds for structural parts of launcher

1294

C. Désagulier et al.

Fig. 5 View of SYLDA structure (height, 4.6–6.4 m) made in sandwich with carbon fiberreinforced plastic designed with adhesive bonds

All these structures are made of composite materials which provide lightness, stiffness, and thermal stability. Metal or ceramic parts are glued on composite materials (see Fig. 6). Junctions with bolts are used for assembling electronic equipment and mechanisms of jettisoning with constraints (dismounting requirements). The central tube is actually the vertebral column made of sandwich using high performance carbon fiber. The equipment platform made of a sandwich structure is linked to the tube via struts also made in composite. The parts are bonded with bolts via metallic interfaces. When the strut is designed with metallic straps, an adhesive bond with a composite part is used. The design of reflectors is driven by high thermal stability requirements. These requirements can be fulfilled by using composite materials with specific carbon fiber and lay-up sequences. All the parts are stuck together (see Fig. 6). Solar panels are made of composite frames where all the solar cells are glued on it. Mirrors or optical systems can be glued on their stands. Very large mirrors like Herschel’s mirror (diameter of 3.5 m) made of silicon carbide cannot be manufactured in one piece. The retained solution was not to use glue to assemble the eight petals but brazing because of its better resistance to dynamic loads during the launch and radiation in its space life.

45.2.4 Which Adhesives for What? For the most part, adhesives are formulated in order to offer behavioral properties related to those of the composite and metal substrates, regarding the fluctuating and extreme stresses flowing between the adherend and the adhesive, both having to work together.

45

Aerospace Industry

1295

Central tube (ASL) Overview of telecom satellite architecture

Platform for equipment Courtesy Airbus

High pressure tank (ASL) Reproduced courtesy of EADS Astrium GmbH

Mirror (Herschel) Courtesy Airbus

Reflector (ASL)

Solar panels (ASL) Fig. 6 Satellite exploded view – various cases of structural bonding on all the parts displayed

General Design for Launchers The structures using composites such as interstage, intertank, equipment bay, and fairing are submitted to static, dynamic (vibration, acoustic, shock), and thermal loads. Pushed by the competition with metallic solutions, designers try to use as much as possible aeronautics composite materials. The aeronautical industry uses pre-impregnated epoxy resins. Because of their limited temperature resistance, composite parts submitted to thermal aerodynamic fluxes are protected. One of the solutions is to glue thermal protection panels with silicone or epoxy-based adhesives (see Fig. 7a). But this efficient method is expensive and a projection process is preferred (see Fig. 7b). However, all the thermal protection cannot be sprayed. The spray process can be easily automatized with robot assistance and does not need adhesive since only a surface preparation or a primer is required. The same approach is followed with the thermal insulation of the cryogenic tanks (see Fig. 7c). Cryogenic insulation could be done by gluing foam panels on the aluminum alloy tank or by spraying the foam. Manufacturing of composite sandwich structures can be done with a two-step process. Preliminary cured skins are glued with a honeycomb core. In order to reduce

1296

C. Désagulier et al.

a) Thermal protection of interstage with cork panel Courtesy of Airbus Defence and Space Systems Spain

b) Example of sprayed thermal protection (ASL)

c) Ariane 5 cryogenic tank –thermal insulation bonding (CRYOSPACE image) Fig. 7 Example of thermal protection of the structural parts of the launcher

cost, the structure can be done in one step usually called co-curing process. In this case the skins are cured together with the core with or without an epoxy adhesive film in between. Again when composite structures are made with a skin reinforced with stiffeners, both solutions are also used depending on the complexity of the part. In case of co-curing process, stiffeners made of resin pre-impregnated carbon fiber are pre-shaped, sometimes semi-cured and laid down on the mold. The second step consists in making the skin with same process of pre-impregnated deposition. At the end, both skins and stiffeners are cured together. In general an epoxy adhesive film is used. The transmission of the loads between composite structures is generally done with an aluminum interface. The rationale comes from the separation requirements of the different structures of the launcher during its flight. The junction of the metallic interface with the composite could be done with two composite or aluminum splices bolted or glued together. Because of the toroidal moments generated at the interface of each stage, designers are reluctant to use adhesive joint solutions to transmit the loads. The

45

Aerospace Industry

1297

performance of adhesion is driven by shear, and peeling effects due to bending must be avoided. In some cases, adhesive solutions are selected as illustrated by SYLDA design (see Fig. 5). Drilling holes in composite structures leads systematically to local reinforcement of the structure, and when weight optimization is searched, adhesive bonding is one of the solutions especially for the structures located in the upper part of the launcher close to the satellites where the ratio weight/performance is 1/1. Solid propulsion uses a lot of adhesive bonds: – The first case is one of the most critical interfaces between propellant and internal thermal protection. Both are complex materials and the mechanisms of adhesion are not simple to solve. This interface called liner is generally formulated by the manufacturer of the propellant who has also a great knowledge of the composition of the thermal insulation, generally an elastomer (EPDM, NBR, etc.)-based material reinforced with fillers and sometimes with chopped fibers. – The second one is the interface between the protection and the vessel which can be metallic (e.g., steel for Ariane 5) or composite (e.g., Vega). Specific adhesives are also required to be compatible with the elastomer and epoxy resin used in the composite. Adhesives formulated with phenolic resins are generally efficient. – The third one concerns the composite motor case design. The composite vessel is bonded to the composite skirts via a compliant material, the role of which is to adapt the deformation mismatch of the two structures (mainly hoop strain of the vessel and the longitudinal strain of the skirt). The junction of the composite structures with the compliant material is done with an epoxy adhesive. The nozzle design requires proofed adhesive joints because of the harsh thermal environment. All the pieces such as the throat made on carbon–carbon composite and the divergent made of carbon and or glass fiber-reinforced phenolic resin are glued to metallic parts. Epoxy-based adhesives are generally used. The last example can be illustrated by the flexible bearing. This a laminate made with elastomer reinforced with sheets of metal or composite. All the elements are glued together. Flexible bearings are used for the steering nozzle. This concept can be also used for the load transmission from the booster to the cryogenic tank (A5 concept) with a specific function of filtering vibration and then participating in the payload comfort. One can summarize that launchers which are designed with composite structures use adhesive solutions when no alternative solution is available as seen for the motor case with its nozzle for solid propulsion or when significant advantage in weight reduction is gained compared to bolted or riveted solutions. Adhesives such as epoxy, phenolic, and silicone-based products are usually provided by different suppliers (3M, SOLVAY, STRUCTIL, HUNSTMAN, LORD, DOW, etc.). Some adhesives could need a specific formulation to meet the required thermomechanical performance or to be adapted to specific processes. This last point is crucial because adhesive suppliers are reluctant to develop specific adhesives for a narrow or niche market (complex formulation and low volume).

1298

C. Désagulier et al.

General Design Purposes for Satellites Satellite designers use a lot of solutions based on adhesive bonds. Adhesive bonds allow: – – – – – – –

To save weight To easily assemble tiny pieces To not degrade composite structures To improve vibration damping To cope with tightness requirement To avoid electrochemical coupling effect To link different types of materials which have significant different thermal expansion coefficients

Sandwich structures are manufactured with standard processes or with co-curing process in the same manner than sandwich structures for launchers. But we can highlight that high modulus (polyacrylonitrile based) or very high modulus (pitch based) carbon fibers are preferred to increase the rigidity and the thermal stability. For electrical constraints, aluminum skin is used to manufacture sandwich panels. The central tube of more than 5 m height can have more than 100 interfaces. A lot of internal (propellant tanks junction) and external (communication modules webs, Earth panel) metallic inserts (aluminum, titanium, invar, etc.) are glued with epoxybased adhesives (see Fig. 8a, b). Aluminum inserts, frames, and titanium telecom reflector parts are bonded by bicomponent epoxy adhesives that cure at room temperature (see Fig. 8c, d). Control of heat exchanges inside the satellite is crucial to provide thermal stability of the optical instruments and electronic devices. The internal temperature is around 25  C, and the external temperature can vary from 100 to 100  C depending on the mission, the orbit. The conditions of the environment with respect to the orbits are reported in Fig. 1. Heat exchanges with the environment being only by solar radiation, the structure of the satellite is protected with solutions providing the appropriate thermo-optical properties (multilayer insulation system, paintings, etc.). Optical surface reflectors are small pieces glued on the surface exposed to solar radiation. The list of adhesives which can sustain the harsh environment met in space is limited, and silicone-based adhesives are often selected because of their stability at high temperature (up to +250  C) and are easy to handle. For instance, solar panels are glued on the substrate with a silicone-based adhesive. But silicone is not selected to stick mirror because of contaminant risk. Reentry vehicles are designed with a front shield able to sustain the high aerothermal fluxes met during the high speed flight in the atmosphere of the planet (Mars, Venus, Saturn’s satellite Titan, etc.). Ablative thermal protection is usually used and the tiles are glued on the structural part. Epoxy adhesives and room temperature vulcanized adhesives are commonly used. For instance, a space-grade silicone-based adhesive has been used with success for various spacecrafts

45

Aerospace Industry

1299

a) Aluminum insert embedded in sandwich panel (ASL)

c) Rear reflector face bonded to composite monolithic backing structure (ASL)

b) Wedges glued on sandwich panel (Courtesy Airbus Satellite)

d) Titanium alloy ASL reflector antenna fitting after surface treatment and bonding

Fig. 8 Example of assemblies with adhesive bonds

developed by Airbus Safran Launchers such as Beagle and ExoMars for Mars exploration and Huygens for Titan exploration (see Fig. 9). The same adhesive was used for bonding and also for sealing the junction lines between panels or tiles. This adhesive choice is driven by a complex dynamic distribution of the thermal flux with regard to the time against the deviation of the thermal insulation properties. The choice of room temperature vulcanized silicone has more to do with the thermal resistance of that particular adhesive family than shear strength. In such an application, a minimum of 0.5-MPa-out-of-plane resistance is currently specified. In applications at such high temperatures, the choice of a silicone adhesive combines the mechanical resistance with a thermal insulation. No toughening properties are required since there is no creeping against the mechanical load. Engineers in charge of the selection of adhesives must be aware of the contaminant risk which is a major source of satellite failure: – – – –

Weakening of the performance of the optical system Damage of the electronic equipment due to electrostatic discharges Weakening of the electrical production of the solar panels Modification of the thermo-optical properties (emissivity, absorption, etc.)

1300

C. Désagulier et al.

Bonding of cork tile on Beagle 2 font shield (ASL)

Exo Mars front shield (ASL)

Beagle 2 heat shield assembly (ASL)

Huygens front shield 2.7 m in diameter, 60 half-angle conic spherical tiles (ASL)

Fig. 9 Examples of reentry vehicle design with ablative front shield

On can summarize that different types of adhesives are used depending on the following applications: – – – –

Load transmission Thermal continuity Electrical continuity Optical compatibility

For long mission in space, organic materials must be protected against the radiation. This is the reason why an aluminized blanket covered the thermal shield of Huygens during the 7-year journey through the solar system via Venus, Earth, and Jupiter up to Saturn and during the 22 days of ballistic phase between the jettisoning of the Cassini orbiter. Cassini spacecraft was launched from Earth on October 15, 1997; Huygens landed on Titan on January 14, 2005. Keep in mind that while anything can be bonded with anything else, there is no universal adhesive despite the slogan for space applications: the main driver is the temperature of the joint. This means a selection of adhesive which is compatible of the thermal expansion mismatch between the adherends. Films or pastes are selected according to the environmental specifications to which they are subjected, as already said (mechanical and thermal), with respect to the design of the parts and the implementation of the bonding assemblies. For instance, the dimension of the part, especially if the parts are at the scale of the

45

Aerospace Industry

1301

launcher, 5.4 m in diameter, can dictate the process, i.e., the choice of a room temperature adhesive, for economic reasons (there is always an economical aspect in bonding to be taken into account): this prevents investing in an autoclave which a hot temperature adhesive would require. But this investment can appear as mandatory for technical reasons; one can understand that for the most parts, the higher the curing temperature, the higher the thermal resistance.

45.2.5 What Are the Adhesive Selection Drivers? As one has seen, adhesives for space applications are designed to withstand loads of mechanical origins during launch and from thermal origin in orbit, while the fitted parts have to deal with different thermomechanical loads. Bonding is a dedicated process to assemble dissimilar components and requires a rigorous approach to selection according to its properties, not only chemical but also physical, such as viscosity, in order to deliver the controlled thickness of the joint during implementation and restrain the creeping during polymerization. Aside from that, the bonding line control means to point out, at least in practice for a two-component room temperature (RT) paste, (1) the ratio of the mixing, (2) its homogeneity, (3) the pot life of the mixing at RT, (4) the pressure for full intimacy during the hardening, and (5) the minimum waiting time for hardening before handling, all these precautions observed and carried out under regulated temperature and hygroscopic conditions. To give a preliminary example, it is not rare that one has to bond a low CTE titanium part to a high CTE two-part epoxy adhesive. The issue that the design office and the material and process experts have to deal with is to guarantee a cohesive failure at the predicted margin level.

45.2.6 And the Adhesive Tapes? The space industry makes extensive use of adhesive tapes. The major formulator companies propose such space-grade tapes as Kapton, Tedlar, and Teflon, polyester or foam coated with various very high bonding capabilities, sensitive to pressure. Kapton tape is the most appreciated with respect to the high temperature resistance of Kapton, intensively used as dielectric insulator of solar cell panels. The Kapton tape is associated with acrylic or fully reticulated silicone adhesives: outgassing is the driving factor for the choice of the adhesive part of the tape, the unwanted event being the bubbling between the tape and the adherends under vacuum. Acrylic foam is also widely used in association with Velcro fastening. The Velcro fasteners, nylon based and easy to manage, were bonded 10 years ago with space-grade silicone (Dow Corning 93500). Its high release properties led to implement the silicone at the end of all epoxy bondings. Occasionally, with respect to the manufacturing flow chart, this was time consuming and required a mandatory dedicated location to insulate the silicone adhesive implementation from other adhesive families, in space and in time. Since Velcro coated with self-adhesive very high bonding acrylic

1302

C. Désagulier et al.

foam is available, it can be easily implemented at anytime and anywhere according to the breakdown of the integration process by simply pressing it on by hand.

45.3

Material and Process Characterization

45.3.1 Structural Bonding or Not? Roughly, two kinds of bonding can be identified: – Structural bonding where high mechanical and thermal loads are required, designed by the strength of cohesive failure of the adhesive or the adherends – Bonding subjected to loads where the properties of the adhesive far exceed the requirements of the adherend assembly whatever the strength to failure, in the adherends and in the adhesive Whatever the definition might be, for space applications, the loads taken into account for the design are those calculated, increased by the margin factor for the considered application, and the cohesive failure mode (in the adhesive or in the adherends) will be the mode expected. As a result, one will admit that bonding shall be referred to as “structural” when the bonded joint is designed with respect to the safety policy margins, the higher the residual allowance, the better the design. In other words, it means that, most of the time, the adhesive joint should have high thermomechanical properties often approaching the ultimate ones in the interests of optimization, i.e., mass saving, the higher residual allowance being not far from the ultimate. Considering metal to metal bonding (aluminum alloy, titanium alloy, stainless steel, Invar), composite to metal (carbon, glass, aramid fiber-reinforced resin, epoxy or cyanate based), and composite to composite, the adhesion level will depend on the ability of the adhesive and the surface adherend preparation to sustain the combination of various mode stresses (3D static, dynamic) with loads displayed in peeling, cantilever, and pull-out configurations, rarely according to a pure shear strength to which the adhesives are dedicated: there lies the challenge.

45.3.2 What Specific Properties Do Space Applications Call for? Before implementation in space equipment, adhesives are subjected to extensive characterization, from the elementary state according to standard tests and in context through representative working configurations, in terms of geometry and adherends, for validation and finally tested at the real scale for qualification, all properties having to comply with the specifications according to the predicted space environment enhanced by security margin company policy.

45

Aerospace Industry

1303

Selection and validation of adhesives are based on their intrinsic properties which must be compatible with engineering and manufacturing process requirements (ECSS-Q-70-71A rev 1). Therefore, the following parameters are generally determined: – Pot life ASTM D-2471 – Viscosity (key parameter for processing which is sensitive to environmental conditions: temperature, humidity); ASTM D 1824 – Thixotropy (property required when adhesive deposition is handmade) – Gel time (determination of time of adhesive processing before it reaches solid state, interaction with tooling and pressure application) – Degree of curing (compatibility with the manufacturing process of the hardware, sensitivity to stoichiometric deviation in case of combining multiparts) The key properties of adhesives for space environment are: – Glass transition temperature (Tg): ASTM D-3418 – Outgassing This test expresses collected volatile condensed matter (% CVCM), total mass loss (% TML), and recovered mass loss (% RML) according to ECSS Q-70-02A. This test is performed in vacuum with or without thermal cycling. ECSS-Q-70-20A specifies 100 cycles. The temperature range covers the application conditions. It could be more than 150 to +150  C. The adhesive for space qualification does have an amount of RML and TML less than 1% and a CVCM less than 0.1%. These acceptance limits are submitted for derogation depending on the absolute quantity of adhesive implemented: – Thermal coefficient of expansion (CTE): ASTM D-696 – Moisture ingress: ASTM D-1151 and D-570 It is well known that a plasticization phenomenon occurs in epoxy polymers which affects the overall characteristics such as Tg. General mechanical properties of the adhesive (Young’s modulus, shear modulus, elongation at break) are obtained by performing: – Tensile tests (dog bone sample ASTM D-628, EN ISO 527-2 -butt joint; ASTM D2095) – Shear test (thick adherent shear test ISO 11003-2; double lap joint ASTM D-3528) Single-lap shear tests (ASTM D-1002, ASTM D-3165, ISO 4557) are very often used because test samples are easy to manufacture and to test.

1304

C. Désagulier et al.

However, it is well known that stress field in a lap joint is rather complicated and that relationship between failure and true bond strength between adhesive and substrate cannot be determined with confidence. This test is generally used for quality control acceptance of adhesives: – Shore D Hardness (ASTM D-2240) In general, the validation of the adhesive is based on several test samples with a minimum of five tests for each type of test, and some of the tests are done for three batches of adhesive to cover the variability of the product. The validation phase of the adhesive is also linked to the final application. Shear tests are, for instance, tested at different temperatures and for different conditions of curing. Moreover two additional tests are sometimes performed. The Arcan test allows combining tension/shear and compression/shear modes. This test is well appropriated to validate failure criteria of the adhesive and to identify plasticity. Due to the complex behavior of the adhesive, this test must be accompanied with simulation being able to take into account plasticity phenomena. Plastic yield behavior depends on both the deviatoric (von Mises) and the hydrostatic pressure. This model allows describing the difference observed when the adhesive is in tension or in compression. Such criterion proposed by Raghava et al. (1973) is rather effective to predict quite well biaxial test results. More sophisticated models such as those proposed by Drucker and Prager (1952) and Prager (1995) are not yet fully used by design offices. They prefer to deal with conservatism approaches based on experimental test campaigns. To justify damage tolerance initiated by a lack of adhesive or the presence of porosity, the energy of crack propagation in mode I (GIC) is also characterized. Double cantilever beam test (ASTM D-3433, ISO 2009) is the most common test to measure GIC. This test is not easy to master due to the strong influence of the boundary conditions. Such test can be performed with composite adherends. Best results are obtained with unidirectional composite adherends; otherwise, damage generally occurs between the first and the second ply inside the laminate. In space applications, high modulus carbon fibers are often used to improve the ratio stiffness/density enabling a better performance in buckling. However, due to the lower adhesion properties with epoxy resin compared to composite made with high resistance carbon fiber, the performance in adhesion with very high modulus carbon fiber is often less. Fracture test on bulk adhesive coupons is not a common test because it is not always possible to manufacture such coupons without microporosity and introducing machining defects. Other characteristics are determined depending on the application. It could be: – Electrical resistance: Composite materials and adhesive joints are not good thermal and electrical conductive materials compared to aluminum alloys. This is one of the drawbacks of composite materials.

45

Aerospace Industry

Thermal conductivity Electrical resistivity

1305

Aluminum alloy 150 W/m. C Epoxy resin (0.18–0.20 W/m. C) CFRP 54 W/m. C in the fiber direction, 1 W/m. C in the transverse direction Aluminum alloy: 4.6–4.8 μohm.cm Epoxy resin: 1E21–6E21 μohm.cm CFRP: 97,100–287,000 μohm.cm in the fiber direction and 210,000–1.58E6 μohm.cm in the transverse direction

Research programs have been launched to improve these properties by adding carbon nanotube fillers. If increasing electrical conductivity has been obtained at laboratory scale, no industrial application has come out. The solution used by designers to guarantee the continuity of the electric conductivity is to fix metal braids with conductive glue enriched with silver fillers: – – – –

Thermal cycling ECSS-Q70-04 Offgassing of toxic materials: ECSS-Q70-29 Flammability: ECSS-Q70-21 Particle and UV radiation: ECSS-Q70-06

The verification of the adhesive bond is based on several tests performed on samples which have to be representative of conditions applied on the real part. This is the key issue because the statistical analysis of the main nonconformances of adhesive bonds is generated by wrong cleaning or bad surface preparation (particulate and molecular contaminations) and poor adhesive processing (insufficient training of the operator, bad monitoring of the curing process, etc.).

Tests for Adherence Verification The objective of these tests is to verify the adherence level (load or energy to separate the bodies linked together) in different environments met during the whole life of the parts. Several tests generally recommended by space agencies are used by the space industry. ESA publishes numerous ECSS documents elaborated with the industry and laboratories in Europe. Peel and shear tests are commonly performed and cleavage as well. For sandwich structures, flatwise tensile tests (ASTM C-6297) and edgewise compression tests (ASTM C 364) are usually performed. For thermal protection glued on the surface (metal or composite), flatwise tensile tests or peeling tests are performed (see Fig. 10). The adherence verification is based on a test matrix which must be compliant to the justification requirements. Peeling tests (ASTM D-1781, D-1876, D-3167, D-3933, ISO4578) are very often used at the beginning of the development giving a first qualitative estimation of the adherence. Illustration of a good or a bad adhesion based on the load curve and failure mode is given in Fig. 11. But care must be taken with peeling test definition

1306

Peeling test (Bell method) ASTM D-3167 floating roller peel test

C. Désagulier et al.

ASTM D-1002: Single lap test specimen

ASTM D-3528: Double lap test specimen for thermal protection

Flatwise tension test ASTM C-6297 Overview of the test set up

Edgewise compression test used for sandwich ASTM C364

Double cantilever beam specimen (DCB), providing a means of determining G1c ASTM D-3433: Cleavage flat adherend specimen

ASTM D-3528: Double lap test specimen for elastomer adhesion

Flatwise tension test ASTM C-6297 Detailed view of the sandwich sample test sample

Arcan test set-up

Fig. 10 Examples of main tests used for adherence measurement

especially when it concerns adherence with elastomers. The peeling test is a geometry-dependent test. This is the reason why the verification of the adherence is much more reliable when determined with double shear tests and flatwise tensile tests for which cohesive failure is required for both.

45

Aerospace Industry

1307

700 600 500 400 300 200 100 0 0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150 160 170 180

Load (N) /displacement (mm) curve Example of good adhesion (rupture of the elastomer without peeling)

500 400 300 200 100 0

0

100

200

300

400

500

Load (N) /displacement (mm) curve Example of good adhesion (rupture inside the adhesive)

400 300 200 100 0

0

100

200

300

400

Load (N) /displacement (mm) curve Example of bad adhesion (rupture in between the adhesive and the elastomer)

Fig. 11 Illustration of different peeling behaviors of adherence between metal and elastomer

1308

C. Désagulier et al.

The environmental effects are taken into account, and the main difficulty is to find out the right method to accelerate aging when performed on ground which is representative of the real life in space environment.

On Ground Moisture ingress combined with temperature and mechanical loads is the main contributor of adhesive performance reduction. It is important to highlight that water is not an active factor of aging but rather than a powerful actor to reveal and to enhance defects at the interface or the interphase of adhesive joints. If accelerated aging is generally performed, natural aging is sometimes considered for specific cases. Aging cycles combining temperature and moisture are determined via Arrhenius’s law. Care must be taken to choose adhesives from which moisture ingress is reversible (Fickian diffusion) in order to avoid any damage of the molecular chains, especially at the interface with the adherends. The effects of moisture ingress are threefold: – Plasticization of the adhesive and adherends in some cases. This affects the mechanical properties (Tg, for instance) and modifies internal stress distributions in the joint. – Hygroscopic expansion which could affect the stress field in the adhesive. – Weakening the interface/interphase. The wedge test (ASTM D-3762, ISO 15107) is an effective test to determine the bondline durability as a function of the adherend surface preparation whether it is metal or composite. This test allows discriminating though adhesives with the environment effect from brittle ones. Though adhesives enable the crack generated by the wedge to grow slowly and stop, for brittle adhesives the crack grows faster running out of the specimen length. The stability of the crack is time dependent with respect to the environment, usually the combination of temperature and humidity (see Fig. 12). Durability tests performed by Popineau (2015) on composite/aluminum adhesive joint have revealed the mode of failure changed during aging. Cohesive rupture observed in the adhesive in the initial state became adhesive with a rupture at the interface adhesive/aluminum. This test shows that care has to be taken with hybrid joints where the mechanism of osmotic pressure can be responsible for the performance reduction of the adhesive joint.

In Space For satellite design, the following environmental factors are systematically taken into account: – – – –

Vacuum (outgassing) Thermal cycling Electromagnetic radiation (UV, X, g) Particles (electron, proton)

45

Aerospace Industry

1309 700 600

40 G1C J.m2

Tg°C

500 35 30

19°C

400 300 200

35°C

100

25 0

200

400

600

800 1000 1200

t0,5 a) Bulk EC2216 adhesive Evolution of Tg (°C) with time

0

0

20 40 60 80 100 120 140 160 180 200

days b) Wedge test Evolution of G1C with time

Fig. 12 Effect of moisture with time (a) Tg evolution with EC2216 adhesive when immerged in water at 35  C, (b) evolution of energy release rate GIC of composite/aluminum joints immerged in water at 19  C and 35  C (Popineau 2015)

– Atomic oxygen – Micrometeorites For manned spacecraft, two additional tests are required: Offgassing represents the substances liberated by materials in microgravity that are harmful to the crew. Offgassing has to be avoided (ECSS-Q-70-29). The flammability test measures the capability with which a material is set on fire (ECSS-Q-70-21, ISO 22538 1-4 dealing with oxygen safety). For satellites, the main problems to overcome are linked to: (a) Thermal cycling in vacuum The most critical properties (mainly shear properties) are evaluated before and after thermal cycling according to the aging test agreed upon in the space industry. ESA specifies standard aging of 100 cycles under 10 5 torr vacuum, between 100 and +100  C with a heating and cooling rate of 10  C/min with a 5 min dwell (ECSS Q 70 04 A). It is not rare that 2,000 cycles are implemented. The thermal cycling is subjected to deviations with respect to the particular operational needs. For information, this thermal cycling is to bring closer the LEO external equipment typically subjected to a [ 90  C, +90  C] temperature range, under [10 4 torr, 10 9 torr] vacuum range, with 16 cycles/day and that of the GEO is exposed to a [ 150  C, +120  C] temperature range, under [10 9 torr, 10 10 torr] vacuum range, with 1 cycle/day. These figures depend on the spacecraft mission, the location of the equipment on the spacecraft regarding its exposure to the Earth and the Sun, and the properties of the materials in relation to their thermal balance. So, if the properties of the materials, i.e., the adhesives, are assessed after thermal cycling, the equipment at its final stage of manufacturing is also thermally cycled under vacuum. For instance, the telecom antenna reflectors in GEO are qualified in the [ 180  C, +180  C] thermal domain, requirements due to a stabilization of the composite on Earth at the

1310

C. Désagulier et al.

end of manufacturing (monolithic composite stress relieving, completion of the chemical conversion of the resins and adhesives, complete degassing of all the materials and assemblies). For the cases of telecom reflectors, the final thermal cycling aims to guarantee a misshaping variation under operational temperatures compliant with the focus control of the radio-frequency signal against the temperature directly exposed to the Sun or completely shadowed by the Earth, ranging with the seasons. In this respect, the thermal load cases for distortion analysis are assessed in order to highlight the most severe cases with respect to the maximum and minimum temperatures on the different parts of the reflector shell. With respect to this, the various cases below are distinguished: – – – – –

EQBOL, equinox period at the beginning of reflector’s life SSEOL, summer solstice at the end of reflector’s life EQEOL, equinox period at the end of reflector’s life WSEOL, winter solstice at the end of reflector’s life OSRBOL, optical surface reflector reflection at the beginning of reflector’s life

Depending upon the telecom mission, an EQBOL may correspond to the shell cold case and the maximum gradient, and the OSRBOL may correspond to the shell hot case. It is impossible to reproduce the space environment for ground testing of space system elements because of the variety and complexity of the environments which can generate more or less degradation of the material. The reliability of the test results depends on simulating the critical effects of the space environments for a particular mission (ISO 15856 Simulation guidelines for radiation exposure of nonmetallic materials). Since the International Space Station is available for such experiments, a lot are mounted on board. Among other programs from the USA and Japan, the European Technology Exposure Facility (EuTEF) was mounted outside the station for carrying out experiments that required exposure to space environment for monitoring degradation dynamics in low Earth orbit and acquiring information in terms of contamination, atomic oxygen, ultraviolet and X-ray radiation, micrometeoroids, and debris (see Fig. 13). These international experiments, where the results are shared by the scientific community, increase knowledge and guide the development of new materials (like nanomaterial-based materials). If new materials are to be qualified in real conditions right from the outset, the qualification of large functional parts is not for the near future, and space equipment is still really only declared good for qualification when they start to work well. (a) The generation of contaminants has a strong effect on: – Modification of heat exchange by modifying thermal optical (emissivity) – Thermal and physical properties (influence on thermal coefficient of expansion which could affect thermal stability of the structure)

45

Aerospace Industry

1311

Fig. 13 European Technology Exposure Facility (EuTEF) mounted on the International Space Station (NASA images)

– Electrical perturbation of the electronic devices (dielectric discharge, perturbation of the measures, etc.) It is well known that polymers are sensitive to radiation, particles, and atomic oxygen generating micro-cracks, surface modification, and production of contaminants. The effects are limited to the surface with varying penetration depth (few microns for protons, several millimeters for electrons). Ceramic fibers are less sensitive, and composites made with carbon fiber are not much sensitive to ultraviolet radiation. For long-term exposure, composite structures are overwrapped with a protective film. Concerning outgassing which generates contaminants, resins less sensitive to moisture are selected such as the new generation of epoxy resins displaying moisture uptake less than 1% or even less, but for specific missions, cyanate resins with moisture uptake less than 0.2% are selected. The atomic oxygen test is only performed for materials which operate at low Earth orbits. For structural applications, the adhesive is usually protected by the adherends. The evolution of the adhesive properties with UV and particle radiations concerns those used for the optical devices or system glued on the surface for thermal control, for instance.

Tests on Technology Coupons The process validation of adhesive joints is carried out through demonstrators where the different bonded parts are manufactured to scale 1 and the bonding

1312

C. Désagulier et al.

implementation defined: surface preparation, adhesive conditioning, film implementation or multipart mixing, tooling assembly under pressure, curing parameters, temperature- and moisture-regulated conditions for room temperature adhesive (typically 22  3  C, 50  10% moisture content), autoclave curing cycle (120  C or 180  C in the case of co-bonding process), and inspections before and after bonding. All these methodological steps are to ensure an effective and homogeneous adhesion. The adherence verification of the adhesive joint is not limited to simple shear or peel tests, but to more complex tests which are representative of the real design, the real manufacturing processes and of course the real surface preparation are needed to reproduce load configurations. These tests are generally well instrumented in order to validate numerical simulations (see Fig. 14). Tensile and shear are generally performed for the inserts. Web, wedge, or cleat junctions are tested in tensile, shear, and bending. Then, comes the qualification of the equipment for delivery to the customer, bonding included, according to the dynamic, acoustic, thermal loads, under nitrogen and vacuum atmosphere, with safety margins.

Out-of-plane strength to failure test on elementary bonded cleats (ASL)

Reduced scale structure loaded in tension to validate the performance of adhesive skirt / vessel junction (ASL) Bending test on elementary adhesive bond (ASL) Fig. 14 Examples of technology test sample able to load adhesive joints in representative conditions

45

Aerospace Industry

1313

Damage tolerance justification is generally done at subscale level or even with coupons. For adhesive bonds, this consists in including the bondline, defects which are mainly representative of lack of adhesive and of process variation.

Part Qualification The qualification of structures is carried out via scale-one tests generally in static conditions at ambient temperature (in some cases with temperature) with a numerical justification for margin of safety (see Fig. 15). The qualification model is loaded up to rupture. In the framework of qualification with proto-flight model (approach quite often used for satellites and rather seldom for launchers), the limit load is applied, and the structure must be healthy without any damage at the end of the test campaign. One hundred percent of the structure is then controlled with nondestructive inspection. In Production Phase Structural adhesive joints are 100% controlled. This could be via nondestructive inspections or via proof tests. In some cases, such as adhesive bonds of thermal

Fig. 15 Illustration of some tests performed on scale-one structures for qualification

1314

C. Désagulier et al.

protection, the health of the bond is determined by the control of the parameters which monitor the manufacturing process. For satellites the acceptance of adhesive bonds is in fact done via series of tests performed on the scale-one part. The health of adhesive bonds is controlled after thermal cycling tests and dynamic or static tests. For some brittle parts like antenna reflectors, acoustic tests are also performed.

45.4

Surface Preparation

45.4.1 The Case of Composites When faced with a bonding case, there is a motto which should not be forgotten: “The Devil is in the details.” Among all the details, surface preparation is the decisive one. The aim of such a preparation is to remove anti-adhesive contaminants like grease or release agent residues which are in the air of the workshop (aliphatic hydrocarbon) and to promote a chemical–physical roughness deemed favorable to adhesion. Composite preparation is realized by sanding (silicon carbide paper, alumina oxide fabric, grade 240 up to 1,000) and wiping with solvents (acetone, ethanol, isopropanol), now often substituted by wiping with alternative liquids without volatile organic components according to the new REACH (Registration, Evaluation, Authorization and restriction of CHemicals, the new health and safety and environmental regulations enacted by the European Union) regulations. All these operations are performed with disposable talc-free gloves if carried out by hand or via an automated process in the case of extensive surface preparation. It should be stressed that surface preparation by hand depends on the skill of the operator. This is a major issue concerning bonding in the workshop. An alternative solution is to cure composites with a peel ply covering the surface to be glued. Peel ply is removed (tear-off) from the surface before gluing. This solution is very costeffective and efficient. The main drawback is the risk of leaving on the surface undesirable products (small pieces not detectable and pollution), since the operation is done manually. Among the reliable surface preparations which can be automated and which are taken into consideration to contribute to the improvement of quality assurance, without material drawbacks, are the low pressure plasma to break organic molecular bones, laser, and blasting methods. All these automated processes are from now progressively introduced in spacecraft equipment manufacture since they are growing in size. The automated inspection of the surface preparation also remains an issue. Several research initiatives have been done investigating several techniques. Among them, a recent program has been carried in the framework of European project called ENCOMB (extended NDT for composite bonding – FP7 – third call). The most representative contaminants met in the workshop have been studied, among them hydraulic fluid, release agent (pollution generated by release film),

45

Aerospace Industry

1315

and moisture (see Table 1). The most pertinent techniques proposed by the laboratories involved in the consortium were tested. Automated infrared spectroscopy probe surface analyses are a promising approach, but aerosol wetting technique (AWT) is more mature and applicable in the workshop. Up to now, the surface inspection (the adhesiveness through the surface energy displayed by the chemical–mechanical activation) is checked by the classic and simple so-called water break test or water drop test: after wetting with an aerosol on the surface of the composite which has just been prepared for adhesive application, low surface energy deemed unfavorable for bonding shall translate as separate droplets of water, whereas high surface energy meaning good surface preparation shall translate as a continuous film (the reader shall note that the words employed are qualitative and not quantitative, and verbs are to be understood in the conditional: because bonding is an art before being a science). This test is not so simple, as it requires more visual analysis performed by an inspector who can only rely on his memory (see Fig. 16). Currently, within the framework of a European research program, Airbus Company is conducting extensive industrial bonding studies. Among those, a robust and reliable surface monitoring method based on the manned model has been developed: after spraying water droplets at the range of about 0.1–0.3 mm, the wetting is characterized by the mean droplet size and their distribution. The image is recorded by a camera and analyzed by software. A hot air gun then dries the surface of the composite, if needed. THz/GHz reflectometry is very sensitive to moisture but this emergent technique is not yet used in space industry.

45.4.2 The Two-Part Epoxy Adhesive It is not necessary to put the importance of the surface preparation and its sensitivity under the spotlight, especially for composite surfaces. For economic reasons, manual sanding is still often the way for (1) removing dust and surface decontamination (aliphatic hydrocarbon, release agent residue contaminants) and (2) chemical Table 1 Main methods investigated in ENCOMB project to detect defects Method Laser-induced breakdown spectroscopy X-ray fluorescence spectroscopy Infrared spectroscopy Laser scanning vibrometry THz/GHz reflectometry Aerosol wetting technique (AWT) Optical stimulated electron emission

HF + + + + + + +

RA +

MO

+ + + + +

+ + + +

+ Possible detection HF hydraulic fluid/water contamination of CFRP, RA release agent (silicone-based) contamination on CFRP surfaces, MO moisture uptake in CFRP

1316

C. Désagulier et al.

Fig. 16 Wetting tests on a composite part (discontinuous droplets on (a) side, continuous water film on (b))

activation (electron peeling, micro roughness generation), and everybody knows how sensitive epoxy adhesives are to deviations in the surface composition. These operations will be entrusted to skilled workers. In the case of two-part room temperature adhesives, the adhesive is applied on the facing sides of the adherends after manual or automated mixing during the open time period of the mixture (45–90 min), and afterward, pressure tools are placed on the assembly for delivering the pressure required for bonding (between 200 and 1,000 hPa). An illustration is given of a bonding process after such a preparation in the case of the Ariane 5 satellite adaptor (see Fig. 17).

45.4.3 The Metallic Parts The metallic parts on spacecraft are mainly made of aluminum and titanium alloys because they are lightweight. Prior to structural bonding, they are subjected to proven surface standard treatments from the aeronautical sector. Due to the short open time period for bonding after the surface treatment (8–24 h), the metallic parts are often coated with organic primers to postpone their ability to bond and to procure or increase their resistance against corrosion if at risk (atmospheric, galvanic). The application of a primer raises the same issues as the adhesive, a primer being an adhesive coating, but in this case, two interfaces are to be held in place, the first one between the metal and the primer and the second one between the primer and the adhesive. In all the cases, because of the great variety of compositions of metal parts, combined with the extensive number of hardening and annealing treatments, there is

45

Aerospace Industry

1317

Fig. 17 Room temperature bonding sequence for composite assembly (from top to bottom, from left to right: two-part adhesive tube, mixing device, bond deposit on monolithic composite part, on sandwich part, docking of the two part, pressurization, assembly achieved ready for inspection (ASL)

not “a” standard treatment by alloy family but various possibilities, even if aluminum alloys have historical feedback from experience. However, preliminary tests should be conducted to determine the performance of the bonding in the case in hand.

45.4.4 Aluminum Alloys The surface preparation of aluminum before fitting, such as painting and bonding, is mandatory because aluminum alloys offer high sensitivity to superficial oxidation with the natural and instantaneous appearance of a brittle aluminum hydroxide coating.

1318

C. Désagulier et al.

This phenomenon which can be compared to what happened with a highly sensitive photographic film upon exposure to the light is more or less pronounced, depending on the ambient atmospheric conditions, i.e., the temperature and the moisture level. With respect to bonding, the challenge was in producing a compatible and suitable surface for the application of the adhesive prior to bonding operations. The idea was to chemically favor the growth of a stable and hard thin coating of aluminum hydroxide. As a result, two main different pretreatments are used, the first one in the USA based on phosphoric acid anodizing and the second one in Europe, consisting in a sulfochromic acid etching by a chromic acid anodizing (ASTM D-2651-01). For structural bonding, it has been demonstrated that the optimum values are achieved when the anodizing is partially sealed, despite a loss of corrosion protection. See also Part B – Surface treatments for more details on this subject. Alternate surface preparations are now introduced due to the new REACH regulations prohibiting implementation of listed chemical elements without enhanced health and safety protection and the release into the environment without significant recycling, in particular for the cancer-inducing hexavalent chromium. The chromium substitution was deployed a few years ago in the automotive industry, and the aeronautical industry can benefit from the studies performed following such substitution. Some alternate formulation packs are proposed mainly based on Cr3+ solutions, and they have been compared via extensive tests with regard to all the requirements for aluminum applications on aircraft. The space industry in turn benefits from these advances, where it appears that there is not only one alternate chemical solution but several, depending on the specific application. For structural bonding cases, the complicated ones are not necessarily the best. Green surface pretreatment qualifications are still in progress. These alternative Cr6 free solutions are in their final step of qualification and are progressively introduced in surface treatment supply chain. There are new axes of research in order to evaluate new treatment generation totally exempt of heavy metals.

45.4.5 Titanium Alloys The surface pretreatment for titanium is the well-known nitric hydrofluoric acid etching by a sodium bifluoric solution (ASTM D-2651-01). Some proprietary deviations from this basic treatment are taken by manufacturers. For example, hold release mechanism brackets in titanium alloy (Ti6Al4V) are used for mounting on antenna reflectors as parts of pyrotechnic mechanisms in order to deploy the reflectors in orbit before entering in operation.

45.5

Nondestructive Inspection

In space industry, high quality of the structures is required because no failure is allowed. The demonstration of the high level of reliability of the structures must be proved either by demonstrating that the process is well mastered (well-known

45

Aerospace Industry

1319

operation done by certified operators) or by the demonstration of the health of the structure. The second issue can be assessed by performing proof tests or by performing nondestructive controls. In fact, a combination of these three methods is carried out. If a large panel of techniques is now available based on standard physical principles (perturbation of electromagnetic wave motion, mechanical wave motion, heat wave motion) to control the health of metallic and composite structures, the determination of the health of adhesive bonds is still a challenge. The main sources of difficulties come from: – No direct access to the joint and the recorded signal can be disturbed by the heterogeneity and anisotropy of the adherends which are also different in nature (composite/metal). – Complexity of the shape of the piece. – Contactless method is sometimes needed to avoid any contamination. If a design to cost approach is followed by designers, design to control is also to be considered to avoid further trouble in nondestructive investigation. This could be illustrated with a simple example showing the detection of local debonding in titanium/CFRP composite joints with pulse-echo ultrasonic mode (see Fig. 18). Control via the titanium adherend is the worse method which could be imposed by the design (no access to the composite part). In some cases, when overlap of the reflected signals occurs, the diagnosis can be improved, thanks to the numerical simulation of the wave propagation. Simulation can improve the reliability of the diagnosis but also to replace the numerous and costly test samples being manufactured with defects. One can refer to CIVA software able to simulate control with X-ray technique, eddy current, and ultrasounds. Such tool is able to establish the probability of detection curves by simulation. The main nondestructive methods used in the aerospace industry are: – Ultrasounds – Infrared thermography Water CFRP composite Glue Z2=4,5 Z3=3 Z1=1,5

Titanium Z4=27

Water Z1=1,5

+0,58

+0,5 Good bond

+0,89 Good bond

-0,15 -0.26 +0,5 Decohesion of the first interface +0,5 Decohesion of the second interface

Titanium Z4=27

CFRP composite

Z2=4,5

+0,01 -0,15 -0,06

+0,89 Decohesion of the first interface

-0,75

Glue Z3=3

-0,19 Low signal

-0,15

-0,72

Reverse signal phase

Favorable situation of control disbond in an adhesive bond with pulse-echo ultrasound method –generation from the CFRP side

Decohesion of+0,89 the second interface

-0,15

-0,07

Not recommended situation of control disbond in an adhesive bond with pulse-echo ultrasound method –generation from the metal side

Fig. 18 Illustration of potential difficulty in defects detection in a hybrid adhesive joint

1320

C. Désagulier et al.

– Shearography (interference of coherent electromagnetic waves) – X-ray All these methods are able to detect debonding between the adhesive and the adherends or a lack of adhesive (see Fig. 19). Ultrasound techniques are able to do more with the detection of porosity and cracks or even a lack of curing of the adhesive. The defects that can be detected with the infrared response (IR) of the bonded equipment subjected to heating are lack of adhesive, bubbles, and discontinuities between the adhesive and the adherends (meaning, for instance, discontinuous adhesive distribution at the interfaces), air blade between skins to honeycomb core, or composite to composite or composite to metal bubbles in the joint. The numerous literature sources reviewed recently by Ehrhart et al. (2010) clearly show no nondestructive industrial methods are available to discriminate a “structural bond” from a “kissing bond,” even if efforts are continuously done at research level.

Fig. 19 Illustration of the principal nondestructive techniques used at ASL

45

Aerospace Industry

1321

This is the reason why adhesive bond technology is not used so much for highloaded structures in aeronautic because of the cost to demonstrate robustness and reliability of the adhesive joints. See also ▶ Chap. 42, “Nondestructive Testing” for more details on nondestructive testing. Two principal definitions of kissing bond are met in literature: (a) Two surfaces in contact but not bonded (b) Smooth bond consisting of two surfaces separated by a thin layer of liquid such as contamination Finally, kissing bond reflects the cases where the two parts are in intimate contact but with a weak adhesion force. According to the Ehrhart et al. (2010), ultrasonic alternative methods (US spectroscopy, nonlinear US, guided waves, oblique incidence, shear wave resonance, etc.) as well as laser proof tests (LASAT, LBI, laser ultrasonic) and other methods like shearography and active thermography have high potential for the measurements of adhesive bond strength, thanks to the new generation of sensors which are more sensitive (mK for IR thermography and few nanometer for shearography). The shearography technique consists of generating a small deformation on the surface of the structure and measuring the out-of-plane deformation (mechanical loading with low depression). This technique is similar to a proof test but performed at low stresses. This fulfilled technique is very relevant to control the health of the adhesion of the thermal protection. This solution is cost-effective, thanks to an automation control with robot. Recent researches in the detection kissing bonds have been initiated with ultrasound-guided wave pushed by aerospace industry needs. One of the potential solutions is to generate specific guided waves able to capture information provided by the interface/interphase which has in principle a different rigidity than the bulk adhesive. Castaings (2014) and his coworkers have deeply investigated this route, and significant improvements have been achieved, thanks to advanced computation tools which enable better understanding of wave propagation (see Fig. 20). Validation of the models has also been improved with experimental measures based on 3D laser velocimetry. The expected results are not possible without the development of appropriated sensors providing the generation of the desired guided wave at the right range of frequencies and the right level of power. The research has started with the detection of defects in single-lap shear test samples with no dispersive materials (aluminum plates). SH0 guided wave mode was generated and detected via Electro Magnetic Acoustic Transducers (EMAT) in 3-mm-thick aluminum plates with a 50 mm overlap in single-lap shear tests (see Fig. 21). Three samples were investigated, each with the aluminum/adhesive interfaces prepared differently (with or without sanded aluminum surfaces) applied, the third was polluted using an oily agent, and the last one the adhesive was a thin glass–epoxy self-supported film. Signals transmitted past the various lap joints revealed good signal-to-noise ratios, good uniformity in the

1322

C. Désagulier et al.

Fig. 20 Simulation of SH0 wave propagation in single-lap shear test samples: ABAQUS computation

Fig. 21 EMAT sensors in pulse and catch configuration over the bondline in single-lap shear samples (dot red line, different positions of EMAT sensors)

adhesive bonds, and high sensitivity of the investigated transmitted SH0 mode to the adhesion quality. This academic study has shown a good reproducibility of the signals measured for five positions of the sensors in the width. Thanks to EMAT sensors development, the signals are different with respect to the type of defects. But the shape and the amplitudes of the signals are not easy to analyze without the help of simulation where the main blocking point is still the quantification of the ultrasound properties of the interfaces (see Fig. 22).

1323

2

Spectral amplitude (u.a)

Aerospace Industry

Spectral amplitude (u.a)

45

(f)

Reference 1.5

1

0.5

0 0.15

0.2

0.25

0.3

0.35

2

1

0.5

1.5

1

0.5

0.2

0.25

0.3

0.35

Frequency (MHz)

0.4

Spectral amplitude (u.a)

Spectral amplitude (u.a)

(f)

With pollution on surface

0 0.15

0.2

0.25

0.3

0.35

0.4

Frequency (MHz)

Frequency (MHz) 2

(f)

1.5

0 0.15

0.4

Sanded surface

2

(f)

With rienforced adhesive 1.5

1

0.5

0 0.15

0.2

0.25

0.3

0.35

0.4

Frequency (MHz)

Fig. 22 Evolution of SH0 guided wave signals generated with EMAT 250 KHz sensors; reproducibility of five measures (Castaings 2014)

The same study has been done with quasi-isotropic composite materials as adherends. Even if the signals are repeatable, they cannot discriminate the defects with the same accuracy. Therefore, new development of sensors and more sophisticated data processing tool in relation with simulation are needed for composite materials. Other authors investigated nonlinear ultrasounds (Yan et al. 2009). It is assumed that when a large amplitude ultrasonic wave propagates through the kissing bond region, the system will behave nonlinearly, and so detection and quantification may be possible. They showed that an imperfectly bonded specimen exhibited much higher nonlinearity than that of a perfectly bonded specimen. These experiments are still performed with academic samples made with metallic adherends. Laser shock adhesion test (LASAT) is a new promising technique. Considered as proof test, the thermoelastic excitation produced by the multiple pulse laser induces an ultrasonic surface and bulk wave propagation in the material. The combination of the incident and reflected wave can create tension stresses inside the material enough to test interface/interphase adhesion and to detect weak bond (see Fig. 23). The main target is to find out the appropriated laser shock parameters which strongly depend on the adhesive bond definition to position the maximum traction stresses at the interface (Perton et al. 2010; Gay 2011).

1324

C. Désagulier et al. 4

3

Time in µs

2

1

0

8 plies Carbon Fiber Reinfoced Plastic / Adhesive / 8 plies Carbon Fiber Reinfoced Plastic Numerical simulation of wave propagation inside the junction generated by the shock (red in compression, blue in tension)

Decohesion Adhesive EA9394

Shock generating 0.46 GPa pressure No decohesion

Adhesive EA9394 Decohesion

Shock generating 0.62 GPa pressure Decohesion of the two interfaces

Fig. 23 Micrographic observations of adhesive joints of two eight plies laminate adherends after laser shock (Gay 2011)

45

Aerospace Industry

1325

All the difficulties hinge on mastering the stress field in terms of intensity and location. Too much energy provided by the laser can damage the surface of the material especially for composite materials. Therefore, the determination of the conditions which guarantees the relevance of the proof test of adhesive bonds must be done with a good understanding of the phenomena, the development of accurate numerical simulation validated by accurate measures, thanks to real-time diagnostics with laser Doppler velocimetry of the free surface and transverse shadowgraphy. Even if some devices are now commercialized (mobile laser bond inspection (LBI) system), LASAT it not yet used in production line. The diagram of wave propagation shows in this case the maximal tensile stress is located nearby the opposite site of the laser impact and the adhesive joint is less loaded in tension (see Fig. 23). In this case, the resistance of the interphase between adhesive and composite is lower than the interlayers and intra layers of the composite.

45.6

Strength, Failure, and Durability

Due to the fact that it is impossible to reproduce representative tests on Earth and that all is based on the reliability of the prediction, it is agreed in the space industry that the durability of the equipment shall be guaranteed by a policy of high safety margins: the margin given by the chemical behavior at high temperature (at the molecular scale) up to the mechanical properties at room temperature (at the macroscopic scale). It is also agreed that the equipment is qualified in simulation facilities according to conventional aging. This attitude is essentially due to the specificity of composite materials, where the stresses are endured first by the interfaces, at the resin-to-matrix/fiber interface, at the resin pre-impregnated inter-plies, and finally at the bonding interfaces. Companies are increasingly investing in automatic lay-up, heating-up, and line manufacturing devices to render the implementation of composites more reliable and robust despite the fact that the production rate of space industry does not justify by itself. To illustrate a commercial bonding case, two generic bonding cases are presented below: a sandwich construction with an adhesive film cured at high temperature and then the room temperature bonding of two composite parts.

45.6.1 Sandwich Structure Bonding Construction As a result, the compression resistance along the plane and the pull out-of-plane resistance, always as a function of temperature, are the drivers for sandwich structures. The selection of the adhesive film need to comply with all the thermomechanical needs at the elementary level, the outgassing space requirements, and

1326

C. Désagulier et al.

last but not least, the adhesive needs to satisfy the manufacturing process chart from the industrial point of view, technically (120  C class for typical sandwich curing assembly, 180  C class co-curing of the skins to honeycomb core, light areal weight supported or unsupported) and economically. A good film adhesive choice shall be a choice, where the compression load withstanding shall depend only on the composition of the skin (aluminum alloy, aramid, glass, or carbon high resistance or high modulus fibers), the thickness of the skin (in terms of space equipment, the count unit is 0.1 mm), and the honeycomb core definition (cell diameter often from 2 to 6 mm and density from 16 to 50 kg/m3 varying with the foil thickness of the honeycomb core). From a technological point of view, the compression resistance need shall be guaranteed by a minimum pull-out resistance, bearing in mind that there is no pure compression load but a combination of various angular loads supposed to contribute to the failure strength. To predict the behavior and prevent durability issues, as mentioned here above, the sandwich structure is fully characterized after numerous thermal cycles under inert gassing (to prevent nonrepresentative oxidation effects on Earth) such as nitrogen or helium (depending on the range of temperature (see Fig. 1). The expansion characteristics (coefficient of thermal expansion, coefficient of moisture expansion) and the thermal properties (through-the-thickness conductivity and the calorific capacity) are checked in addition to the mechanical properties. In most cases, a 1 MPa pull-out strength associated with a cohesive failure in the honeycomb or in the composite skins is often considered as acceptable for satellite sandwich parts. With such a result, a value of 300 MPa compressive strength associated with a Young’s modulus of 150 GPa delivered by the implementation of ultra-high modulus carbon fiber is displayed by satellite composite sandwich structural parts designed and experimented. Regarding the sandwich failure mode when subjected to a compressive load, the various failure modes have been set out in the ESA Handbook (ESA PSS 03 203). The failure modes generally observed are (1) facing failure which is the adequate mode rendering correct equilibrium of the spreading of the stress flux provided that the failure load complies with the predicted value, (2) intracellular buckling when the flexural behavior of the skins regarding the cell size induces such behavior according to a certain critical level of loading, and (3) outward face wrinkling when the secondary out-of-plane vector component of the main compressive strength is higher than the flat tensile resistance of the skins to honeycomb.

45.6.2 Monolithic Room Temperature Bonding In most cases, a two-part room temperature adhesive with a tensile lap shear of 20 MPa at room temperature and 10 MPa at 120  C shall be appropriate for bonding composite to composite and composite to metal on spacecraft. The choice of the adhesive shall be considered as appropriate when the failure is cohesive instead of adhesive, either in the bonded line, when the shear strength of the adhesive is lower than the shear strength of the adherends (metal or composite), or within the

45

Aerospace Industry

1327

composite thickness, when the inter laminar cohesion (the ability of the resin-to-fiber interfaces to withstand the loads) is lower than the adhesive cohesion. One specific attention is required when characterizing this type of adhesive. Due to the cold curing mechanism, the kinematic is slow, and time/temperature equivalence should be used with a lot of precautions in order to determine the appropriate mechanical performance of the assembly in real conditions.

45.6.3 A Commercial Bonding Case The case of applying bonding in real life is illustrated by the case of a radio-frequency antenna reflector. This reflector is dedicated to telecom satellite broadband service such as television, Internet, and wireless access networking services in the Ku band (12–18 GHz). This reflector belongs to the category of the high-throughput satellite (HTS), meaning, in order to designate any satellite (Ku band or Ka band, 20–40 GHz), it uses very small spot beams in order to obtain multiple orders of additional throughput compared to classic fixed satellite service. This can be achieved according to a multisource antenna principle or shaped antenna principle according to the multi-beam signal monitoring. Airbus Safran Launchers designs and manufactures both (see Fig. 24).

Fig. 24 Multisource antenna reflector 2.8  3.5 m (a, b) and multi-spot 1.6-m-diameter antenna reflector (c, d) – ASL

1328

C. Désagulier et al.

The backing structure is structurally bonded to the reflector shell of the Airbus Safran Launchers Ku-band reflector named ULR (for ultralight reflector): 2.2 m diameter CFRP sandwich shell and monolithic backing structure and titanium and Invar fittings for a total mass of 10 kg  1 kg/m2. This reflector concept has been mounted on telecom satellites since the beginning of the third millennium. The working environment is the following: (1) to withstand the dynamic and acoustic launching loads with a sine vibration environment of 10 g/out-of-plane axis and 15 g/in-plane axis, 7.90 g is reached on a shell equivalent to 1,000 N pulling out, a dynamic stress time of 1/2 h, and a few minutes acoustic stress time up to 150 dB, (2) to withstand the geostationary orbital thermal loads [ 50  C, +50  C] up to jettisoning of the launcher fairing, and (3) to withstand a thermal stress time of 15 years. First, the joints are tested at the elementary level and the loading cases predicted and margined according to out-of-plane, peeling, and cantilever directions after space-like aging thermal cycling. These tests will validate the design of the adherends and the industrial manufacturing process for the implementation of adhesives. Therefore, the mechanical tests demonstrate the ability of the bonded fitting to withstand the margined predicted loads. Then, it is necessary to validate the design at the margin loads. This is done through mechanical tests operated on representative artificially aged samples (preliminarily space aged by thermal cycling): static pull-out-of-plane test and cantilever for peeling tests. The industrial step consists in using optimized tools to reproduce the elementary bonding conditions which have been established to provide the required properties. The more optimized (reliable and faster) it is, the cheaper it will be – the ever-present obsession at all stages of industrial productions, space industry included. The cleats are bonded with an array of tools for ensuring a stress-free assembly after the surface preparation (sanding/ethanol wiping), the adhesive being applied in situ with a tube mixing device, with the dedicated static pressure application during hardening. With the cure time over, an ultrasonic inspection is carried out for bubble detection in the joint. Repair is required if the defect exceeds the specification (filling of the bubble or debonding and replacement of the cleat). Once the backing structure is bonded to the shell, the qualification tests are carried out to check that the equipment can withstand the launching loads and the thermomechanical behavior as predicted. First, the reflector undergoes thermal cycles within the extreme temperatures under vacuum to reveal possible weakness at the bonding interfaces, before shaking through the acoustic test performed in a chamber reproducing the launcher conditions. Finally, the reflector is shaken according to the dynamic spectrum of the launcher. At the qualification stage, the reflector being designed to work in all the launching cases, the dynamical and acoustical vibration spectra are maximized in order to cover the spectra of all the launchers existing (Ariane, Proton, Delta, Longue Marche). A final ultrasonic check performed after thermal cycling will sign off the qualification for delivery to the customer.

45

Aerospace Industry

1329

There is no room for error in space, no return ticket, and the possibility of repair is unusual; the Hubble telescope in low Earth orbit still remains an exception. In summary, the qualification tests performed on each flight model include thermal cycling and thermal stress distortion measurement, dynamic and acoustic tests representative of the launch margin load, and ultrasonic inspection after tests on each cleat junction for the structural bonding parts. Each reflector antenna, i.e., equipment, is considered as a prototype model where all the qualification tests are performed before launching.

45.7

Conclusions

45.7.1 By Way of Summary The assemblies are key issues for mastering the cost of space structures as well as performance. A non-careful mastering of the quality of the process and non-appropriate design can lead to a dramatic impact in late assembly phase or during the in-life service of the product. The challenge is to design and manufacture, while ensuring optimal thermomechanical returns over mass investment keeping in mind the minimization of the production cost. As surface bonding has been preferred to local fastening for spreading the loads with the result of optimizing the design to mass, the challenge then consists in the selection, for a joint where the resistance is governed like the cogs in a clock by the specific behavior of two adherends, the glue on one hand and the working environment on the other. The design office associated with the material and processes experts deals everyday with these inputs for succeeding the correlation between the predicted margined level and the real level withstood, sign of an optimal combination and trust in structural bonding.

Launchers Upper-stage structures are made in composite materials and the load transfer between stages is still done via bolted solutions. Few parts are designed with adhesive bonds. The adhesive bond solution is only proposed when mass optimization is the main driver. Apart from sandwich manufacturing which is now obtained via co-curing process, the main drawbacks of the adhesive bond solution are linked to the low confidence of the designer regarding kissing bond detection not yet solved and the qualification cost. The main applications of adhesive bonds are met in thermal protection of structural parts, thermal protection for cryogenic vessel, thermal protection of the stages, and thermal protection of the solid rocket motor. The success of the adhesive solution is based on stringent manufacturing process requests to perform a lot of controls of raw materials and process parameters and last but not least to have skill and qualified operators. Rigorous procedures established in close collaboration between designers, material and process specialists, and manufacturing specialists are the key to successfully mastering adhesive bonds.

1330

C. Désagulier et al.

The major difficulty encountered by material specialist concerns adhesion with elastomers where the mechanisms of adhesion with metals and composites are complex to master.

Satellites Adhesive bonds in satellite design are the mandatory technology for fastening parts enabling the optimization of the structures. Each gram of structural material saved is a gram of propellant added and, consequently, satellite lifetime extended (remember that for satellites “the lighter it is, the more profitable it is”). The selection of adhesives is based on the mission of the satellite governed by the constraints of space environment (orbit around planet and deep space trip). If moisture ingress (outgassing constraints) and thermal cycling are the key parameters, radiation resistance is required for limited applications. In principle, adhesive joints are not directly exposed to the aggressive space environment.

45.7.2 And Finally “It is not the glue which makes the gluing” (Max Ernst) but preparation, inspection, and deep control of the process. The adhesive shall determine the maximum loading allowable for the bonding assembly, but structural bonding shall occur when the adherend surfaces have been suitably adapted after the dedicated preparations to provide conditions of interphase formations (this slight mixed layer on an atomistic scale between the adherends and the adhesive). The cohesive failure mode is the only mode which provides reliable prediction since one knows the separate properties of the adherends and the adhesive on the whole. If somebody had to summarize the particularities of bonding for space applications, he would formulate the following equation: space bonding = high mechanical loads (launcher and satellite) over a short time + high thermal loads (satellite only) over a long time, and it is the challenge with which one is faced. The feedback from experiments in bonding is crucial because, apart from everything one does know, there is everything one never knows, which is huge. Sometimes, it hinders technological breakthroughs having to balance the cost with risk reduction up to almost zero, with the financial issues, the reduction of manufacturing costs, and added payload mass.

References Castaings M (2014) SH ultrasonic guided waves for the evaluation of interfacial adhesion. Ultrasonics 54:1760–1775 Collier C, Ainsworth J, Yarrington P, Lucking R (2010) Ares V interstage composite panel concept and ring frame spacing trade studies. 51st AIAA/ASME/ASCE/AHS/ASC structures, structural dynamics, and materials conference, Orlando Drucker DC, Prager W (1952) Soil mechanics and plastic analysis or limit design. Q Appl Math 10 (2):157–165

45

Aerospace Industry

1331

ECSS-Q-70-71A rev 1, Space product assurance. Data for selection of space materials and processes Ehrhart B, Valeske B, Muller C-E, Bockenheimer C (2010) Methods for the quality assessment of adhesive bonded CFRP structures –a Resumé. 2nd international symposium on NDT in aerospace ESA PSS 03 203 Structural materials handbook Gay E (2011) Comportement de composites sous choc induit par laser: Développement de l’essai d’adhérence par choc des assemblages de composites collés, PhD thesis Perton M, Blouin A, Monchalin J-P (2010) Adhesive bond testing of carbon – epoxy composites by laser shockwave. J Phys D Appl Phys 44(3):034012 Popineau S (2015) Durabilité en Milieu Humide d’Assemblages Structuraux collés type Aluminium/Composite. PhD thesis Prager W (1995) The theory of plasticity: a survey of recent achievement. Proceedings of the institution for mechanical engineers Raghava R, Caddel RM, Yeh GS (1973) The macroscopic yield behavior of polymers. J Mat Sci 8:225–232 Yan D, Drinkwater BW, Neild SA (2009) Measurements of the ultrasonic nonlinearity ok kissing bond in adhesive joints. NDT E Int 42:459–466

Automotive Industry

46

Klaus Dilger, Bernd Burchardt, and Michael Frauenhofer

Contents 46.1 46.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . General . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.2.1 Press Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.2.2 Body Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.2.3 Paint Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.2.4 Trim Assembly . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.2.5 Power Train . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.3 Specific Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.3.1 Adhesively Bonded Sheets in the Press Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.3.2 Adhesives in the Body Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.3.3 Adhesives in the Paint Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.3.4 Adhesives in the Assembly Line/Trim Shop . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.3.5 Adhesives in the Power Train . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.4 Other Specific Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.5 Surface Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.6 Strength and Durability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.7 Durability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.8 Type of Failure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.9 Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.10 Repair and Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1334 1334 1336 1336 1336 1336 1337 1337 1337 1338 1344 1344 1346 1347 1348 1349 1354 1363 1364 1365 1365 1366

K. Dilger (*) Institute of Welding and Joining, TU Braunschweig, Braunschweig, Germany e-mail: [email protected] B. Burchardt Weiningen Zürich, Switzerland e-mail: [email protected] M. Frauenhofer Audi AG, Ingolstadt, Germany e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_46

1333

1334

K. Dilger et al.

Abstract

Modern cars contain about 15 kg of adhesive (IVK 2017). Adhesives are used to avoid corrosion, to stiffen the car body, to obtain a better crash performance, to join different materials that are not weldable, and for sealing purposes. Adhesives are used in powertrain and car body applications. In the car body, up to 200 m of structural adhesives are applied. The trend toward a new lightweight design leads to the necessity to join different materials in a material mix. Here, the use of high strength steels, light metals, and fiber-reinforced plastics has expanded extremely in the last years. In terms of adhesive bond, one has to deal with different surfaces and thermal expansion. Additionally, the adhesive has to act as an electrical insulation to prevent the joint from forming contact corrosion. In this chapter, typical applications for automobiles are shown. Additionally, suitable adhesive properties and surface treatments are discussed.

46.1

Introduction

Adhesive bonding of automobiles started with sealing applications like the substitution of the windscreen gasket by a bonding seam of a rubber elastic adhesive or gap filling with an adhesive to prevent corrosion. The sealing adhesives led to a stiffening of the car body because the bonded windows could act as a shear field and the adhesively bonded joints were stiffer than spot welded flanges due to the complete connection. To further improve the stiffness, an increasing number of adhesives with a high modulus was used. Since these adhesives were not suitable to improve the behavior of the joint in a crash situation, they could not be used as a real structural element, which would enable to reduce the number of spot welds. The third generation of adhesives with enhanced crash properties finally offers the opportunity to use adhesive bonding as a real structural joining alternative in an automobile construction. We can generally distinguish between applications in the automobile body and the power train. Regarding bonding in the car body, you have to take into account that the body has to be painted, meaning that the adhesives have to be applied before or after certain painting processes. This is of high importance in terms of the selection of the suitable adhesive. In Fig. 1a, typical bondings on raw metal in the car body are shown. Fig. 1b gives examples of bondings on painted panels. Regarding the strength of the adhesive bonds, one can distinguish between high strength and high stiff bonds and bonds with lower strength and a higher flexibility. To produce high strength joints, often epoxy adhesives are used with a low bondline thickness of about 0.2 mm. Windows are bonded with a highly elastic polyurethane adhesive with a bondline thickness of more than 2 mm (Fig. 2).

46.2

General

The process chain in an automobile production can be separated in the production of the car body and the components, respectively the powertrain. The stages of the body production are the press shop, the body shop, and the paint shop. The final

46

Automotive Industry

Fig. 1 (a,b) Typical bondings in a modern automobile

Fig. 2 Typical semistructural and structural bondlines

1335

1336

K. Dilger et al.

production step to finish the automobile is the trim assembly. While in the press shop besides of a possible use of bonded blanks no adhesives are used, adhesives are applied in the three other areas mentioned and have to undergo the following process steps.

46.2.1 Press Shop Classically, sheets made of steel – uncoated or zinc coated – and aluminum alloys are cut and cold formed in a deep drawing process. To prevent the metals from corrosion and to facilitate the pressing process, the blanks have an oily surface. A modern approach is the use of a warm forming process where the deformation and the hardening of the steel (e.g., 22MnB5) is performed at the same time. Therefore, the sheets have to be heated to a temperature of about 850  C, which could cause the formation of a thick oxide layer. To prevent the parts from this oxidation, a special coating is used, which has to be taken into account for the subsequent joining procedure.

46.2.2 Body Shop The formed parts are brought together to assemblies in the body shop. Here, almost every joining technique is used. The standard method used for joining of steel is spot welding, whereas for light metals mechanical joining like clinching and riveting is commonly used. Independent of the material used and of the joining technique, the application of adhesives is necessary or beneficial to prevent the parts from corrosion. Adhesive bonding is used in steel and aluminum bodies to stiffen the parts and to improve the crash behavior. In a mix material design, the use of adhesives is essential to avoid the contact corrosion that could result from the contact of materials that have a different electrochemical potential. The adhesives applied in the body shop will be cured in the oven, which is needed to cure the electro-coated paint. Therefore, during all of the processes performed in the body shop, the adhesive is liquid and the handling strength has to be obtained by an additional joining technology.

46.2.3 Paint Shop Before the paint is applied, different pretreatments like cleaning and phosphating are executed. The paint is applied in different layers. It starts with the application of the electrolytic cataphoretic paint and the curing of that layer during about 30 min at a temperature of 180  C. Afterward, more paint layers are applied and cured at temperatures from 110  C to 150  C. The resulting layer structure is shown in Fig. 3.

46.2.4 Trim Assembly In the trim assembly, the windows and different panels and other optical elements are bonded to the painted body. Thus, the paint layer becomes part of the structure, which

46

Automotive Industry

1337

Fig. 3 Different layers of paint due to the current painting process of a car body

has to be taken into account if there is a demand for a higher strength. Additionally, the chemical base of the paint, the color, and the painting process are of a high importance in terms of the resulting properties. Because of the heat sensibility of the applied materials and the assembly process, only cold curing adhesives can be applied.

46.2.5 Power Train In the power train, adhesives are used for shaft-to-hub connections, for sealing purposes and as a thread lock. Special demands for the adhesives used result from the high temperatures and the contact with fuels and lubricants.

46.3

Specific Requirements

46.3.1 Adhesively Bonded Sheets in the Press Shop To reduce weight, metal sheets with a varying thickness are used for forming parts. In this case, the wrought products are made of sheets with different thicknesses by welding or differential rolling. Hence, to produce the so-called tailored blanks, different sheets are welded together in general by laser beam welding before they are formed in the pressing process. This is very common for linear and nonlinear joints. A new possibility to improve stiffness and corrosion resistance is the use of so-called bonded blanks where two sheets are bonded together prior to the pressing process. In this context, it is important that the adhesive has a sufficient green strength to ensure the right position of the blanks during the pressing process, but nevertheless there is enough plasticity to avoid the formation of cracks in the adhesive layer during the deformation process. One possibility to achieve this behavior is the use of an adhesive with two hardening procedures. The green strength results from a viscosity increase due to the cooling of the hot applied adhesive, and the final strength is formed by crosslinking due to the chemical reaction after the

1338

K. Dilger et al.

pressing process that takes place in a furnace at 180  C (Wisner et al. 2014). An example of a part made from bonded blanks is shown in Fig. 4.

46.3.2 Adhesives in the Body Shop Adhesives are typically used in the body shop for structural flange bonding spot weld bonding, clinch bonding, etc. in a so-called hybrid process. Here, the adhesive is applied on one sheet and after joining of the parts the additional joining technique takes place. Further applications are hem flange bonding and antiflutter bonding. Adhesives used in the car body structure have very specific and demanding requirements: 1. They need adhesion on oiled bare, electrogalvanized, or hot-dip galvanized steel and on aluminum metal sheets. 2. They undergo the complete pretreatment process for the electro coat including the heat curing in the EC oven. The adhesives are designed to cure under these conditions. 3. A temperature resistance of 200  C over half an hour is needed, but due to the variety of production processes between the different manufacturers and even between different plants up to 220  C is specified, or at the lower end under low bake conditions 160  C for 15 min. Therefore, the curing properties of adhesives have to be adapted to the specific process conditions of the individual production line. 4. In certain cases, parts with uncured adhesive are stored or shipped. Under these conditions adhesives may absorb moisture from the atmosphere and when curing at 180  C the absorbed moisture will cause bubbles, thus reducing the strength of the adhesive.

Fig. 4 Bonded blank technology (Wisner et al. 2015)

46

Automotive Industry

1339

5. Adhesives must be applied in a liquid stage to achieve a proper wetting, and thus they usually do not withstand the conditions of the pretreatment baths. Therefore, a precuring or pregelation is often required. Due to separate handling, this is easier for the bolt-on parts like doors and hoods since adhesives used for them are often cured in a separate prebaking oven or by induction curing. The newest generation of structural adhesives does not need such a precuring any more. 6. Since all body shop adhesives are applied directly on the metal surface, corrosion resistance is crucial. And it is even more demanding than the corrosion resistance of coatings, since there is no specific pretreatment, and the structural adhesives replacing welding operations must withstand much higher loads than a coating. Regarding the corrosion tests, every car company has its own test requirements, because they have individual longtime records with their specific test conditions. But in general, they have the following requirements in common: the combination of salt spray, high humidity, and temperature change. These conditions give usually similar results depending upon the time and the number of cycles.

Structural Flange Bonding (Hybrid joining) Typical configurations of the use of different joining technologies at one flange in a hybrid joining process are shown in Fig. 5. To build up the structure of the car body, welding is the standard technology to join metals and especially spot welding is used for all steel cars. In the last 30 years, the requirements for car design have changed due to the pressure of lower fuel consumption and higher safety for the occupants, which revealed the limits of the spot welding technology in terms of durability, crash resistance, and production speed in connection with thinner high strength metal sheets or in a more obvious case, with aluminum substrates. To overcome the drawbacks of spot welding, other joining technologies have been evaluated including riveting, self-piercing riveting, clinching, and bonding. The structural adhesives marketed for the automotive industry at the beginning of 1990 were rather brittle and not really resistant under high-impact conditions. Consequently, the adhesives industry developed the new generation of highperformance structural adhesives in this decade. These combine the strength of standard epoxy adhesives with higher deformability before break. These are often formulated as warm-melt adhesives

Fig. 5 Hybrid joining of a flange in a car body

1340

K. Dilger et al.

and applied at temperatures up to 60–80  C, and after cooling, they get highly viscous enough to withstand the pretreatment operations. But these warm-melt properties have to be well balanced, because if the viscosity of the adhesive is too high after cooling, spot welding will lead to pocket formation as demonstrated in Fig. 6. The welding will need more electrical and pressure power since a thicker layer of adhesive acts as an insulator. Usually, this is addressed as squeezability, which means the property of the adhesive to flow under a small pressure when the metal sheets are put together. The challenge for the adhesive formulator is to balance a low viscosity at the application temperature of 80  C with a rather high viscosity at room temperature to achieve a good washout resistance and at the same time a good squeezability (Stepanski 2010). Another critical point is linked with the reactivity of the adhesive. It needs heated pumps and tubes for the application at temperatures between 50  C and 80  C. But due to the inherent reactivity of such one-component adhesives, there will be a slow increase of its viscosity over time. The flow in the tubes will then be reduced and when increasing the temperature to keep the constant flow, the viscosity increase continues and the tubes will be blocked. This is important in cases where the consumption of the adhesive is rather low or the tubes are long. Therefore, not only the design of the adhesive is important, but also when planning the application equipment, it should be clear that the adhesive in the tubes should regularly not stay for a longer period at elevated temperature. Since this is specific to the adhesive and the equipment of the production line, it has to be evaluated in detail by the responsible engineers for every single application. For all adhesives application, it is recommended to use heated pipes since this guarantees a rather constant application viscosity. The need to reduce the weight of a car leads to the use of alternative materials that have to be joined with a similar material or a different material in a so-called hybrid construction. Here, traditional joining and welding processes are often not suitable due to different reasons. One is the difference in between the thermal expansion coefficients of the materials, another is the electric potential of the material. Adhesives are suitable for bonding of almost any material and reduce the mentioned problems, thus increasing the use of adhesives in automobiles significantly in the last 10 years. In Fig. 7, the material mix in a modern car structure is shown.

Fig. 6 Pocket formation when welding through a highly viscous adhesive

46

Automotive Industry

1341

12% 23% 3% 47%

steel (cold forming)

steel (hot forming)

Al-sheets

Al-castings

15%

Alu-extrusions

Fig. 7 Material distribution in the Audi Q7 (Audi)

Here it is necessary to compensate the different elongations of the parts due to the different thermal expansion (CTE-Mismatch) by a thick and elastic adhesive layer. However, such an adhesive layer can reduce the stiffness of the part significantly so that adhesives with a middle modulus of 10–100 MPa are developed.

Hem Flange Bonding Bonding of hem flanges has now been practiced for more than 30 years. At the beginning, the reason was more corrosion protection than the increase in stiffness. Meanwhile, the hem flange bonding contributes to the structural stiffness of hoods. And it is an important element for the resistance of the doors against side impact. The adhesive prevents the inner door panel with the reinforcement bar to slip out of the hem flange upon impact. Regarding application, the hem flange bonding needs very specific processes. To fulfill the strength and corrosion requirements, the flange should preferably be filled completely. But in this case, the hemming tools and also the parts will soon become soiled, which would then require an additional manual cleaning step. Therefore, the application is carried out on the inner side of the outer panel to avoid the adhesive to be squeezed out. There is another difficulty. The high pressure of the hemming tools press the gap to almost zero, but the metal bounces back due to its elastic properties thus creating voids. Some adhesives use small glass bubbles as a filler in order to keep a gap of 0.3 mm, which reduces the number of voids. In addition, the doors are seam sealed to achieve a proper corrosion protection. But when the air is entrapped in the hemming area, bubbles will be formed in the hem flange sealant on curing, while reducing the visual quality of the sealing bead. And if not sealed, the hem flange has a high risk for corrosion because of these voids. This problem has not been solved yet to the satisfaction of the engineers. A possible solution is the application of the seam sealer onto the hot hem flange when using induction heating, but then the cycle time will get too long, which means additional expensive tools. The different steps of hem flange bonding are shown in Fig. 8. Antiflutter Bonding A third important application is antiflutter bonding. To avoid rattling noises, the gap between the outer panel and the inside structural metal sheet of a hood must be filled

1342

Step 1: Adhesive application

K. Dilger et al.

Adhesive Outer sheet

Inner sheet Step 2: Joining

Step 3: Hemming

Outer sheet

After hemming ideal situation

Void Step 4: Before sealing

After hemming real situation

Hem flange sealant

Step 5: After sealing

Entrapped air

Bubble Step 6: After curing

Expanded entrapped air

Fig. 8 Main process steps of hem flange bonding

with a suitable material. This often is soft rubber foam to avoid any surface marking, but in this case no increase in stiffness is possible. With the push toward lightweight design, the gauge of the outer metal panels has become thinner and thinner. And when using a “semi-structural” adhesive, the stability of such a hood can be increased. This is also a structural adhesive because it contributes to the stiffness of the hood, but due to the requirements regarding surface quality with no marking on the outside panel, a balance between strength, modulus, and elongation must be found. If the adhesive becomes too strong, there will be an unacceptable surface marking (read-through) on top of the bonded area, and if the E-modulus of the adhesive is too low, there is no stiffening effect. The best solution would be a high modulus adhesive with no read-through. Unfortunately this “read-through” is caused only to 10% by the shrinkage due to the chemical reaction of the adhesive itself, the other 90% is due to the physical shrinking from 180  C to room temperature.

46

Automotive Industry

1343

The surface marking (Fig. 9) can go in both directions, but due to the thermal shrinkage of the adhesive, in most cases it is indention. In the phase of the heating-up in the oven, there are differences of up to 60  C between the inner and outer metal sheet and when the adhesive cures at this stage, surface marking after cooling can be observed. Tests with a two-component adhesive curing at room temperature showed that with cold curing the marking was significantly reduced. Unfortunately, it is very difficult to get adhesion on oily metal sheets using two-component adhesives. Figure 10 shows typical applications at a front hood. There is a hem flange area either around the whole perimeter or only in the front and the back, and on the side of a spot-welded area but with the same structural hem flange adhesive. The reinforcement bars of the inner panel are bonded with thick adhesive layers (usually around 3–5 mm thickness). In areas where higher pressure onto the outside panel is expected when closing the front hood by hand, a larger area is supported. To apply the adhesive on such a larger area, the multidot application is preferred. This has the Fig. 9 Surface marking by antiflutter bonding

Outer panel

Adhesive bead

Inner panel

Cross section

Multi dot adhesive application

Fig. 10 Antiflutter bonding applications

Hem flange area

1344

K. Dilger et al.

advantage that there are gaps between the dots, and the electrocoat can flow into these gaps and provide the necessary corrosion protection. Adhesives used in a car have to work at temperatures between 40  C and 105  C. More demanding than the temperature is the resistance toward humidity because hot and humid conditions can reduce the strength of adhesives significantly depending on formulation and chemistry. There is also a differentiation between inside and outside parts. Outside parts have not only to pass high temperature and high humidity tests but also the corrosion test cycles like the VDA 621-415. Inside parts are exposed to higher temperatures (in some areas up to 120  C) than outside parts but are not exposed to a corrosive environment (Table 1).

46.3.3 Adhesives in the Paint Shop In the paint shop, there are only some minor applications of antiflutter bonding and seam sealing with materials which need similar properties like adhesives in terms of good weathering resistance and good adhesion. These sealants are polyvinychloride plastisols in most of the cases, as they are used for underbody coating and do not provide high strength. In this case, the adhesion occurs on a coated surface, which offers more opportunities in the formulation and no issues regarding corrosion. Since the paints are cured in the paint ovens, their heat can also be used to cure the sealants. Temperatures are typically between 140  C and 160  C with the tendency to baking at lower temperatures. Since the paint process follows the body shop operations, all body shop adhesives have to withstand the baking conditions of the paint. There are efforts to move all non-paint operations out of the paint shop, thus bonding in the paint shop will be the exception.

46.3.4 Adhesives in the Assembly Line/Trim Shop The largest variety of adhesives is used for all the parts assembled in the trim shop. Since many parts are produced at suppliers or subsuppliers and are installed in the car either by screws, bolts, or clips, there are only a few applications performed directly at the assembly line. This is also due to the fact that adhesives need a specific process for curing or hardening, which in most cases do not fit into the required short Table 1 Test conditions of VDA 621-415 corrosion test cycle (Available from VDA website www. VDA.de/en/publikationen) One cycle Day 1 2–5 6–7

Duration 24 h 4  24 h 2  24 h

Condition Salt spray DIN 50021 SS DIN 50017 KFW 23  C, 50% relative humidity

This cycle of 7 days is run ten times. The overall duration is 10 weeks

46

Automotive Industry

1345

cycle time of often less than 1 min. Most bonded parts are preassembled and built into the car as finished parts. Adhesive bonding of these separately built-on parts is not within the scope of this chapter. The most important structural bonding in the final assembly is the installation of the windows. Even if the adhesive for this purpose must be flexible enough to prevent a breakage of the glasses due to high stresses through deformation, the resulting increase in torsion stiffness (a high quality property of a car) depends on the car design in the range of 50–80% with almost no additional cost and no increase in weight. Therefore, direct glazing has replaced the rubber gaskets in almost all car production lines and has become an important element for the torsion stiffness of the car. This installation principle of direct glazing can also be transferred to the installation of other panels like a separate lightweight roof made of aluminum or carbon fiber composite. The most important application at the assembly line is the bonding of the glasses. Figure 11 shows a typical cross section of a frame with the adhesive bead. Usually, a triangular bead with a height of 8–10 mm and a width at the base of 10 mm is applied onto the glass which is then installed into the car. This is necessary to cope with the tolerances of the frame and the glass. The use of 1-C polyurethane adhesives is state of the art. Standard grades are applied at room temperature and need fixation for at least 1 h to get sufficient strength to hold the window in place. Warm-melt windshield adhesives are applied at temperatures between 60  C and 80  C and on cooling they build up sufficient initial strength to hold the windows in place during curing with no additional fixations. The adhesives are fully cured after 3–7 days depending on the climate conditions. The warm-melt properties must be balanced very thoroughly. If the adhesive is setting too fast, it does not wet the surface enough and there will be disbonding under torsion stress. To increase the slow curing speed in cold and dry atmospheres, two component systems are available, which are either real 2C systems or 1C adhesives accelerated with water paste. These are used when the bonded structure has to bear load at a very early stage. An important detail is the conductivity of the windshield adhesives. If such adhesives were formulated with higher amounts of carbon black as filler, the direct contact of such an adhesive with aluminum would lead to severe and fast contact corrosion and would destroy the aluminum in the case that the isolation through the paint is damaged. Also, if a radio antenna is integrated into the glass (e.g., a rear window), the impedance of the adhesive will play an important role for the quality of radio reception.

Fig. 11 Cross-section of windshield bonding

Glass Outer shell

Ceramic coating

Inner frame

1346

K. Dilger et al.

When windshield adhesives are directly exposed to sunlight, the thin adhesion layer directly on the glass will be destroyed by UV-radiation. Therefore, a UV protection is a must. In the beginning of windshield bonding, this was achieved by a black primer, but under severe UV-radiation, even this rather stable UV-blocking primer was destroyed over time. With the application of a screen-printed ceramic frit (see Fig. 11) to the bonding area of the glass, the necessary high-quality UV protection could be achieved. More than 30 years of success in windshield bonding has proven the reliability and durability of this technology. The replacement of windshields must be possible, too. At the beginning, this was an important criterion against the introduction of direct glazing. But with specifically designed cutting knives or with high strength steel wires, the removal for repair purposes was possible in a short time. To facilitate the repair, a minimum thickness of the adhesive layer must be guaranteed, preferably more than 3 mm. The success of windshield bonding encouraged the designers to bond other panels onto the frame. In particular, the roof is in study, because saving weight on top with an aluminum or carbon fiber composite roof, will lower the point of gravity and improve the road handling. An overview on elastic bonding can be found in the booklet of Burchardt et al. (1998) and Cognard (2006). For the fixation of model names and door trims as well as rubber gaskets, pressure sensitive adhesive tapes are commonly used. Such pressure sensitive adhesives (PSA) are highly sensitive to stress. The parts should be attached in a completely stress-free state and with a defined pressure. After the application, the adhesive continues to wet the surface and to build up the final adhesion. Brockmann (Brockmann and Grüner 2003) investigated the creep behavior of such tapes and he found out that under load the tape may creep over its whole overlap length and is still bonding with the same shear and peal strength as in the unloaded stage. Besides the areas described above, adhesives are used for different purposes in the engine and the gearbox. These applications are described below.

46.3.5 Adhesives in the Power Train In the power train, adhesives are used for shaft-to-hub connections, for sealing purposes, and as a thread lock. For shaft-to-hub connections mainly anaerobic acrylats are used. In most cases, the adhesive bond is supported by a shrinking of the hub to the shaft. By this hybrid joining process, both a high axial strength and a high torque can be achieved. A secondary effect of the shrinking is residual stresses in the parts, which can lead to a reduction of the fatigue strength. Especially for rigid materials like sinter metal this can cause problems; therefore that the shrinkage should be reduced and the adhesive part of the strength should be increased [17, 18] Lit.-Stellen ergänzen. To produce these bonded shrink fits, the hub should not be heated up to more than 200  C to avoid damage of the adhesive and the shaft should not be cooled to avoid condensed water.

46

Automotive Industry

1347

Figure 12 gives an overview about the shrink fit process to produce a gear-shaft connection. Thread locks are made by either anaerobic or microencapsulated adhesives. These adhesives have a defined initial breakaway torque. Silicones are often used for As gaskets formed at a place in the power train. To enhance the curing time, there are UV-curing products on the market. It is important that – because of the bad paint adhesion on silicone-contaminated surfaces – silicones are only used for power train applications and it has to be ensured that no contamination of body parts is possible.

46.4

Other Specific Requirements

Adhesives used in a car body need a temperature resistance of at least 105  C and must resist high humidity conditions as well as temperature change between 40  C and +80  C. Such adhesives are often tested under the so-called cataplasma condition, that is, 70  C in a water saturated atmosphere. Each car manufacturer has its own test regime but they rely on the combination of high humidity and temperature. In addition, the UV resistance of the system must be checked either with exposure to sunlight in areas like Florida or in test chambers with UV-lamps in combination with heat and/or moisture. The interpretation of such artificial weathering needs experience, since the extrapolation for the real-life weathering cannot be done automatically. On top of technical requirements additional properties are requested. One is low fogging of small ingredients or plasticizers in the adhesive and another is low odor, since customers became more sensitive toward the smell of adhesives. For outside application, fuel resistance, resistance to cleaning agents, and washing fluid for the windshield and headlights are required in addition to water and corrosion resistance. Fig. 12 Bonded shrink fit

1348

46.5

K. Dilger et al.

Surface Preparation

The right surface preparation is the key to successful adhesive bonding. Therefore, it is at first sight surprising that bonding in the body shop with the highest strength requirements occurs on oily surfaces. But to be sure, these systems have been tested extremely thoroughly and only those combinations of adhesives, oils, and metal surfaces are allowed to use which have passed all test. And then, it is up to the quality control system over the whole value chain from the suppliers of adhesives, oils, and metals to the stamping plant to make sure that all process steps are well under control. And all changes have to be communicated before implementation. The fact that body shop adhesives are cured at temperatures about 180  C helps the adhesive to absorb the oil and get in direct contact to the surface with good wetting. The weakest area of the surface is the thin oxide layer which is on all metal surfaces. The phosphatation process which creates a good adhesion layer for the paint cannot be used since the adhesives are applied at an earlier stage during the assembly of the metal parts. In the future, precoated metal sheets could probably be used, but then the spot weld bonding would be no more possible. In addition, one has to keep in mind that the adhesive strength on such coatings is lower than that of high strength structural adhesives. Even the adhesion quality of a hot dip galvanized steel coating has to be tested since with some structural adhesive the zinc layer debonds from the metal surface. In case of aluminum, a pretreatment is necessary to achieve a long-term-corrosion-resistant adhesion. Whereas in the aerospace industry, anodizing processes are state of the art, automotive industry has been looking for more cost-effective solutions. A Cr (VI) passivation provided the best results and is now used as benchmark for new chrome-free pretreatments necessary due to the environmental impact of chrome. The best and most cost-efficient results are reached with a passivation of the aluminum in a coil coating process. But this implies that the lubricants needed for the drawing of the aluminum sheets is compatible with the adhesive. Otherwise a separate washing process of the stamped parts would be necessary which is not really economic. A new article from Rechner et al. (2010) discusses different surface pretreatments. In addition to solvent cleaning, new methods with laser, CO2, and plasma treatment are available. A short overview has been written by Schulz (2010) (See also Part B). As the oven for the EC coat can be used for curing, almost all body shop adhesives are one component systems. The heat curing helps to absorb the oil thus providing the necessary adhesion. In contrast to such one component systems, it is very difficult to get adhesion on oily surfaces with 2-C adhesives. Due to the immediate start of the curing after mixing of the two components, the time to replace the oil on the surface is too short in comparison with the curing speed, because the monomer or prepolymer components of the adhesive resin and hardener react faster with each other than they are able to wet the surface. Only with very long open time adhesives there is some chance to get adhesion. Significant efforts are made to use 2-C adhesives for

46

Automotive Industry

1349

bonding of hoods to reduce the thickness of the outer sheet without surface marking. Also a challenge is the adhesion on painted surfaces. This has been intensively tested with the introduction of direct glazing. Since there are so many different paint surfaces due to the chemistry, process conditions, and requirements for the paints, there are three different approaches to achieve good adhesion for the windshield adhesive: 1. Bond on the surface of the filler or directly on the EC coat. This requires a cover tape applied before the coating of the paint. The masking tape must be then removable without any residues on the surface even after several bake cycles. And in addition, the paint must stick on the tape and should not chip when removed. The masking is an additional process with additional costs. The EC coat provides the highest adhesion strength with values in the range of 15 MPa. 2. The use of a primer to prepare the paint surface for the windshield bonding. This is a commonly used process. It is mostly a manual operation with the risk of contamination of other areas with the primer. But with specific formulated primers it is possible to achieve adhesion on all paint surfaces. 3. Bond directly on the top coat. This requires a thorough testing of all paint surfaces and a specific formulated adhesive adapted to that paint type. It means more quality control tests during production. But also in this case it is necessary to mask those car bodies which undergo a repainting process, because additional paint layers usually have less adhesion than the first one, which would create a weak interphase for the bonding system. 4. In general, the adhesion within the paint layers determines the strength of bonding on paint. High modulus windshield adhesive can build up stronger bonds that the paint which leads then to adhesive failure within the paint layer instead of failure within the adhesive itself. This applies also for other parts like, for example, a roof made of aluminum or carbon fiber composite installed at the assembly line.

46.6

Strength and Durability

Strength and durability is not inherent to an adhesive but it must always be seen in the context of the whole design of the bonded joint, the substrates, the expected permanent maximum and cyclic loads in combination with environmental influences like climate, radiation, and oxidation. Figure 13 shows a relation between strength and deformation for the world of adhesives. It visualizes in which strength range different adhesive types usually work, even if there may be exceptions. And it helps to think about the balance between strength and deformation. The range of strength of structural adhesives used in automotive production is marked at one end by windshield adhesives with rather low modulus and high elongation and at the other end by very high strength epoxy adhesives. The top end of adhesives for automotive bodies are crash-supportive epoxy-hybrid adhesives

1350 % 400

K. Dilger et al.

Elongation at break Windshield Type adhesive (PUR)

200

2C PUR; rubber based Acrylic, Rubber based and PUR

100

Crash supportive Epoxy hybrid

50 25 G-Modulus

Epoxy 3 3–8

50 5–12

100 14–18

300 30–40

1000 60–70

Level of lap shear strength (MPa)

Fig. 13 Landscape of adhesives for structural automotive applications (PUR polyurethane)

which have a somewhat lower G-modulus than very high strength epoxies, but significantly higher elongation. In between, there is a large area where the mechanical properties of the adhesives can be varied. In automotive applications, this broad range can be covered with 1- and 2-C polyurethanes, rubber- or acrylic-based adhesives, and epoxies or modified epoxies. There are also PVC- or acrylic-plastisol type products, but these provide due to their thermoplastic properties no real structural strength. The right approach for engineers is to define first the required mechanical properties of the adhesive for given substrate combinations with respect to the expected mechanical loads and then select the adhesive with the corresponding profile. This differs from the standard approach where the adhesive with the highest strength is selected for structural applications without really taking into account other important adhesive properties and the mechanical behavior of the substrates themselves. When designing bonded joints, the right bonding area and overlap length is the key element for a successful bonding to guarantee the required load transfer with respect to the given mechanical properties of the substrates. For example, when bonding metal to glass, a too strong adhesive in combination with a large bonding area or length will lead to glass breakage when exposed to very low temperatures. Or a too strong adhesive for the bonding of glass fiber or carbon fiber composites will lead to the delamination of the matrix resin from the fiber thus not providing the desired bond strength. By reducing the modulus of the adhesive and increasing the bonded area, the required load can be transferred without such delamination.

46

Automotive Industry

1351

It is important to know that adhesives need a certain area to be able to transfer the desired load. Adhesives with very high modulus and low elongation at break are limited when enlarging the bonding area since they have no uniform stress distribution within the bond line. By increasing the bonding area, adhesives with lower modulus and higher elongation at break can transfer even higher loads than hard and brittle high strength adhesives and are in addition more tolerant toward impact stress. What is the real strength required for structural bonding? This central question cannot be answered in a simple way. The most important property is not the lap shear strength but the G-modulus or E-modulus, which can be used in calculations for the stiffness of a bonded structure. The bonding of a sun roof lid gives an idea about this effect. In this example, the lid is bent as shown in Fig. 14 with a force of up to 500 N, and depending on the modulus of the antiflutter and hem flange adhesive, the deformation is measured and calculated. The calculation made by Deimel (1993) of the deformation depends on the modulus shows that the bending could be reduced from 7.8 to 4.8 mm at the given bending force of 500 N (Fig. 15). The maximum allowed bending was 5.8 mm. This is of real importance since at a car speed of more than 200 km/h the air would deform the lid too much and blow it away. The real bending measured with adhesives with 0 MPa (a deformable rubber), 5 MPa (a windshield adhesive), 20 MPa (a modified polyurethane (PUR) adhesive), and 100 MPa (a flexibilized 2C epoxy adhesive) showed a rather good correlation with the calculation. The stiffness increase was slightly higher than calculated which can be explained by the fact that the modulus was measured as secant modulus at one-third of the maximum shear strength. And in reality, at very low deformation, the modulus is higher. This example shows that for structural antiflutter bonding, a modulus of 50 to maximum 100 MPa will be enough to get the necessary stiffening effect. With a modulus of 1,000 MPa or more, the risk of surface marking will be much higher especially when using thinner metal sheets. And be aware that high strength steel sheets will not have less surface marking because it is the E-modulus of the steel which determines this property and not the yield strength. In this real case, the requirement from the design was the necessity of a very thin lid to save space overhead. Plastisol-based antiflutter material did not provide the required stiffness, and when using a very strong epoxy adhesive to get the maximum increase in stiffness, surface marking occurred. The solution was a modified PUR-adhesive with a G-modulus of 50 MPa limiting the bending to 5.2 mm but with no surface marking. The key to design bonded joints is the knowledge of E-modulus or G-modulus. These values can be used to calculate and predict the strength of a bonded joint. In Fig. 14 Bending of sun roof lid

F

1352

K. Dilger et al.

mm 8.0 7.6

Bending at 500 N

7.2 6.8 6.4 6.0 5.6 5.2

Calculated Measured

4.8 4.4

10−9 10−8 10−7 10−6 10−5 10−4 10−3 10−2 10−1 100 10+1 10+2 10+3 10+4 10+5 Shear modulus [MPa]

Fig. 15 Stiffness of sun roof lid depending on shear modulus

contrast, the value of the lap shear strength of a joint is of no use since this depends on the specific properties of the substrate and the geometry of the samples and thus not inherent to the adhesive properties. A good overview of the design and calculation principles of bonded structural joints can be found in a series of articles by Schlimmer (Bornemann and Schlimmer 2004a, b; Bieker and Schlimmer 2004). Design of joints with structural adhesives and sealants is also treated in ▶ Chaps. 27, “Design Rules and Methods to Improve Joint Strength,” and ▶ 28, “Design with Sealants,” respectively. A short example of a calculation is used to demonstrate this effect. When bonding a lap shear sample made from mild steel with yield strength of 280 MPa, a metal gauge of 1 mm, and a bonded area of 300 mm2 (12  25 mm2), the force to reach the yield strength is 1 mm  25 mm  280 MPa or 7,000 N. The lap shear strength of the adhesive will be calculated with 23.3 MPa. Due to the small increase of the strength of the steel on yield, in practice not more than 25 MPa can be expected. And there will be no adhesive providing more than 25 MPa, even if the inherent strength of an adhesive may be much higher. The other example will be a lap shear joint sample with 1 mm steel and an adhesive with 5 MPa lap shear strength (e.g., a windshield adhesive). In this case, when the bonding area is 1,800 mm2, which corresponds with an overlap length of 72 mm, there is also a failure in the steel. The lap shear strength is a measure which depends on the sample design whereas the G-modulus is an inherent adhesive property. The benefit of a lap shear test is a simple specimen for quality control of the combination of adhesive, adhesion, and surface quality.

46

Automotive Industry

1353

40

Stress t (MPa)

30

Section 3 20 Section 2

10

Shear modulus at 1/3 t max around 550 MPa

Section1 0 0

20

40

60

80

100

Strain g (%)

Fig. 16 Stress–strain diagram of a crash-suitable adhesive Fig. 17 Failure of a lap shear sample with a crash-suitable adhesive

The diagram in Fig. 16 shows three sections. Section 1 is the area where the adhesive shows almost linear behavior and the strength and the deformation of the bond can be predicted and calculated for the lifetime. Section 2 is an area where the adhesive shows resistance against short time high load without damage. And it can be seen that the maximum load in this section is almost in the range of 40 MPa, which is far above the strength level of the steel in the calculated example above. Sections 1 and 2 are typical for high strength structural adhesive, also for the brittle types. Section 3 shows the additional benefit of a crash-suitable adhesive because such adhesives start to deform significantly without break absorbing enough energy to keep the parts together. Figure 17 shows that these adhesives really work with a broken lap shear sample which proves the calculation above. Two areas are remarkable in Fig. 17. In area 1, the metal broke after it exceeded its tensile strength, and in area 2, very high stress concentration occurred due to the metal deformation. At failure, it was observed that at the edge in area 2 the adhesive started to crack but due to the built-in nanoscale crack stoppers it did not break. Such adhesives are designed that during curing phase separation occurs and phase of elastomeric nanoparticles is built, which act as crack stopper published by Burchardt et al. ( 2009). Such a property is crucial in case of a sudden overload through an impact. In particular, hard and brittle adhesives with very high strength (higher than 1,000 MPa

1354

K. Dilger et al.

Fig. 18 Impact peel test (crash) – ISO 11 343 Fig. 19 Impact peel test specimen

G-modulus) fail under such loads very fast and these are not recommended for automotive applications where impact resistance is needed. The impact performance is tested with an impact peel tester shown in Fig. 18, where a weight on a pendulum drives a wedge through the specimen and the forces are recorded. The test specimen is a wedge sample as illustrated in Fig. 19. The diagram of an impact peel test in Fig. 20 shows the difference of a crashsuitable adhesive in contrast with a very good structural automotive adhesive. The first peak of the curves is the initial peel resistance and the jagged curve gives the average value of the impact peel strength. Whereas the standard structural adhesive ends after 15 ms with an average peel force of 400 N, the crash-suitable one withstands for almost 20 ms providing a peel force of 900 N. The area under the curve is a measure of the energy absorption of the adhesives itself. Of course, the small amount of adhesive in a bonded structure is not absorbing the crash energy. It is the structure which is held together by such a high-performance adhesive thus providing better results than a spot welded part. More information on impact is given in ▶ Chaps. 21, “Impact Tests,” and ▶ 29, “Design for Impact Loads.”

46.7

Durability

A key question of bonded joints is the durability since adhesives are not metals and their performance is more dependent on temperature and environmental influences. And an engineer has the responsibility for the function of the structure over the

46

Automotive Industry

1355

2000 Structural adhesive Crash suitable structural adhesive

Force [N]

1500

1000

500

0 0

5

−500

10

15

20

Time [ms]

Fig. 20 Diagram of an impact peel test

designed lifetime even under harsh environmental conditions. How can the durability of an adhesive be evaluated? The durability of adhesive bonding is determined by four elements: 1. 2. 3. 4.

Durability of the substrates Durability of the adhesive itself Durability of the boundary layer Durability of the bonded system

The durability of the substrates and the bulk adhesives can be tested separately with shouldered test samples. These tests give an indication if the bulk material can provide the required performance. Testing of the adhesion within the boundary layer needs more effort. Metal adhesion in automotive industry always means adhesion on the oxide layer. This layer is attacked by moisture in combination with oxygen which leads to disbonding over a certain period of time. All surface preparations and specific ingredients of the adhesives have the goal to slow down this deterioration of the oxide interphase as well as possible. This is usually tested by cycles like the VDA test. And it can be seen that under these conditions, the oxide layer will be destroyed over time. In order to have a correlation of this effect in comparison with real-life corrosion results, every car manufacturer has its own cycles. An important contribution for a durable adhesive is the exclusion of moisture. This does not mean that the adhesive layer cannot get wet, but it must be able to dry up again and in cases where moisture may remain longer, it is recommended to protect the adhesive with a sealant, especially in the lower areas of the car. This is

1356

K. Dilger et al. (N) 6,000

1,200 MPa

Before storage

5,000

Shear stress

4,000 500 MPa 3,000

After storage 2,000

1,000

0 0

1

2

3

4

5

6

7

8

9

10

11

12

Storage for 6 week at 40°C and 100 r.h.

13

14 15 % strain

Fig. 21 Decrease in shear modulus of an epoxy adhesive after humidity storage

anyway a good solution, since the corrosion starts from the cut edges of the metal sheets which are not really protected by the adhesive squeezed out in an uncontrolled way. To simulate the effect of humidity on the adhesive, the samples should be conditioned in a humidity chamber (for example, according to DIN 50017) before testing, because the saturation of the adhesive with moisture will lead to a plasticizing effect and reduce the strength and modulus of the adhesive but reflect better the reality (see Fig. 21). Figure 21 shows the decrease of the modulus of a standard epoxy adhesive after storage at high humidity. Of course, this effect is dependent on the formulation of the adhesive, but all adhesives are affected by humidity to a certain degree. This effect cannot be detected with simple lap shear samples because the lap shear strength is almost unchanged (see calculation above). There is sometimes even a slightly higher lap shear strength observed since the plasticizing effect of the humidity reduces the brittleness of the adhesive and the failure occurs at slightly higher forces. The durability for a bonded system can be tested by applying cyclic loads on either simple lap shear specimens, more sophisticated H-specimens or on the whole bonded structure (e.g., the whole car body). It is recommended to run a corrosion test or at least a high humidity test upfront ( Fig. 22). Since adhesives are plastic materials, they may creep under high loads and therefore the expected static load is a determining factor for the durability. Compared to lap shear strength or tensile strength measured with shouldered test bars, the resistance toward static load is significantly lower. It is recommended to use only in the range of 3% to a maximum 10% of the maximum lap shear strength for the calculation of the sustainable permanent static loads for the use phase.

46

Automotive Industry

1357

Applied force Width of specimen

50 mm

Adhesive layer l b

500 mm

30 mm

Overlap length 82 mm

Test specimen hat profile

H-shaped specimen

Fig. 22 Test specimen for durability

Under dynamic stress, adhesives can withstand much higher loads (up to 30% of the maximum load). When measuring fatigue, the strength level of a windshield adhesive after more than 20 million cycles is in the range of 0.2–0.3 MPa, which is in the range of 3–5% of its initial lap shear strength (Fig. 23). When tested under the corresponding static load, nearly the same result occurs, but it is not clear if this correlation is valid for all adhesives. For windshield bonding, an additional durability issue is the resistance to UV radiation. The directly exposed bonding zone of the adhesive is destroyed by UV radiation if not protected by a screen print on the glass with sufficient UV absorption or with other methods of masking. Using only a black primer is not sufficient to guarantee a durable adhesion for the predicted lifetime. The UV transmission must stay below a value of less than 0.2%. Inspection at cars after 8–10 years lifetime in various environments built in 1987 and 1988 have shown that there was reduced adhesion in areas where the UV transmission was higher, because at that time the quality of the screen print on the glass was not on the level of today. Nevertheless, the windshields were still bonded to the frame without leakage which shows the robustness of this technology. Even after more than 30 years of use, polyurethane windshield adhesives show no relevant decrease in strength and do not lose their mechanical properties significantly. A precondition for a durable adhesion is, as with the body shop adhesives, the exclusion of permanent moisture. The combination of bonding with other joining methods like spot welding, riveting, or clinching called hybrid bonding or hybrid joining gives additional benefits. The punctual joining methods provide immediate fixation and they also take away the static load from the adhesive which gives superior dynamic fatigue strength due to an even stress distribution, especially in combination with very thin metal sheets. Hybrid bonding

1358

K. Dilger et al.

Maximum shear stress (MPa)

10

1

0.1 100

101

102

103 104 105 Number of cycles

106

107

108

Fig. 23 Fatigue curve of a windshield adhesive

will allow lightweight design by using thinner metal sheet made from very high strength steel without reducing strength and durability. Due to the material savings, the additional process cost will not be higher since the number of welded spots or rivets can be significantly reduced and the application speed of an adhesive is higher than that of spot welding. Therefore, the overall number of robots in a real production would be the same. The weight increase by the additional adhesive is in the range of a maximum of 2 kg for a whole car body, but the weight savings can be in the range of 20%. Thus, the savings in material for steel compensates the cost for the adhesive. The Wöhler-diagram in Fig. 24 demonstrates the difference in durability of different adhesive types. The 1C epoxy starts with the highest strength level but after ten million cycles, the remaining strength is lower than that of a polyurethane toughened epoxy which has a lower initial strength but due to its reduced crack sensitivity, a higher durability. The rubber-based adhesive and the 2C epoxy adhesive have a similar behavior. These tests were run with aluminum AlMg5Mn single lap shear samples and a surface treatment with Ti/Zr. The overlap length was 12.5 mm and the size of the metal plates 110  48  1.5 mm. The adhesives were conditioned with a 480 h salt spray test according to DIN 50021 before testing. Spot-welded samples would have already failed at the initial load level of all these adhesives. Real-life experience from cars shows that spot weld bonded areas provide a much better corrosion resistance than spot welded areas alone because the adhesive fills the joint preventing the intrusion of moisture. When bonding metal sheets with the typical thickness for automotive use in the range between 0.6 and 1.5 mm, bonded samples can typically bear higher loads during more cycles. The diagram in Fig. 25 illustrates this case. Spot weld bonded H-samples (see Fig. 22) start the fatigue test at a stress level of around 60 kN and end after six million cycles at around 30 kN, which is the starting level of the Metal Active Gas (MAG)-welded sample.

46

Automotive Industry

1359

10

Load level (KN)

mod. 1C EP

1C EP 1C Rubber 2C EP

1 103

104

105 Number of cycles

107

106

Fig. 24 Fatigue strength after 480 h salt spray test (EP epoxy)

Stress amplitude FA (kN)

100

Spot weld bonding Welding 10

Self piercing riveting Spot welding

1

104

105

106

107

Number of cycles until failure Fig. 25 Fatigue strength with different joining methods

Spot-welded single lap shear samples under the same conditions show a much lower force level at the end of the test. Since the performance of punctual fixations like rivets or spot welds is more sensitive to the thickness of the metal sheets, adhesive bonding exhibits even more advantage with thinner metals. Since this is a double

1360

K. Dilger et al.

logarithmic diagram, the difference in the load level of spot welding and bonding is 6–32. Therefore, with such hybrid bonding, the bond strength under fatigue is determined by the adhesive bonding alone. Premium car producers like AUDI, BMW, Mercedes, and even VW use bonding in combination with high strength steel. For example, the newest AUDI A 1 reports a bond length of 60 m (ATZ extra 2010). Whereas Fig. 17 shows the influence of adhesive properties on the durability, the diagram in Fig. 26 gives a comparison between the effect of adhesive bonding on steel and aluminum. In this case, the samples are hat profiles preconditioned at 70  C and 92% RH for 240 h and the distance of the spot welds is 50 mm. The load is expressed in the torsion moment, and it can be seen that a spot-welded aluminum profile with 2 mm thickness gives the worst results, even if it ends up at nearly the same level as the spot welded steel with 1 mm thickness. Comparing the result of 1 mm spot-weld bonded steel with 1.2 mm spot-weld bonded aluminum, the aluminum shows already a better durability. What is also obvious is how much improvement can be seen between the spotwelded 2.0 mm aluminum and the 2.0 mm spot-weld bonded aluminum. The distance between the spot welds is 50 mm which is longer than the usual 30 mm. With shorter distances between the spot welds, the results for the spot-welded samples would be better but will not reach the level of bonded samples. This is not only observed with these test specimens, but it was also proven in the comparison of real spot weld bonded cars, as shown in Fig. 27. Depending on the design of the car, the increase in static torsion stiffness is for steel cars between 15% and 25%. The 38% are a case where due to the longer spotweld distance and the weaker structure the improvement was significantly higher. Car 3 was a convertible with additional stiffening bars at the bottom; therefore, the improvement was lower. But what was rather unexpected were the results obtained for cars 5 and 6. Car 5 was a conventional steel car with a high structural stiffness,

Cyclic change in tosional moment

Nm/° 280 AlMg3; 2.0 mm spot weld bonded

240 200

AlMg3; 1.2 mm spot weld bonded

160

Steel;1.0 mm spot welded bonded

120

Steel;1.0 mm spot welded

80 40

AlMg3; 2.0 mm spot welded

0 104

107 Number of cycles

Fig. 26 Fatigue strength: comparison of spot welded and bonded aluminum and steel

46

Automotive Industry

1361

65

70 60

Increase %

50 38 40 27

30 20

23

20

14

10 0 1

2

3 4 Car number

5

6

Fig. 27 Stiffness increase of spot weld bonded cars measured before and after curing of the adhesive

and the improvement through bonding was with 20% in the expected range. The car 6 was the same design but made from aluminum. And since it was an experimental car, the aluminum sheets were stamped in the same tools like the steel sheets, which means the increase of the metal thickness for aluminum versus steel was in the range between 0.1 mm and 0.2 mm. But the increase in stiffness was 65%, which is much higher than with steel car and confirms the results of the hat profiles. And the weight for the car body in aluminum was 55% lower than for the same steel car. This concept is realized in the current Jaguar XJ. As already mentioned, the lap shear strength of an adhesive is not very relevant to the strength of a bonded joint because when enlarging the bonded area, the force can also be transferred with adhesives of lower strength, but remember the G-modulus, which is important for the stiffness of the structure. In the diagram in Fig. 28, the strength of spot welded and spot-weld bonded lap shear samples with 45 mm width and 16 mm overlap plus one spot weld is shown. It shows that strength level is only determined by adhesive bonding. After the failure of the bond, the spot weld can still bear the additional load as was shown in the research report from Hahn and his group (Hahn and Wiβling 2007). The calculated lap shear strength of the bond based on this diagram is about 11.1 MPa. One spot weld accounts for 3,500 N, and with a spot weld distance of 30 mm such a welded flange with a length of 1,000 mm, for example, would provide failure load of 116.6 kN. Assuming a lap shear strength for the adhesive of 11 MPa and an overlap of 12 mm for the bonded area, the corresponding failure load of the bonded flange with a length of 1,000 mm will be 132 kN. According to this calculation, a lap shear strength of 10 MPa is sufficient for bonding thin steel sheets in comparison with spot welding. The crack resistance of an adhesive is more important than its maximum strength.

1362

K. Dilger et al.

10,000

Lap shear strength Adhesive failure

Force (N)

8,000

6,000 Spot weld failure

4,000

2,000 Spot weld bonded

Spot welded

0 0

1

2

3

4

5

6

7

8

9

Elongation (mm)

Fig. 28 Failure diagram of bonded and spot weld bonded lap shear sample

And for the stiffness, only the G-modulus is important and with the rather thin adhesive layers of 0.3–0.5 mm in a spot-weld bonded flange, there will be no significant difference between an adhesive with 100 and 1,000 MPa G-modulus in the overall torsion stiffness of the car body. The diagram also shows the benefit of the hybrid joining since the failure of the spot weld occurs at higher elongation. Another important property for durability is the glass transition temperature, abbreviated as Tg. Since adhesives are organic polymers, their properties are much more temperature dependent than metals. And a measure for the temperature stability is the Tg. In Fig. 29, the dynamic mechanical analysis (DMA) of a crash-suitable epoxy adhesive is shown with a Tg of about 90–100  C depending on how the Tg is determined. Usually for body shop adhesives, the Tg should be above 80  C to avoid a strong change in the modulus of the adhesive because the modulus changes about a factor of 100 between 60  C and 120  C (see diagram in Fig. 22). There are two different ways to determine the Tg. One is the intersection point of the two tangents at 85  C and the other is the point of inflection of the storage modulus curve at 100  C. Since the Tg is not an exact temperature, it is better to look at the storage modulus curve to see the real difference in modulus below and above Tg. On the other hand, a low Tg (below 50  C) does not mean that the adhesive does not fulfill the requirements. Typical windshield adhesives always work at temperatures above their Tg, as it is shown in Fig. 30. But the decrease in the modulus for this type of adhesive in the working range of 40  C to +100  C is from 20 to 2, which is only about the factor of 10 lower.

Automotive Industry

1363

Storage Modulus E (MPa)

46

−20

0

20

40

60 80 Temperature (°C)

100

120

140

Torsion modulus (MPa)

Fig. 29 Determination of the Tg with DMA analysis of a structural adhesive

100

10

1 G - curve

−60

−40

−20

0

20

40

60

80

100

120

140

160

180

Temperature (°C)

Fig. 30 Torsion modulus over temperature of a 1C PUR adhesive

More information on the durability of adhesive joints is given in Part F.

46.8

Type of Failure

The same failures as in other bonding applications apply to adhesive bonding in automotive. They are:

1364

K. Dilger et al.

Failure of one of the substrates Failure of the adhesive Adhesion failure When designing a bonded joint, the goal is to achieve at least a break within the adhesive. This guarantees that the adhesion is higher than the cohesion and the adhesive is fully utilized. If failure in one of the substrate occurs, the bonding fulfills the requirements. Such failures can be much better predicted due to the inherent material properties of the substrates or the adhesive. But in reality, failure occurs in most cases within the adhesion zone. This is the main reason that engineers have some doubts to use adhesives in structural applications. When bonding metal substrates, it is the deterioration or corrosion of the oxide layer which destroys the adhesion zone. With plastic substrates, effects like plasticizer migration or residues of release agents are often reasons for failure, whereas moisture plays a less important role. In real cars, bonded structures show overall an even better durability than spot welds since the overlap of the metal sheets is much better protected against corrosion. When bonding plastic parts like Sheet Moulding Compound (SMC) or even carbon fiber composites, moisture will have the same plasticizing effect on the matrix resin as on the adhesive.

46.9

Quality Control

As mentioned in the introduction, adhesives are process materials. This is similar to a welding process. For these processes, it is very difficult to control the quality of a bonded joint in a nondestructive way. Therefore, it is crucial to control the process parameters of each production step, to check the properties of the adhesive with the real substrates and it is necessary to control the surface conditions of the substrates itself. And even with nondestructive testing methods, it is not possible to check the quality and durability of the adhesion itself. This can only be done with thorough adhesion tests upfront with original substrates and original process conditions. Therefore, the way to a durable bond goes along with a survey of all relevant process parameters. If this is guaranteed, bonded joints have the same life expectancy as other parts of the car body structure. It is recommended to run separate lap shear samples along the whole process which can then be tested separately to have a continuous process survey. But this does not differ from the quality control system of the paint process. The quality of a bonding process depends to a high degree on the quality of the personnel. Therefore, a similar level of education is needed as for the welding specialists. The Institut für Angewandte Materialforschung (IFAM) in Bremen offers education courses for three different levels of expertise and everyone who wants to use structural adhesive bonding. See Part H for more information on Quality control.

46

Automotive Industry

1365

46.10 Repair and Recycling It is true the repair of bonded joints is different from that of bolt-on parts. But in comparison with riveted, spot welded, or welded structures, the removal of bonded parts may be even easier. When repairing bonded metal structures, the removal of a damaged part is done by mechanical abrasion or with hammer and chisel. The adhesive is then removed by abrasion like it is done with painted surfaces. More challenging in the repair case is to bond a new part on the old structure. It is not possible to use the same adhesive as in the production line since the process conditions are different. In this case, 2-C high strength adhesives with improved impact resistance can be used as published by Lutz (2009). And to process these adhesives correctly, well-educated personnel is required. Windshield adhesive can be removed either with a steel wire or with specific cutting tools. This is nowadays a standard procedure in repair shop and it can even be done within 1 h on the road. In this context, it is important that the bonding area is designed in a way that enough space for the cutting tools ore wire is available. There is no need to remove the old adhesive layer completely. It is recommended to bond on a maximum 1 mm thick layer of the old windshield adhesive after the proper surface preparation with a primer which gives additional corrosion protection when the paint layer suffered damage when removing the windshield. With the idea of recycling, adhesively bonded parts are under pressure since the common perception is that such parts cannot be separated again. And it is true if dissimilar materials are bonded together the separation of the different substrates needs more effort. There are activities to develop adhesive which disbond on command, but it is almost impossible to have durable adhesion under all relevant conditions and in addition an easy disbonding, for example, under high temperatures or by chemical reaction. See ▶ Chap. 59, “Recycling and Environmental Aspects”. For bonded metal parts, this is in the recycling case not a real issue since the amount of adhesive in a car body structure is with about 2 kg lower than the weight of the paint. When shredding the metal parts, adhesives as plastic materials can be separated in the same way as the paint and other plastic residues.

46.11 Conclusions Adhesive bonding for structural applications has become an important joining technology in the production of cars. It is for decades a well-established method to join doors and hoods and install glasses. And with a new generation of impactsuitable structural adhesives, premium cars have introduced hybrid joining to build lighter cars with increased impact performance. This chapter may help to exploit the full potential and the benefit of adhesive bonding in combination with other joining methods for cars. In particular, the engineers of premium cars in Germany have already started to trust such a technology and use it to a larger extent. It takes time for

1366

K. Dilger et al.

the implementation because the full benefit can only be achieved when the car body is designed for this technology. The main obstacles are the lack of education and experience of the design engineers who have to implement adhesive bonding.

References ATZ extra (2010) Presentation of the new AUDI A:1 Bieker C, Schlimmer M (2004) Berechnung und Auslegung von Klebverbindungen (Teil 3). Adhäsion, Kleben & Dichten 7:38–42 Bornemann J, Schlimmer M (2004a) Berechnung und Auslegung von Klebverbindungen (Teil 1) Adhäsion. Kleben & Dichten 5:30–33 Bornemann J, Schlimmer M (2004b) Berechnung und Auslegung von Klebverbindungen (Teil 2) Adhäsion. Kleben & Dichten 6:40–42 Brockmann W, Grüner KJ (2003) Betrachtung zum Kriechverhalten von Haftklebstoffen, 17. Internationales symposium Swissbonding, Rapperswil Burchardt B, Diggelmann K, Koch S, Lanzendörfer B (1998) Elastic bonding. Verlag Moderne Industrie, Landsberg Burchardt B, Schulenburg JO, Linnnenbrink M (2009) New building blocks for lightweight structures. Adhesion, Adhesives & Sealants 10:22–27 Cognard P (2006) Handbook of adhesives and sealants, vol 2, General knowledge, application techniques, new curing techniques. Elsevier, New York Deimel A (1993) Fachhochschule München, Einfluss der Klebstoffsteifigkeit auf das Verformungsverhalten geklebter Fahrzeugkomponenten, Diploma work Hahn O, Wiβling M (2007) Methodenentwicklungzur Berechnung von höherfesten Stahlklebeverbindungen des Fahrzeugbaus unter Crashbelastung. Abschlussbericht Forschungsvereiningung Stahlanwendung e.V, Düsseldorf Industrieverband Klebstoffe IKV 2017: www.klebstoffe.com/die-welt-des-klebens/anwendungsg ebiete/fahrzeugbau.html Lutz A (2009) Crash resistant two-component adhesives for repairs “As good as new”. Adhesion extra Adhesives & Sealants 10:28–29 Rechner R, Jansen I, Beyer E (2010) Laser- und Plasmaverfahren im wirtschaftlichen Vergleich. Adhäsion, Kleben & Dichten 1–2:36–43 Schulz D (2010) Gut gereinigt ist halb geklebt. Adhäsion, Kleben & Dichten 7–8:18–22 Stepanski H (2010) Punktschweisskleben im Automobilbau. Adhäsion, Kleben & Dichten 5:30–35, 6:35–41 Wisner G, Stammen E, Dilger K, Spiekermeier A, Halanesh M, Hübner S, Behrens B-A (2015) Kleben und Umformen von Stahlblechen in Bonded-Blank-Technik für den AutomobilRohbau. In: DVS-Berichte Band 315 (2015). DVS Media GmbH, Düsseldorf, S, pp 171–177 Wisner G, Stammen E, Fischer F, Dilger K, Zillessen A, Brodel M. (2014) Accelerated adhesive bonding of wood panels for prefab houses. In: Martínez JMM (ed) EURADH 2014 10th Europeanm adhesion conference – proceedings. 22.04. – 25.04.2014, Alicante/Spain 2014. pp 307–310, 2014, ISBN. 978-84-616-8627-8

Adhesive Bonding for Railway Application

47

Yasuaki Suzuki

Contents 47.1 47.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesive Bonding Application for Current Rolling Stock . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.2.1 Railcar Structures and Their Fabrication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.2.2 Application of Adhesives in Rolling Stock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.2.3 Use of Sealants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.2.4 Pressure-Sensitive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.3 Application of Adhesives for Prototype Railcars in Japan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4 Application of Structural Adhesive for Rolling Stock over the World . . . . . . . . . . . . . . . . 47.4.1 Britain (Seeds 1984) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4.2 France . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4.3 Germany . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4.4 Switzerland . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.4.5 Indonesia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47.5 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1368 1368 1368 1370 1377 1378 1380 1383 1383 1387 1388 1389 1389 1390 1390

Abstract

Adhesion is an indispensable joining technology for railway industries. Adhesives are used for the fabrication of almost all rolling stocks. In the steel main structure of conventional rolling stocks, adhesives are applied for bonding decorated aluminum sheets of wall and ceiling to frames, bonding of floor covering to floor plate, and fixing heat insulating material to the inside of carbody. Large amount of sealants is also used even in the conventional rolling stocks. Recent high-speed trains have modern light-weight structures consisting of aluminum hollow extrusions, sandwich panels, or carbon fiber reinforced plastics. For the structures, adhesion plays an important role as a structural joining method. Even for conventional rolling Y. Suzuki (*) Suzuki Adhesion Institute of Technology, Ichinomiya, Aichi, Japan e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_47

1367

1368

Y. Suzuki

stocks, the use of adhesives is increasing. For instance, weldbonding is nowadays experimentally applied to the aluminum body structures. Moreover, adhesive is used to smooth the outsides of each carbody and between the outersurfaces of carbody and window. As a result, aerodynamic noises inside and outside the train and energy consumption are considerably reduced. Maglev trains will be increasingly use in the future, and the use of adhesives will expand due to the demand of reducing weight and energy consumption. In this chapter, the cases in Japan are mainly explained. World-wide cases are shown in Sect. 4.

47.1

Introduction

Railways have progressed with three important missions such as high speed, safety, and mass transportation. In particular, a high speed rail system called “Shinkansen,” which means new railway trunk line, has been developed in Japan over the last four decades and has significantly contributed to the development of Japan. Nowadays, the network is prolonged to many subsidiary regions including Hokkaido island through an undersea tunnel even though it is far from the main island of Japan. In addition, Chuo Shinkansen as a detour of Tokaido Shinkansen that utilizes the Superconducting Maglev will be constructed (Central Japan Railway 2009a). Rolling stock, which means railcars of broad sense, consumes the smallest energy per person-distance among all the transportation systems. It is about half that of buses and one-sixth of that of cars (Kamiura et al. 2000). In addition, since the power source is mainly electric, the consumption of fossil fuel by rail systems is relatively small. Therefore, we might say that trains are the most environmentally friendly transportation system. Recently, high speed, passenger comfort, safety, and being environmentally friendly are indispensable for railway systems, and they lead to the importance of weight saving, high strength, high rigidity, noise insulation, vibration damping, thermal insulation, incombustibility, and recyclability. In addition, as the service life and model-change intervals of railcars are shortened, the cost reduction has also become an important factor (Miyamoto 1998). To meet such vast demands, introducing multimaterial structures, in which different materials are joined together and used respecting their own characteristics, is a good solution. The most promising method to join the multimaterial structures of trains is adhesion.

47.2

Adhesive Bonding Application for Current Rolling Stock

47.2.1 Railcar Structures and Their Fabrication The main structure of railcars is called the “body structure.” It consists of six panels: side constructions, roof construction, end constructions, and an underframe. Stainless steel and aluminum alloy are increasingly used instead of carbon steel for weight

47

Adhesive Bonding for Railway Application

1369

reduction and easy maintenance as the main structure. They are mainly fabricated by welding. By eliminating the thickness increment for corrosion, stainless steel body structure can be lighter than carbon steel body. For stainless steel body structures, panels are joined using spot welding on a frame joined by welding. Figure 1 shows an example of the body structure of 205 series electric railcar for commuting made of stainless steel (Matsuzawa 1986). As shown in this figure, corrugated stainless steel plate are spot welded on an underframe. Aluminum alloy cars are fabricated with another method. Large-scale extrusions are used in order to reduce the total length of welding and they are joined by metal inert gas (MIG) or laser welding technique. For instance, E2 series and 700 series Shinkansen cars have a double-skin structure consisting of large-scale aluminum hollow extrusions welded continuously along the longitudinal direction, as shown in Fig. 2 (Abiko and Kobayashi 1998). This method can use automatic processes because there are very few columns or beams and that contributes greatly for cost reduction. Recently, another novel method called Friction Stir Welding, whose thermal input to joints is smaller than the conventional welding methods, has been gradually adopted for new types of aluminum cars instead of MIG welding in order to reduce the thermal deformation and enhance the fabrication accuracy. In the body structures, modules such as heat insulating materials, floor deck, floor covering (made of non-rigid polyvinyl chloride (PVC), polyolefin, or rubber), inside panels (made of decorated aluminum sheet with melamine resin, aluminum panels pasted non-rigid PVC film, or fiber reinforced plastics (FRP)), windows (made of glass or polycarbonate), interior parts such as seats, lighting equipments, air conditioning equipments, and wiring and piping are equipped. In addition, hard wiring

Belt rail

Entrance post

Carline Roof board Running board & Rain gutter

Cant rail Frieze board Window header

Entrance frame End plate End post

Pier panel Stiffener Side post

End beam Centre sill Bolster

Cross beam

Steel deck

Wainscot panel

Rocker rail

Side sill

Fig. 1 Body structure of 205 series electric railcar for commuting made of stainless steel (Matsuzawa 1986)

1370

Y. Suzuki Roof board (A6N01S–T5)

Frieze board (A6N01S–T5)

Pier panel (A6N01S–T5)

Floor (A6N01S–T5) Cross beam (A7N01S–T5)

Side sill (A6N01S–T5)

Fig. 2 Body structure of 700 series Shinkansen made of large-scale aluminum hollow extrusions (Abiko and Kobayashi 1998)

and heavy equipments such as transformers or compressors are placed under the floor, that is, the complete sequence of rolling stock fabrication.

47.2.2 Application of Adhesives in Rolling Stock In current rolling stock, adhesives are mainly utilized to join interior equipments. Table 1 summarizes typical adhesive applications for rolling stock use (Suzuki 2006, 2009). Figure 3 shows the share ratio of adhesive types and purposes in the cases of suburban electric cars and Shinkansen cars (Suzuki 2006, 2009). The main applications of adhesives are: 1. Ceiling board and air ducts The ceiling board of cars has a sandwich structure consisting of decorated aluminum surface plates with melamine and incombustible lightweight board cores bonded with two part epoxy adhesives, as shown in Fig. 4 (Suzuki 2007). The space between the ceiling board and the roof sheet is used as air conditioner ducts in the case of electric railcar for commuting, as shown in Fig. 5 (Suzuki 2007). 2. Partition panel The partition of Shinkansen cars has a sandwich structure made of similar materials as the ceiling mentioned above. The bonding of the sandwich panels is performed with two part epoxy adhesives too.

47

Adhesive Bonding for Railway Application

1371

Table 1 Use of adhesives in present rolling stock (Suzuki 2006, 2007) Position Roof Ceiling board Air duct Partition panel Heat insulating material Decorated wall sheet

Entrance Floor

Door

Seat

Inspection lid Parts inside of the car

Adherends Electric insulated roof sheet/roof sheet Decorated sheet with melamine resin (Al)/ incombustible lightweight board/Al or FRP panel Air duct plate(Al)/polychloroprene sponge or soft polyurethane foam Decorated sheet with melamine or PVC resin (Al)/incombustible lightweight board/plate (Al) Oxidized acrylic fiber or carbon fiber/outside plate (steel, SUS, Al)

Adhesives Nitrile rubber Epoxy

Application Spray Spatula

Polychloroprene rubber Epoxy

Paint brush or spray Spatula

Nitrile rubber

Spray

Spacing plate (incombustible hardboard, rigid PVC)/frame (steel, SUS, Al) Spacing plate (incombustible hardboard, Rigid PVC)/decorated sheet with melamine or PVC resin (Al) Liner under threshold plate (rigid PVC)/side sill Resilient floor covering (nonrigid PVC)/floor plate(Al) Resilient floor covering (nonrigid PVC)/floor compound material (epoxy resin) Resilient floor covering (nonrigid PVC)/floor coating (polyurethane resin) Resilient floor covering (polyolefin resin)/ floor plate(Al) floor plate(Al)/Al honeycomb/floor plate (Al) (sandwich panel) Plate(SUS)/paper or Al honeycomb/ decorated plate (Al) Plate (SUS)/paper or Al honeycomb/plate (SUS) Plate (Al)/paper or Al honeycomb/plate (Al) Decorated side plate (Al)/extrusion (Al) Seat frame (Steel, Al)/soft polyurethane foam

Nitrile rubber Nitrile rubber

Paint brush Paint brush

Nitrile rubber

Paint brush

Nitrile rubber Nitrile rubber Polyurethane Polyurethane Film type epoxy

Spray Spray Spray Spray Hot-press

Epoxy Epoxy Epoxy

Spatula Spatula Spatula

SGA Polychloroprene rubber SGA SGA

Gun Paint brush

Decorated sheet with melamine resin (Al)/ extrusion (Al) Incombustible lightweight board/plate (Al) Keeping box of cassette tape for announcement on the train (PMMA/PMMA) Magazine rack in the first class car (PMMA/ PMMA) Circuit-breaker cover (PMMA/PMMA)

Methylene dichloride (solvent bonding)

Gun Gun Syringe

(continued)

1372

Y. Suzuki

Table 1 (continued) Position PVC pipe Etc.

Adherends Joint part of rigid PVC pipe Metal/metal Rubber/metal

Adhesives PVC SGA Chloroprene rubber

Application Paint brush Gun Paint brush

Al aluminum, SUS stainless steel, PVC polyvinyl chloride, SGA second generation acrylic adhesive, PMMA polymethylmethacrylate

43.6% 1.3% 4.2% 50.9%

1: Nitrile rubber 2: Acrylic resin 3: Polyvinyl chloride resin 4: Polychloroprene rubber

1: Nitrile rubber 2: Polyvinyl chloride resin 3: Polychloroprene rubber

49.8% 3.8% 46.4%

Air duct 9.9% Rubbers 5.4%

4

Rubbers 4.6%

3 Pipe 3.8%

2

Air duct 18.8%

Heat insulating material 27.5% Suburban 75kg/car

Floor sheet 31.2% Shinkansen Decorated Al & 82kg/car spacing plate

Pipe 4.2% 3

18.7%

2 Resilient floor covering 27.5%

Etc. (including SGA) 1.3%

Resilient floor covering 31.2%

1

1 Electric insulated roof sheet 0.9%

Decorated Al & spacing plate 15.2%

Fig. 3 Share ratio of adhesive types and purposes for suburban electric car and Shinkansen car (Suzuki 2006, 2009)

3. Ceiling of decks and inspection lid The ceiling of decks and inspection lid for Shinkansen and limited express cars are made of aluminum corrugated panels (Fig. 6) bonded with an epoxy adhesive (Sugano 1992). Limited express is a special train faster than ordinary expresses in Japan. 4. Heat insulating material Heat insulating material (thickness:30–50 mm) is usually installed between painted side plates and roof sheets because the temperature of the outside plates can rise up to 80  C and drop below –30  C. Typical materials for heat insulating are oxidized acrylic fibers or carbon fibers. They are bonded on the back of outside plates with heat resistant nitrile rubber adhesives. In some cases, 20-mmthick heat insulating material is bonded on the back of decorated inside walls.

47

Adhesive Bonding for Railway Application

1373

Fig. 4 Sandwich structure for ceiling and duct board (Suzuki 2007) Polyurethane coating Roof board

Heat insulating material (oxidized acrylic fiber)

Heat insulating material (polychloroprene sponge bonded with polychloroprene rubber adhesive) Air duct Aluminum sheet

Melamine decorated aluminum sheet

Polyurethane adhesive

Core board (incombustible light-weight board)

Fig. 5 Air duct between the ceiling board and the roof board (commuting electric car) (Suzuki 2007) Aluminum sheet (thickness: 0.5 mm)

5.3 mm

Epoxy adhesive

Polyvinyl chloride decorative sheet

Corrugated aluminum sheet (thickness: 0.2 mm)

Fig. 6 Sandwich panel for door deck ceiling and inspection lid (Sugano 1992)

1374

Y. Suzuki

5. Inside wall decorating sheets Inside walls, which are made of aluminum sheets decorated with melamine resin or non-rigid PVC sheets, are joined to the frame using adhesives or screws. In almost all cases, a spacing plate made of flame retardant hardboard or rigid PVC is inserted between the inside walls and the frames to adjust their flatness and to prevent generation of creaky noise. The spacing plate is bonded on frames and the back of walls with nitrile rubber adhesives. Decorated aluminum sheets with PVC are used for Shinkansen and limited express cars. Aluminum sheets are joined to the frame using adhesives or screws, and a primer is applied on the surface. Nonrigid PVC sheets having a pressuresensitive adhesive layer are bonded on the surface. 6. Floor plate and floor covering Typical floor structures of cars are shown in Fig. 7 (Suzuki 2007). The case of 700 series Shinkansen car is shown in Fig. 7a. A heat insulating material made of oxidized acrylic fibers is installed on the aluminum extrusions. Floor plates of aluminum-aluminum honeycomb sandwich structure bonded with a film type epoxy adhesive are screwed on it. Floor covering made of nonrigid PVC is bonded with a nitrile rubber adhesive. Floor structures of limited express type, suburban, for commuting, nongovernmental railroad, and subway electric railcars are shown in Fig. 7b. In them, floor compound material made of epoxy resin and light weight aggregate

Resilient floor covering (Non-rigid PVC) Floor panel (Al/Al honeycomb/Al) Film type epoxy adhesives Floor compound material (Epoxy resin) Duct

Duct

Resilient floor covering (Non-rigid PVC)

Oxidized acrylic fiber

a

700 series Shinkansen car

b

Steel deck Limited express type, suburban, for commuting, nongovernmental railroad, and subway electric railcars

Resilient floor covering (Non-rigid PVC) Polyurethane coating Aluminum hollow extrusion

c

Subway electric car

Fig. 7 Typical floor structures of railcars (Suzuki 2007)

47

Adhesive Bonding for Railway Application

1375

is plastered and cured on steel deck, and PVC floor coverings are bonded on the surface of it with nitrile adhesives. Floor structures of subway electric cars made of aluminum are shown in Fig. 7c. PVC floor coverings are bonded with a two-component polyurethane adhesive on polyurethane-coated floor plate of aluminum hollow extrusions. 7. Cover-all hood As shown in Fig. 8, N700 series Shinkansen cars adopt cover-all hoods between cars to reduce external noise and running resistance (Central Japan Railway 2009b). They consist of soft elastic materials bonded adhesively and have led to energy savings as well as reduction of noise outside and inside the car. 8. Other items Door panels have a sandwich structure consisting of aluminum or stainless steel surface plates and aluminum or paper honeycomb core bonded with two part epoxy adhesives. On inspection lids, switchboard covers, and seat side panels, decorated aluminum sheets with melamine resin are bonded with secondgeneration acrylic adhesives (SGAs) to aluminum frames or stiffners. Figure 9 shows the cross section of a railcar with adhesively bonded parts (Suzuki 2007). 9. Carbon fiber reinforced plastic (CFRP) bogie (Nishimura 2014) Figure 10 shows a bogie that adopted CFRP leaf springs developed by Kawasaki Heavy Industries Co. (Nishimura 2014). In this bogie, CFRP is adopted as the main structure of the bogie frame, and as shown in Fig. 11, the CFRP frame is provided with a suspension function so that the primary suspension is unnecessary. This CFRP spring has a three-layer structure of CFRP/GFRP/CFRP using adhesive, and the upper and lower CFRP layers are designed to support the bending load and the intermediate layer GFRP supports the shear load (Nishimura 2014). Since the CFRP springs have structures that only ride on axles, those have the feature that it is easy to absorb the irregularity of the railway, and it reduces the wheel load loss by half. This bogie has been proved to fulfill all requirements of Fig. 8 Cover-all hood used in N700 series Shinkansen car (Central Japan Railway 2009b)

1376

Y. Suzuki

Poly chloroprene rubber sponge (bonded with poly chloroprene rubber adhesive)

Roof board (FRP coating) Rain gutter Airduct

Spacing plate (flame retardant hardboard bonded with nitrile rubber adhesive)

Air duct

Heat instating material (oxidized acrylic fiber bonded with nitrile rubber adhesive) Inspection lid Spacing plate (flame retardant hardboard bonded with nitrile rubber adhesive)

Indicator

Decorated aluminum sheet

Sealant (polysulfide) Advertisement frame Window glass Inspection lid Flame retardant hardboard or non-rigid hard polyvinyl chloride bonded with nitrile rubber adhesive Sealant (polysulfide) Decorated aluminum sheet Heat instating material (oxidized acrylic fiber bonded with nitrile rubber adhesive)

Sealant (oil based caulking compound)

Resilient floor covering (non-rigid polyvinyl chloride bonded with nitrile rubber adhesive)

Outside plate Sealant (polyurethane)

Side sill

Steel deck Floor compound material (epoxy resin)

Fig. 9 Adhesive bonded or sealant filled parts in cross section of a railcar for private railways (Suzuki 2007)

Fig. 10 Carbon fiber reinforced plastic (CFRP) bogie (Nishimura 2014)

47

Adhesive Bonding for Railway Application

1377

Fig. 11 CFRP spring which serves two functions of side beam and axial spring (Nishimura 2014)

regulation 49 CFR 213.333 on traveling safety of rolling stock by the Federal Department of Transportation of the United States of America. The mass per one CFRP spring is about 50 kg, and the weight reduction of about 450 kg is achieved by one bogie (about 900 kg per one car). This bogie was first adopted as a 6000 type train of Kumamotodentetsu Co. in 2014 and was adopted as a 7200 series suburban type DC train in Shikoku Railway Company in 2016.

47.2.3 Use of Sealants Sealants are widely used in railcar structures. Sealants can protect the structure and equipment from corrosion, increase the service life, and keep the airtightness of railcars in air-conditioning. For high-speed trains such as Shinkansen, airtightness is important because the outer pressure of the trains varies when they enter or exit from a tunnel. The same phenomenon can also occur when trains cross each other. In these cases, airtightness can keep passengers comfortable. Table 2 shows examples of sealant application for railcar structures (Suzuki 2007). Figure 12 also shows some sealant applications (Japan Association of Rolling Stock 2005, 2006). Figure 13 shows the ratio of sealant types used in the structures of stainless cars for private railway and Shinkansen (Suzuki 2007). In Japan, there were a national railway owned by the government and private railways owned by companies previously. The national railway has been privatized and separated to JR companies, so that JRs are not national railways anymore and are included in the category of “private railways” in definition. However, JRs are not called “private railways” due to the convention. The word “private” indicates the railways except JRs and municipal transportations in Japan. Generally, silicone sealants have good electric insulation, weather resistance, and good bonding strength. However, they also have disadvantages such as getting dirty due to contamination and low paint adhesion. Therefore, silicone sealants are applied to roof, equipments under the floor or around pipeworks. Since modified silicone sealants can be painted, they are applied to exteriors as waterproof or decorative sealants. Polyurethane sealants have a very good weather resistance and are, therefore, used for waterproof purposes. Although polysulfide sealants are two-part type and need primers, their water and weather resistance are good enough to apply to join and seal glass to metals. Nondrying type oil-based caulking compounds have good insulation and environmental resistance, but their bond strength is not high.

1378

Y. Suzuki

Table 2 Use of sealants in present rolling stock (Suzuki 2007) Position Body structure

Roof top

Window & door

Interior

Equipment

Applied parts Headlight mount End surroundings Entrance post and outside plate surroundings Outside plate seam (stainless car) Entrance threshold plate Bodyside door pocket Outside plate hem – side sill Headlight – outside plate (aluminum car) Pantograph base Ventilator base Rain gutter (steel) – outside plate Up and down footstep Side window unit, end window unit, and cab front window unit Train destination indicator window Sliding door window and hinged door window (outside) Sliding door window and hinged door window (inside) Polycarbonate window – outside plate (Series 100 Shinkansen passenger’s room) Hatten Center pin cover Air duct insulator watertight treatment Piping inside room End treatment of electric cable pipe Outlet port of equipment Riser pipe on floor Canon plug Water pipe fitting

Sealant Modified silicone Modified silicone Modified silicone Modified silicone Polyurethane Polyurethane Polyurethane Silicone Polyurethane, silicone Silicone Polyurethane Polyurethane Polysulfide, modified polysulfide Polysulfide, modified polysulfide Polysulfide, modified polysulfide Polyurethane Silicone Polyurethane, oil-based caulking compound Polyurethane Butyl rubber Oil-based caulking compound Oil-based caulking compound Oil-based caulking compound Oil-based caulking compound, silicone Oil-based caulking compound, silicone Modified silicone, silicone

All the sealants are applied with caulking guns

Therefore, they are not suitable for air tight applications, but for equipments under the floor or around pipeworks. Joining detachable parts is one of the oil-based caulking compound applications.

47.2.4 Pressure-Sensitive Adhesives Typical applications of pressure-sensitive adhesives (PSAs) are shown in Table 3 and mentioned before (Suzuki 2007).

47

Adhesive Bonding for Railway Application

1379

Modified polysulfide or polysulfide Silicone

Oil based caulking compound or polyurethane

Polyurethane

(Unit side window and end window)

(Pantograph base) Polyurethane

Polyurethane Polyurethane or oil based caulking compounds (Hatten) (Entrance threshold plate) Through pipe Silicone Airtight plug

Modified silicone

Upper surface of underframe

(Airtight treatment of vertical riser pipe on floor)

Modified silicone (End surroundings)

Fig. 12 Sealant application in railcar structures (Japan Association of Rolling Stock 2005, 2006)

1. Bonding Decorating Aluminum Sheets to Spacing Plates As mentioned in Sect. 2.2 (5), when decorating aluminum sheets are bonded to spacing plates, double-side-coated PSA tapes are used in some cases. Nitrile rubber adhesives must be applied to the surface of the spacing plate and dried in advance. 2. Aluminum Panels Covered with Nonrigid PVC Film As mentioned before, Shinkansen and limited express type electric railcars have aluminum panels covered with nonrigid PVC film bonded with PSA. The panels are made with a press. 3. Name Plates Name plates are bonded to the structure using both double-side PSA tapes and rivets.

1380

Y. Suzuki Others 4%

Oil based caulking compound 5%

Others 9% Butyl 10% Polyurethane 39% Suburban stainless 50kg/car

Modified silicone 21%

Polysulfide 21%

Silicone 21%

Polysulfide 41% Shinkansen 60kg/car

Butyl 10%

Polyurethane 31%

Fig. 13 Ratio of sealant types used in the structures of suburban stainless cars and Shinkansen cars (Suzuki 2007) Table 3 Use of pressure-sensitive adhesives (PSAs) in present rolling stock (Suzuki 2007) Position Decorated wall sheet

Adherends Spacing plate (incombustible hardboard, rigid PVC)/decorated aluminum plate Nonrigid polyvinyl chloride film/Al plate

Name plate

Name plate/Outside skin, inner decorated plate, polyvinyl chloride films having PSA Color strip (Nonrigid polyvinyl chloride)/outside plate (Al, SUS)

Discriminating strip

Supply Double-side PSA-coated tape Acrylic PSA(back side) Double-side PSA-coated tape Double-side acrylic PSA-coated tape

Al aluminum, SUS stainless steel, PSA pressure sensitive adhesive

4. Discriminating Strips (color strips) In stainless cars and aluminum cars, discriminating colored PVC strips, which have PSA on one side, are pasted along the outer surface of their bodies in the longitudinal direction.

47.3

Application of Adhesives for Prototype Railcars in Japan

1. CFRP Body Structure Weight saving is a very important issue, even for trains. Sophisticated lightweight materials have recently been introduced in railcar body structures. Phenol resinbased carbon fiber reinforced plastics (flame resisting CFRP), for instance, have been applied to a prototype structure shown in Fig. 14a (the outside) and (b) (the inside) (Suzuki and Sato 1993). The structure is 6 m in length and was made by Railway Technical Research Institute, Nippon Sharyo, Ltd., and Toray Industries Co., Ltd. in cooperation, using CFRP plates. The structure is made from curved outside CFRP plates and stiffners as a unit by the pultrusion technique. The

47

Adhesive Bonding for Railway Application

1381

Fig. 14 A prototype carbon fiber reinforced plastic (CFRP) body structure for high speed trains using panels molded by pultrusion method (Suzuki and Sato 1993)

300-mm-width CFRP panels were joined to aluminum alloy frames with Huckbolts®(Alcoa co.) and a two-part room temperature setting epoxy adhesive. According to the results of strength tests and a finite element analysis, the structure has equal strength and stiffness as structures made only of aluminum alloy. 2. CFRP Roof Construction A prototype body construction of Shinkansen was experimentally produced by Tokyu Sharyo Corporation. This had a roof construction consisting of four blocks made of phenol resin-based CFRP, as shown in Fig. 15 (Matsuoka and Aso 1992). Each block had different structures which were compared in terms of strength and weight. Block 1 and 2 had CFRP carlines and purlines crossed in a grid and bonded to a CFRP skin. The frames and skin of block 2 were thicker than those of block 1. In block 3, the structure had a sandwich panel (30 mm thick) of CFRP skins and a polyurethane foam core between carline without purline. Block 4 consisted of only a sandwich structure (50 mm thick) of CFRP skins (1 mm thick) and paper honeycomb core filled with a phenolic foam without frames. The

1382

Y. Suzuki 5700

5700

5700

2505

5700

Block 2

Block 3

Block 4

3500

3500

Block 1

2500 3100

3500

2500 17500

3500

24500 Unit:mm

Fig. 15 Carbon fiber reinforced plastic (CFRP) roof construction divided into four blocks (Matsuoka and Aso 1992 and Matsuoka 1992)

structure of block 4 was most effective to reduce the weight because ceiling boards and heat insulating materials could be omitted and air ducts were simplified due to adoption of curved sandwich structures. The aluminum rims were bonded to the edges of each block using a heat setting epoxy adhesive. A double lap joint in which CFRP side was double with 80 mm lap length was used. The aluminum rims of block were welded to the other of it’s and aluminum side and end constructions, as shown in Fig. 16 (Matsuoka 1992). Due to the use of this structure, the weight (4.3 t) of body structure was lighter than its of the present Shinkansen. The stiffness and airtightness of this prototype body structure were satisfied usefulness. 3. Weldbonding Structures (Sasaki 1995) East Japan Railway Company (JR-East) constructed a prototype train: TRY-Z, which consisted of carbodies of three different type structures. A front car was made of thin-wall aluminum hollow extrusions welded continuously. The other front car had a monocoque sandwich structure in which aluminum honeycomb and plates were brazed. The middle car had a skeleton body structure of aluminum alloy joined by a combined method of spot welding and adhesion by a heat setting adhesive. This method is called “weldbonding.” 4. Body Structure for Maglev Trains Japan Air Lines has developed a type of normal conduction maglev trains called High Speed Surface Transport (HSST), in which several types of sandwich panels were used as shown in Table 4 (Japan Alminium Association 1990). 5. CFRP Bogie Frame A CFRP bogie frame shown in Fig. 17 was experimentally made by Railway Technical Research Institute. The side beams of the frame have four webs made by dry hand lay-up of carbon prepregs (Wako et al. 1992). The vertical webs had

47

Adhesive Bonding for Railway Application

1383

Fig. 16 Joining method of a carbon fiber reinforced plastic (CFRP) roof block to the main structure (Matsuoka and Aso 1992; Matsuoka 1992) Table 4 Adhesively bonded sandwich structures used in prototype HSST car (Japan Alminium Association 1990) Prototype car HSST-03

HSST-04 HSST-05

Position Floor board Ceiling board Side ceiling board Side sliding door Floor board Floor board

Composition of sandwich structure CFRP/Nomex ® honeycomb/CFRP GFRP/Paper honeycomb/GFRP GFRP/Al honeycomb/GFRP Al/Al honeycomb/Al Stainless steel/plywood/stainless steel Al/Al honeycomb/Al

an U shape for a large bonding area. The horizontal webs of flat shape were bonded on the bonding area of the vertical webs and to the honeycomb cores with a heat setting adhesive (121  C). The cross beams of the frame were CFRP pipes made by filament winding. The tip of the cross beam were inserted to the side beam and joined with a room temperature setting adhesive.

47.4

Application of Structural Adhesive for Rolling Stock over the World

47.4.1 Britain (Seeds 1984) Present Use A large amount of adhesives is used in the construction of railcars, including solventbased, anaerobic, and cyanoacrylate products, but these do not form structural adhesives. The uses of structural adhesives have been divided into three categories.

1384

Y. Suzuki Cross beam (filament winding) Side beam (bonded with adhesive cured at 121°C) Taper sleeve

Cross section Inserted and bonded with room temperature setting adhesive Honeybomb core

Fig. 17 Carbon fiber reinforced plastic (CFRP) bogie frame fabricated with adhesives (Wako et al. 1992)

1. Simple Fastening In this group, the normal operating stresses are quite low although high loads may occasionally have to be borne. An example of this type is the use of a two-part epoxy adhesive to bond steel kicking-strips to the base of the aluminum sliding door. 2. Sandwich Panel Construction This represents a major use of structural adhesives in the industry, and a wide range of internal trim panels are produced in this way by bonding melamine to plywood or aluminum. 3. Structural Bonding Two particular examples are worth noting. The first is the bonding of steel stiffening strips with acrylic adhesive to the rear surface of grass fiber reinforced plastic (GFRP) window-surround panels to prevent flexing and to provide additional stiffness under impact. The second example is the bonding of aluminum skin and honeycomb structure to make the sliding doors which are situated on the outside of the latest suburban coaches (passenger railcars without power unit).

Development Work 1. Aluminum Ceiling Panel Figure 18 shows the design of the aluminum ceiling panel intended for use on the latest inter-city coaches (Seeds 1984). The requirement is to fasten the top-hat stiffeners to the upper side of the panel without damaging the epoxy-paint finish on the underside. Two-part separate application acrylic adhesive met the requirements of the joint and a light abrasion plus a solvent wipe gave adequate preparation.

47

Adhesive Bonding for Railway Application

1385

Fig. 18 Aluminum ceiling panel bonded adhesively (Seeds 1984)

Fig. 19 Bodyside (filter) door components of Class 58 locomotive (Seeds 1984)

2. Class 58 Locomotive Bodyside Door The locomotive door is presently made by spot-welding a number of pressed components, as shown in Fig. 19 (Seeds 1984). The use of structural adhesive bonding has several important advantages. The surface finish can be greatly improved by eliminating the spot-weld blemishes; dimension variation in the

1386

Y. Suzuki

components can be absorbed by using a gap-filling adhesive, and therefore a smoother finish could be obtained; and finally, adhesive bonding can provide a faster process for a capital outlay of less than that for spot-welding. Under impact conditions, joints using this steel bonded with conventional toughened epoxy and acrylic adhesives fail preferentially at the zinc/steel interface, and this leads to an unacceptable impact performance. By testing a wide range of adhesives, it was found that the shear modulus had to be less than 100 MPa to enable the adhesive in the joint to deform sufficiently under impact to prevent the over-stressing of the interface. As an environmental durability test, measurement of lap-shear strength after exposure to 95% RH at 40  C was carried out. The only adhesive which met all of the requirements was a two-part acrylic adhesive. That adhesive also had the desired cure characteristics (work life 5 min, de-jig time 30 min). 3. Coach Bodyside The present method for constructing the coach bodyside on suburban railcars uses a spot welding process. The panel consists of 2-mm IZ steel ® (zinc-coated steel) skin and top-hat section stiffeners, and Fig. 20 shows the panel components arranged horizontally on the spot-welding jig (Seeds 1984). The advantages of doing this would be two-fold: firstly, the process could be made several times faster for a capital outlay lower than that for the present spot-welding equipment and secondly the resulting surface finish would be considerably better than the spot-welded panels, thereby saving most of the time presently taken to prepare the bodyside for painting. Adhesive joints should be sufficiently strong and stiff for the structure to operate in 20–40 years. The determination of strength and stiffness requirements has been

Fig. 20 The arrangement of the bodyside module components of suburban coach on the assembly jig (Seeds 1984)

47

Adhesive Bonding for Railway Application

1387

tackled in two distinct ways: first, by the use of structural models of vehicles (e.g., Finite Element Method) and second, by direct comparison with spot-welds at both the joint and structural level. Experiments indicate that in shear (the principal stress in the bodyside panel) the adhesive joints can be twice as strong as the equivalent spot-weld joint, but joint stiffness has been shown to be approximately similar. Using toughened single-part epoxy adhesives which can be cured at 120  C, prototype equipment, able to produce full-size bonded panels in a 2- to 3-h cycle time, will permit at least a twofold increase in production rate compared with the spot-welding process. Several full-scale bonded panels have been prepared, and the results have proved successful in demonstrating the capabilities of structural adhesives. Assembly and service trials will be conducted within the next 15 months.

47.4.2 France Société Nationale des Chemins de fer Français (SNCF) made a full-scale prototype composite structure of Train à Grande Vitesse (TGV) double-decker trains whose maximum speed is 350 km/h (Cléon 1994). It comprises sandwich structures of glass fiber and carbon fiber reinforced epoxy resin skins and honeycomb or foam cores bonded adhesively. A half tube of the structure was integrally formed by a vacuum suction method at 120  C. A pair of the half tubes was joined together along the longitudinal midlines at the top and bottom using adhesives, as shown in Fig. 21. The weight saving of the structure led to noise reduction, dynamic stress decrease in the lower structure, and energy saving due to the decrease of its distance for acceleration and reduction of speed.

Fig. 21 Fabrication of Train à Grande Vitesse (TGV) double-decker prototype consisting of composite materials (Cléon 1994)

1388

Y. Suzuki

47.4.3 Germany German high speed train: Intercity-Express (ICE) has viewing windows and blind windows, whose cross sections are shown in Fig. 22 (Rahn et al. 1991). Since these windows are bonded to the body panel making the outer surfaces of the body smooth, the aerodynamic drag of the train is minimized. The window structure is also effective to improve the airtightness and water vapor impermeability of the car. The window glasses are bonded adhesively into aluminum frames. This method is called “direct glazing.” The complete window units are screwed into the coach body, and the joint between adjacent windows and between window and coach body is plugged with an elastic filling material. This principle has already been applied to the center trailers of the ICE with very good results. The windows were subjected to a long series of fatigue tests in various environmental conditions and passed comfortably. Such flat surfaces are very effective to reduce the running noise in the ICE train because almost all the noise is only due to aerodynamic noise at speed of 300 km/h.

Fig. 22 Cross section of Intercity-Express (ICE) trailer car windows (Rahn et al. 1991)

47

Adhesive Bonding for Railway Application

1389

47.4.4 Switzerland In Switzerland, Schindler Waggon AG made a 1:2.5 scaled prototype of train cars shown in Fig. 23 (Romagna et al. 1997). The structure was fabricated by the wet filament winding method of knitted multi-axial glass fabric tape onto rigid foam cores supported by ring-shaped frames that were made by the filament winding method too. Local reinforcements, heat insulating material, wiring and air ducts, frames for door and window openings, and in situ wound ring frame were integrated into the body structure, which had enough strength, stiffness, and surface flatness. Therefore, the structure was advantageous in terms of costs and productivities.

47.4.5 Indonesia Prototype railcars consisting of multimaterials were fabricated in the Jabotabek project. For instance, the frames were mild steel, the roof construction were painted mild steel sheets, the front end construction was FRP, and the rear end construction and the side construction were stainless steel. The side constructions were bonded to the frames with a two-part room temperature setting epoxy adhesive using positioning jigs. The cars have been in service for already 4 years.

Roof canal

Window frame Exterior laminate sheet Wall canal

Exterior foam core Ring frame Intermediate laminate sheet Interior foam core

Ring frame

Interior laminate sheet

Fig. 23 Novel construction concept for the filament wound, highly integrated carriage body of passenger coaches (Romagna et al. 1996)

1390

47.5

Y. Suzuki

Conclusion

To meet the demands of faster speed, weight saving, noise reduction, cost reducing in terms of production and maintenance, joining methods in train structures are increasingly diversified as well as the materials. Adhesion is a joining method having advantages such as low stress concentration, high stiffness, and large damping, so that the use of adhesion will expand in train applications as research proceeds.

References Abiko Y, Kobayashi G (1998) Rolling stock & technology 1:22 (in Japanese) Central Japan Railway (2009a) Central Japan Railway Company data book 2009, pp 24–25 Central Japan Railway (2009b) Central Japan Railway Company enviromental report 2009, p 6 Cléon LM (1994) Proceedings of the World Congress on railway research 2:673 (in French) Japan Alminium Association (1990) Reaearch report of alminium alloy and lightweight of rolling stock -Research on lightweight of rolling stock as high speed transportation, pp 42–53 (in Japanese) Japan Association of Rolling Stock (2005) Industries standard JRIS R 0206-1; rolling stockguidelines for rigging design-Part 1:underfloor installations (in Japanese) Japan Association of Rolling Stock (2006) Industries standard JRIS R 0144:rolling stock-designing guidelines for sealing material (in Japanese) Kamiura M, Sunaga M et al (2000) Tetsudo Kougaku, Morikita Shuppan. p 213 (in Japanese) Matsuoka S (1992) Denkisha no kagaku, 45(12):30–34 (in Japanese) Matsuoka S, Aso K (1992) Tokyu car technical review 42:6–19 (in Japanese) Matsuzawa H (1986) Introduction for passenger car engineering, Railway System Research, p 106 (in Japanese) Miyamoto M (1998) Kougyouzairyou, 46(3):108 (in Japanese) Nishimura T (2014), Tetsudosharyokogyo (Journal of Japan Association of Rolling Stock Industries), 470:21–24 Rahn T, Hochbruck H et al (1991) ICE Hightech on rails. Hestra-Verlag, p 116 Romagna J, Anderegg K et al (1996) Proceedings of 17th International European Chapter of SAMPE Conference, 57 Sasaki K (1995) Rolling stock & technology, 8:13 (in Japanese) Seeds A (1984) Int J Adhes Adhes 4:17 Sugano T (1992) Nikkei New Materials, July 13:8 (in Japanese) Suzuki Y (2006) Fundamental and application of structural adhesive bonding edited by Miyairi H, CMC, p 343 (in Japanese) Suzuki Y (2007) “Adhesion handbook ver.4” edited by The Adhesion Society of Japan, Nikkan Kogyo Sinbun, p 1011 (in Japanese) Suzuki Y (2009) A textbook of adhesion technology for men aiming a professional edited by The Adhesion Society of Japan, Nikkan Kogyo Sinbun, pp 179–187 (in Japanese) Suzuki Y, Sato K (1993) RRR, 6:9 (in Japanese) Wako K, Umezawa Y, Tokuda A (1992) RTRI report, 6(10):19 (in Japanese)

Marine Industry

48

Peter Davies

Contents 48.1 48.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Specific Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.2.1 Certification . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.2.2 Manufacturing Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.2.3 The Marine Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.3 Surface Preparation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.3.1 Composite Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.3.2 Metallic Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.4 Mechanical Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.4.1 Type of Adhesive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.4.2 Substrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.4.3 Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.5 Durability in a Marine Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.5.1 Degradation Mechanisms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.5.2 Predicting Marine Aging Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.5.3 Testing to Simulate the Marine Environment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.6 Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.7 Repair and Recycling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48.8 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1392 1392 1392 1395 1404 1404 1405 1406 1407 1407 1409 1409 1410 1410 1411 1412 1413 1413 1414 1415

Abstract

This chapter describes the use of adhesive bonding to assemble structures in the marine industry. The marine environment is extremely aggressive, and this has resulted in the widespread use of fiber-reinforced composite materials. Adhesive

P. Davies (*) Materials and Structures Group, Marine Structures Laboratory, Brest Centre, IFREMER, (French Ocean Research Institute), Plouzané, France e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_48

1391

1392

P. Davies

bonding is a lightweight and corrosion-resistant means of joining these materials. The main emphasis of this chapter will therefore be on the assembly of composites, though some examples of metal bonding will also be discussed. Three industrial applications are used to illustrate the use of adhesive bonding: small pleasure boats, high performance racing yachts, and bonded structures in the offshore industry. Each has specific design requirements, and there is no single marine adhesive suitable for all structures, but the requirement for long-term durability in a seawater environment is common to all these applications.

48.1

Introduction

The marine industry encompasses a wide range of structures, from small boats to offshore oil platforms. It covers small boatyards, large production line-based boat assembly plants, naval shipyards, and a host of subcontracting companies providing services from design through to offshore platform repair. Many of these use large quantities of adhesives and sealants, and it is not possible to provide a description of all of them. In this chapter, the role of adhesive bonding in the marine industry will be described through three examples of application areas: small pleasure boats, high performance racing yachts, and bonded structures for the offshore oil and gas industry. The materials involved are mostly fiber-reinforced composites, as these are the materials which are most frequently assembled by adhesive bonding in this industry. They range from low cost glass mat or woven roving-reinforced polyester, widely used in pleasure boats, to high performance carbon fiber-reinforced epoxy and honeycomb sandwich for racing yachts.

48.2

Specific Requirements

In this section, we will first present the particularities of adhesive bonding for this industry. There are a number of specific marine requirements, resulting from regulatory requirements, the way marine structures are assembled, and the environment they meet in service. Thus, when considering adhesive bonding of structures for marine applications, in comparison with other industrial applications, the following points need particular attention: • Certification • Manufacturing conditions • The marine environment

48.2.1 Certification All marine structures require certification by classification societies (Bureau Veritas, Lloyds, Det Norske Veritas (DNV), Germanischer Lloyds, American Bureau of

48

Marine Industry

1393

Shipping (ABS), etc.), so in order to obtain approval, it is necessary to take account of their requirements from the design stage. In addition to extensive recommendations, developed to ensure structural integrity, several of these societies also have requirements for the adhesive materials to be used in assembly. These are intended to ensure conformity with specified requirements, by subjecting test samples to physical, chemical, environmental, or operational stresses in a recognized laboratory. Many adhesive suppliers indicate certification authority approval for marine use in their product data sheets. Table 1 shows typical requirements. It is important to note that these tests are performed on adhesively bonded joints, usually in the form of lap shear specimens, and not bulk adhesives. The approval is therefore specified for particular substrates and includes their surface preparation and the joint manufacturing procedure. Taking the DNV guidelines (DNV 2009) as an example, the approval procedure is based on the requirements of the DNV guidelines for classification of high speed, light craft, and naval surface craft (DNV 1991). The adhesive supplier indicates a “manufacturer’s specified minimum value” (msmv) for lap shear strength and cleavage strength. Standard tests are then performed to check these and obtain a reference value. The marine durability of approved adhesives is then taken into account by stipulating a percentage of mechanical strength retention after weathering under load. Specimens are weathered under different conditions, as shown in Table 2, and tests are performed on the aged specimens in order to determine property retention. Here, the weathering condition involves loading according to ASTM D2919 (test method for determining durability of adhesive joints stressed in shear by tensile loading) to a constant stress and weathering according to ASTM D1183 (resistance of adhesives to cyclic laboratory aging conditions). The latter includes a cycle for exterior marine exposure:

Table 1 Typical requirements for approval of marine adhesives Tests type Lap shear testing

Cleavage peel testing

Fatigue testing Immersion tests Glass transition temperature

Condition Control (room temperature) Elevated temperatures Cold temperatures Control (room temperature) Elevated temperatures Cold temperatures ASTM D3166 Water Fuel ASTM E1356 (calorimetry)

Requirement Manufacturer’s values achieved

Manufacturer’s value achieved

Target value achieved (e.g., 107 cycles) Minimum % retention achieved Above minimum value (e.g., 70
C for military ships)

1394

P. Davies

Table 2 Marine durability conditions for adhesive-type approval (DNV 2009) Test Lap shear ASTM D1002 or D3163 Cleavage fracture ASTM D3433

– – – –

Aging condition Loaded to msmv and weathered (ASTM D1183) 28 days 56 days Loaded to msmv and weathered (ASTM D1183) 28 days 56 days

Requirement Mean >90% reference Mean-2 standard deviations >95% 28-day value Mean >90% reference Mean >95% 28-day value

48 h at 71
C and 3 g/m2), such as, for example, acrylates or some epoxy resins. As a rule, however, these surface contaminations have to be removed prior to bonding (Henning and Moeller 2011). Special attention must be paid to the type of steel used. Stainless steels are passivated by their chromium content in such a way that the reactivity toward corrosive media is reduced. Due to this passivation, the strength of intermolecular bonds between the adhesive and the steel necessary to achieve a good adhesion may also be reduced (Habenicht 2009). However, this statement has not been demonstrated in general (Geiß et al. 2012), and the bond strength needs to be evaluated for any combination of adhesive and steel grades. The bonding to galvanized steels can be problematic in that the zinc layer has a different adhesion to the steel substrate, depending on the type of deposition

1442

S. Böhm et al.

(hot-dip or electrolytic galvanizing). Furthermore, the adhesive strength is influenced by the different orientations of larger zinc crystallites, as a heterogeneous stress distribution is thereby generated within the adhesive layer (Habenicht 2009). Today, in steel construction, polyesters, vinylesters, acrylates, epoxy resins, polyurethanes, or hybrid systems are employed (Dilger 2011). Especially the polymerizing systems (acrylates, esters) prove beneficial, since these readily cure at low temperatures and are more insensitive to mixing errors (Feldmann et al. 2006). Structural thermosetting adhesives, which have been successfully used in car body shops for several years like one-component epoxy-based systems, are not suitable for many steel structure applications because of the required high curing temperatures (between 175  C and 190  C) and the resulting difficulties on construction site.

49.5.3 Applications As mentioned above, gluing in steel construction (as in the construction industry in general) only has a very limited range of use, even though in the past initial attempts have been made to increase its application. Therefore, the examples listed here are rather of an academic nature or part of basic research and are not routinely used in steel construction.

Bridge Building In 1955, a bridge based on hybrid bonding technology (adhesive bonding backed up with screws) was raised over a side channel of the river Lippe in Marl, Germany (Brockmann and Neeb 2001). In this case, rods were adhesively bonded with the lower and upper belt with a polyester-phenolic resin via a flange connection. Besides other examples (Brockmann and Neeb 2001; Piekarczyk and Grec 2012), further developments at the time were scarce due to the limited properties of the adhesives and experiences compared to contemporary joining techniques. However, even today, gluing in bridge construction remains a niche application. This is mainly due to the progressive development of welding technology with regard to the realization of complex structures. In addition, statements about the long-term stability of adhesive bonds under environmental influences can only be made inadequately. A common method for strengthening decks in ship construction is the application of sandwich structures where an elastomeric adhesive fills the gap between two steel sheets. These steel-polymer-steel structures (SPS) possess a promising torsion rigidity. Furthermore, the elastic core is able to relieve stress peaks. Feldmann et al. were investigating such laminates for the post-strengthening of road bridges (Feldmann et al. 2006). They achieved a reduced cover plate bending and thus a higher load capacity.

49

Civil Construction

1443

Metal Framework Facades Recent glass facades demand a highly stiff supporting structure with a maximum bearing capacity which effectively takes up the occurring loads (e.g., wind loads) but which at the same time makes light-flooded architectural design and construction possible (Pasternak and Meinz 2007). Mullion and transom facades which use profiles with a rectangular hollow cross section are a suitable construction variant for this kind of task. However, an additional strengthening in terms of an increased load-bearing capacity and improved stiffness can be achieved by glued-in steel sheets. The effectiveness of this method could be evaluated by means of a fourpoint bending test at the Brandenburg Technical University (Pasternak and Meinz 2007). Especially for industrial halls, facades made of trapezoidal profiles are employed which in general are attached to the support structure by means of nails, rivets, or screws. However, the fastening methods possess several disadvantages like the possibility of rust formation or the local occurrence of dents (Pasternak and Meinz 2007). Thus, Pasternak and Meinz developed a fastening method that makes use of a hook profile adhesively bonded to the trapezoidal sheets. They determined the optimal construction parameters in terms of normal stresses in dependence of hook length or type of the adhesive system used. Hybrid Bonding As already mentioned, adhesives are often not able to compensate larger bonding gaps in the range of several millimeters to centimeters. However, these gaps are unavoidable due to high tolerances in building construction. Scientists at the Karlsruhe Institute of Technology have developed a method that can solve this discrepancy (KIT 2016a, b). For this purpose, inorganic granules are adhered to the steel surface with an organic structural adhesive. The steel surface is simultaneously sealed by the adhesive and protected against corrosion. The components thus treated are subsequently connected by means of a grouting mortar.

49.6

Conclusions and Outlook

In structural engineering, two connection techniques are currently employed. On the one hand, mechanical fixing elements are used. These transfer forces only locally and, in addition, may partially damage the adherends. On the other hand, welding is a widely spread joining technology that results in stiff and resilient structures. However, welding is only suitable for certain material combinations. Thus, many novel design concepts cannot be realized. Adhesive technology leads to a high material utilization and freedom of design and is therefore nowadays already established in some sectors of the construction industry. This includes not only maintenance and repair but also the realization of novel building structures.

1444

S. Böhm et al.

Since concrete itself can be referred to as an (inorganic) adhesive, in concrete construction all processes are basically bonding processes. In addition, the use of organic systems can be beneficial for further applications, e.g., for anchor systems and reinforcement elements. The glass industry best illustrates the high potential of adhesive bonds. SSG facades play an increasing role due to their excellent design possibilities. The development of suitable silicones and acrylic systems as well as special types of glass (prestressed glasses, laminated glasses) are likely to lead to further construction projects that will benefit from the advantages of glued joints. In timber construction, the use of adhesives is also quite common. By the use of adhesion technology, new material combinations could be achieved, which feature many advantages. In addition adhesives are employed for structural bonding as well as for reinforcement applications. In steel construction, adhesive bonding is less common. Nevertheless, several research projects demonstrate its benefits even in this sector. This results not least from the vast diversity of materials and the resultant difficulty of joining by wellestablished techniques. Even if adhesive bonding is already established in many areas of construction, its potential is not yet fully exploited. The reasons for this are the relatively low level of trust in this technology, the lack of knowledge in the use of adhesives, as well as the limited process control. Future developments should address these current deficits in order to successfully expand the application of adhesive technology in the civil construction sector.

References 3M (2013) 3M stamark pavement marking tape, series 380. 3M Traffic Safety and Security Division, St. Paul. http://multimedia.3m.com/mws/media/849277O/3m-stamark-pavementmarking-tape-series-380-flyer.pdf?fn=380.pdf. Accessed 16 Nov 2016 3M (2016) Verlegeanleitung für 3M Stamark Dauermarkierungsfolien auf Betondecken. 3M Deutschland GmbH, Neuss. http://multimedia.3m.com/mws/media/345062O/verlegeanleitung-beton-plusprogramm.pdf?fn=PM_%2028_ESD_IB_DE.pdf. Accessed 16 Nov 2016 Bergmeister K (2001) Kleben im Betonbau – Theoretische Grundlagen und Bemessungsvorschläge. Beton Stahlbetonbau 96(10):625–633 Berset T, Schwegler G, Trausch L (2002) Verstärkung einer Autobahnbrücke mit vorgespannten CFK-Lamellen tec21 128(22):27–29 Biolzi L, Cattaneo S, Rosati G (2010) Progressive damage and fracture of laminated glass beams. Constr Build Mater 24(4):577–584 Brockmann W, Neeb T (2001) Stand der Klebtechnik im Stahlbau. Stahlbau 70(2):106–115 Brunner R (1996) Kleben im Baubereich. Seminar Modernes Kleben, Trends aus Industrie und Forschung, EMPA Dübendorf, 1996. EMPA, Dübendorf, pp 70–79 Caroll B (2014) Selecting and specifying the most effective anchor type. Eng J Dublin. http://www. engineersjournal.ie/2014/10/28/selecting-specifying-effective-anchor-type/. Accessed 14 Nov 2016 Corning D (2011) Structural Glazing Handbuch. Dow Corning GmbH, Wiesbaden. http://www. dowcorning.com/content/publishedlit/62-0979-03.pdf. Accessed 22 Nov 2016

49

Civil Construction

1445

Corradi M, Grazini A, Borri A (2007) Confinement of brick masonry columns with CFRP materials. Compos Sci Technol 67(9):1772–1783 Custódio J, Broughton J et al (2009) A review of factors influencing the durability of structural bonded timber joints. Int J Adhes Adhes 29(2):173–185 Delo (2016) So schön kann das Kleben sein. adhäsion KLEBEN & DICHTEN 60(10):51–52 Dilger K (2011) Kleben. In: Henning F, Moeller E (eds) Handbuch Leichtbau – Methoden, Werkstoffe, Fertigung, 1st edn. Carl Hanser, München Dutta PK, Mosallam A, Vargas J, Mikhael D (2001) A fast method for field evaluation of concrete/ FRP adhesive bonding in extreme temperatures. In: Repecka L, Saremi FF (eds) SAMPE symposium and exhibition. Taylor & Francis, Long Beach, pp 451–462 Feldmann M, Völling B, Geßler A et al (2006) Kleben im Stahlbau. Stahlbau 75(10):834–846 Ferrier E, Lagarde G, Hamelin P (2001) Concrete beams reinforced by fibre-reinforced plastics: the effect of temperature on the adhesive layer. Compos Sci Technol 61:425–431 Gao B, Kim J, Leung CKY (2004) Experimental study on RC beams with FRP strips bonded with rubber modified resins. Compos Sci Technol 64(16):2557–2564 Geiß PL (2006) Klebtechnik übernimmt tragende Rolle im Bauwesen. adhäsion KLEBEN & DICHTEN 50(1/2):14–18 Geiß PL (2008) Geklebte Bauwerke. Blickpunkt Kleben 6 Geiß PL, Fritzsche C, Kleiner F, Peschka M, Rauscher M, Schmale HC, Vogt, D, Zanotti A, Weber C, Boldt F, Wibbeke, M (2012) Merkblatt 382 – Kleben von Stahl und Edelstahl Rostfrei. Stahl-Informations-Zentrum, Informationsstelle Edelstahl Rostfrei Giesecke A (2007) Strukturverglasung – Fusion von Form und Funktion. adhäsion KLEBEN & DICHTEN 51(7/8):24–27 Götz B, Böhme C, Borgmeier R, Burger K, Grote L, Hettich U, Modersohn WJP, Peisl C, Schmitt P, Schneider S, Wendt A, Bisping H, Heinicke F (2010) Technisches Merkblatt – Verankerungen mit Verbunddübelsystemen. Bundesverband Bausysteme e.V, Koblenz. http://bv-bausysteme. de/tl_files/bv-bausysteme/downloads/FV%20Duebel-%20und%20Befestigungstechnik/ duebel_merkblatt_verbundsysteme_0810_36s_web.pdf. Accessed 2 Nov 2016 Grimm R (2013) Durchblick: Welche sind die wichtigsten Fliesenkleber? Verlagshaus Wohlfahrt GmbH, Duisburg. http://www.baustoffwissen.de/wissen-baustoffe/baustoffknowhow/grundlagen/ bauchemie/durchblick-welche-sind-die-wichtigsten-fliesenkleber/. Accessed 5 Nov 2016 Grunewald D (2012) Verstärkung von Betonbauteilen mit geklebter Bewehrung – Auswirkung auf die Baupraxis. Beton Stahlbetonbau 107:A29–A30 Habenicht G (2009) Kleben – Grundlagen, Technologien, Anwendungen. Springer, Berlin Haldimann M, Luible A, Overend M (2008) Structural use of glass – structural engineering documents 10. International Ass. for Bridge and Structural Engineering, Zürich Headley M (2014) EVA finds popularity among decorative fabricators. ME USGlass 49(4). http:// industry.glass.com/USGlass/2014/April/NewsAnalysis_Trends.htm. Accessed 13 Nov 2016 Hollaway LC (2008) Advanced fibre polymer composite structural systems used in bridge engineering. ICE Man Bridge Eng. https://doi.org/10.1680/mobe.34525.0503 Henning F, Moeller E (2011) Handbuch Leichtbau – Methoden, Werkstoffe, Fertigung. Hanser, München Jalham IS, Alsaed O (2011) The effect of glass plate thickness and thickness of the bonding interlayer on the mechanical behavior of laminated glass. New J Glass Ceram 1:40–48 Karbhari VM, Engineer M, Eckel DA II (1997) On the durability of composite rehabilitation schemes for concrete: use of a peel test. J Mater Sci 32:147–156 Keller T, Gürtler H (2006) Design of hybrid bridge girders with adhesively bonded and compositely acting FRP deck. Compos Struct 74:202–212 Keller U, Mortelmans, H (1999) Adhesion in laminated safety glass – what makes it work? Glass processing days. 13–16 June, Tampere, pp 353–356 KIT (2016a) Hybride Klebverbindung. Karlsruhe Institute of Technology, Karlsruhe. https://www. kit-technology.de/de/technologieangebote/details/591/. Accessed 7 Nov 2016

1446

S. Böhm et al.

KIT (2016b) Connection between two joining partners and method for establishing said connection. WO 2016/091377 A1 Kleist A (1998) Verstärkung von Stahlbeton mit CFK-Lamellen. Tiefbau Ingenieurbau Straßenbau 40:35–41 Klosowski J, Wolf AT (2016) Sealants in construction, 2nd edn. CRC Press, Boca Raton Krelaus R, Freisinger S, Schmidt M (2008) Adhesive bonding of structural members at the Gaertnerplatz Bridge in Kassel. In: Fehling E, Schmidt M, Stürwald S (eds) Ultra high performance concrete (UHPC). Kassel University Press, Kassel, pp 597–604 Levy SM (2012) Construction calculations manual. Elsevier, Amsterdam Louter C (2007) Adhesively bonded reinforced glass beams. Heron 52(1/2):31–58 Mailvaganam NP (1997) Construction technology update no. 11 – effective use of bonding agents. Institute for Research in Construction, National Research Council of Canada McKenna JK, Erki MA (1994) Strengthening of reinforced concrete flexural members using externally applied steel plates and fibre composite sheets – a survey. Can J Civ Eng 21(1):16–24 Meier U (1994) Rehabilitation of structures with the CFRP sheet bonding technique – advancing with composites. In: Crivelli-Visconti I (ed) Advancing with composites, Materials and technologies, vol 1. Woodhead Publishing, Cambridge Meier U, Münger F (1995) Repair of structures through external bonding of thin carbon fiber sheets. IABSE Rep 73(1/2):1167–1172 Mihala R (2008) Bauwerkverstärkung mit eingeschlitzten CFK-Lamellen in Beton. Zement + Beton 1:4–5 Oehlers DJ (2001) Development of design rules for retrofitting by adhesive bonding or bolting either FRP or steel plates to RC beams or slabs in bridges and buildings. Composites A 32 (9):1345–1355 Pasternak H, Meinz J (2006) Versuche zu geklebten Verstärkungen im Stahlhochbau. Bauingenieur 81:212–217 Pasternak H, Meinz J (2007) Kleben im Stahlbau – zwei Beispiele aus dem Fassadenbau. Bauingenieur 82:115–124 Piekarczyk M, Grec R (2012) Application of adhesive bonding in steel and aluminium structures. Arch Civ Eng 58(3):309–329 Porteous J, Kermani A (2012) Structural timber design to Eurocode 5. Wiley-Blackwell, Oxford Rizzo P, Spada A, Degala S et al (2010) Acoustic emission monitoring of chemically bonded anchors. J Nondestruct Eval 29(1):49–61 Rug W, Mönck W (2015) Holzbau – Bemessung und Konstruktion. Beuth, Berlin Schober KU, Harte AM, Kliner R, Jockwer R, Xu Q, Chen JF (2015) FRP reinforcement of timber structures. In: Harte A, Rietsch P (eds) Reinforcement of timber structures – a state-of-the-art report. Shaker, Aachen Serrano E (2000) Adhesive joints in timber engineering – modelling and testing of fracture properties. KFS, Lund Shayan A, Sanjayan JG, Swarmy N (2003) Evaluation of an epoxy-based steel plating system for strengthening bridge decks. Int J Mat Prod Tech 19(3/4):284–294 Stahlton (2011) Verstärkung von Tragkonstruktionen. Stahlton AG, Hinwil. http://www.stahltonbautechnik.ch/images/stories/produkte/verstaerkung/Bewehrungen005.pdf. Accessed 14 Nov 2016 Steiger R, Serrano E, Stepinac M, Rajčić V, O’Neill C, McPolin D, Widmann R (2015) Reinforcement with glued-in rods. In: Harte A, Rietsch P (eds) Reinforcement of timber structures – a state-of-the-art report. Shaker, Aachen StoCretec (2016) Tragwerksverstärkung, Höhere Belastbarkeit für Stahlbetonbauwerke. StoCretec GmbH, Kriftel. http://stocretec.de/media/documents/download_broschuere/0_bi_os/140080_ DE-Broschuere_Betoninstandsetzung-Tragwerksverstaerkung_2013htm.pdf. Accessed 21 Nov 2016 Suter R, Jungo D (2001) Vorgespannte CFK-Lamellen zur Verstärkung von Bauwerken. Beton Stahlbetonbau 96(5):350–358

49

Civil Construction

1447

Triantafillou T, Deskovic N (1992) Prestressed FRP sheets as external reinforcement of wood members. J Struct Eng 118(5):1270–1284 Wagner A (2009) Alterungsbeständigkeit strukturell geklebter Stahl- und Glasverbunde. Technische Universität Kaiserslautern Weller B, Schadow T (2007) Forschung und Entwicklung im Konstruktiven Glasbau. Stahlbau 76 (3):162–166 White C, Tan K, Wolf A, Carbary L (2010) Advances in structural silicone adhesives. In: Dillard DA (ed) Advances in structural adhesive bonding, 1st edn. CRC Press, Boca Raton Wörner JD, Freitag C (2012) Glasbau. In: Zilch K, Diederichs CJ, Katzenbach R, Beckmann KJ (eds) Handbuch für Bauingenieure – Technik, Organisation und Wirtschaftlichkeit, 2nd edn. Springer, Berlin Wu Z, Yuan H, Kojima Y et al (2005) Experimental and analytical studies on peeling and spalling resistance of unidirectional FRP sheets bonded to concrete. Compos Sci Technol 65 (7/8):1088–1097 Yiin W (2016) Tourists test the safety of 984-foot-high Chinese glass bridge – with sledgehammers. The Washington Post, Washington, DC. https://www.washingtonpost.com/news/morning-mix/ wp/2016/06/28/tourists-test-the-safety-of-a-chinese-glass-bridge-by-trying-to-break-it-withsledgehammers. Accessed 15 Nov 2016 Zilch K (2013) Verstärken von Betonbauwerken mit CFK. 1. DAfStb-Jahrestagung, Deutscher Ausschuss für Stahlbeton e.V., Berlin. http://www.dafstb.de/application/Jahrestagungen/zilch. pdf. Accessed 21 Nov 2016 Zilch K, Mühlbauer C (2007) Zum Fügen von Ultra-Hochfestem Beton (UHPC) durch Verkleben – Grundlegende Untersuchungen. In: Bauen mit innovativen Werkstoffen, VDI-Berichte 1970, VDI-Verlag, Düsseldorf, pp 37–46

Electrical Industry

50

Kwang-Seok Kim, Jong-Woong Kim, and Seung-Boo Jung

Contents 50.1 50.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Classification of Electrically Conductive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.2.1 Isotropic Conductive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.2.2 Anisotropic Conductive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.2.3 Non-conductive Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.3 Conductive Filler Materials for ECAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.3.1 Hybrid Metallic Fillers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.3.2 Carbon-Based Nanomaterials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.3.3 Hybrid Nonmetallic Fillers (Conductive Polymers and PEDOT:PSS) . . . . . . . 50.4 Specific Requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.4.1 Electrical Conductivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.4.2 Thermomechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.5 Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.5.1 Test Vehicle Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.5.2 Storage Test in Harsh Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.5.3 Mechanical Reliability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1450 1453 1454 1456 1459 1459 1460 1461 1463 1464 1464 1469 1471 1472 1474 1478 1479 1480

K.-S. Kim Carbon & Light Materials Application Group, Korea Institute of Industrial Technology, Jeonju, Jeollabuk-do, Republic of Korea e-mail: [email protected] J.-W. Kim School of Advanced Materials Engineering, Chonbuk National University, Jeonju, Jeollabuk-do, Republic of Korea e-mail: [email protected] S.-B. Jung (*) School of Advanced Materials Science & Engineering, Sungkyunkwan University, Suwon, Gyeonggi-do, Republic of Korea e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_50

1449

1450

K.-S. Kim et al.

Abstract

Electrically conductive adhesives (ECAs) are used for the joining of electrical and electronic components in the electrical industry. ECAs are composites consisting of a polymeric matrix and electrically conductive fillers. The mechanical properties are provided by the polymeric matrix, while the electrical conductivity is supplied by the conductive fillers. Typically, there are three types of conductive adhesives, viz., isotropic, anisotropic, and non-conductive adhesives. These adhesives are classified according to their conductivity, which is controlled by the conductive filler content in the polymer matrix. For their application in the electrical industry, there are two main specific requirements for these adhesives, i.e., electrical conductivity and thermomechanical properties. In this chapter, the mechanism underlying the electrical conduction in adhesive joints is derived, while the thermal and mechanical parameters that should be measured are introduced. Some recent advances related to the improvement of ECAs are also addressed. In terms of the evaluation of the reliability of adhesives in electronics, the basic test procedures, including several specific test methods and analysis techniques, are explained. Recommendations are also given to select a suitable test method.

50.1

Introduction

Conventionally, the joining of components in the electrical and electronic industries was mainly achieved by employing metallic materials, such as low melting point Sn-Pb solders. Sn-Pb solders are generally used to join similar or dissimilar metallic materials, such as Cu to Cu or Cu to Ni, etc., which are used as interconnect materials in electrical/electronic systems. When using Sn-Pb solders for joining, Sn plays the major role in the joining process, while Pb is included to lower the melting temperature and increase the ductility, thereby enhancing the thermomechanical reliability, preventing whisker formation, and decreasing the thickness of the intermetallic compound layers which could be formed between the solder and other metallic materials. Although Sn-Pb solder alloys provide a low melting point and good mechanical reliability, the toxicity and environmental incompatibility of the lead contained therein are forcing changes in the world market for electronic products which will dramatically reduce and possibly even eliminate their use altogether. As a consequence, the industry is changing its primary interconnect technology from Pb-based solder alloys to Pb-free alloys and electrically conductive adhesives (ECAs) (Liu and Lin 2005; Kim et al. 2005; Yin et al. 2006; Zhang et al. 2006; Plumbridge 2004). Although various Pb-free solders have already been developed, most current commercial Pb-free solders, such as Sn-Ag and Sn-Ag-Cu, have higher melting temperatures than that of Sn-37Pb solder, which raises the soldering temperature and, thus, reduces the integrity, reliability, and functionality of the printed circuit boards, components, and other attachments. On the other hand, ECAs generally have a lower processing temperature, which minimizes the thermal damage of the

50

Electrical Industry

1451

packaging components, as well as many other distinct advantages that solder alloys cannot offer. They are flexible, capable of fine pitch interconnections, environmentally friendly, and cheaper to manufacture (Kim et al. 2008a). The first employment of ECAs in the electrical and electronic industries was realized in the early 1970s for attaching semiconductor chips and other passive devices onto ceramic substrates in the fabrication of hybrid microcircuits. Epoxy technology pioneered the use of isotropic conductive adhesives for die attach applications in 1966. These early adhesives were filled with gold and silver powders and flakes which provided electrical and thermal pathways for the bonded devices (Lau 1995). However, those rare metal fillers must be reduced to allow their application in ECAs from the aspect of cost. Researchers are recently focusing on high aspect ratio and highly conductive fillers such as metallic nanowires and carbon-based nanomaterials instead of excess weight loading of rare metal fillers. Nowadays, ECAs are mainly employed in advanced electronic packaging technologies, such as chips on flex (COFs), chips on glass (COGs), and flip chip attachment on printed circuit boards (PCBs). Figure 1 shows schematic images of these electronic packages. Before proceeding to the discussion on the concept and applications of ECAs, a brief introduction to electronic components and packaging technologies is helpful.

a

b

Chip Chip

LCD Panel

Glass substrate

Flexible substrate

c

Chips PCB

• LCD: Liquid crystal display • PCB: Printed circuit board

Fig. 1 Schematic description of various electronic packages with adhesives: (a) chip on flexible substrate, (b) chip on glass substrate, (c) chip on rigid board

1452

K.-S. Kim et al.

The technical developments realized in the industry in the last two decades of the twentieth century were enabled by the enormous progress made in microelectronic and semiconductor technology, due to the miniaturization of integrated circuits (ICs). Reducing the circuit dimensions reduces the transmission distances and increases the achievable speed of operation, and, in this way, the performance of the chips could be enhanced dramatically. The chips in the circuit communicate with each other through an input/output (I/O) system of interconnects, and these delicate chips and their embedded circuitry are dependent on the package for support and protection. At this time, the major functions of the electronic package are to provide a path for the electrical current that powers the circuits on the chip, distribute the signals onto and off of the silicon chip, or remove the heat generated by the circuit and support and protect the chip from hostile environments. In the microelectronic packaging industry, flip chip technology, which was introduced for ceramic substrates by IBM in 1962, provides the highest packaging density and performance and the lowest packaging profile. Flip chip technology involves the connection of a chip to a substrate with the active face of the chip facing toward the substrate. Generally, flip chip technology can provide a cost-effective packaging method in various applications, such as telecommunication, computers, display products, and automobiles. The interconnection in flip chip packages that was firstly suggested by IBM was achieved by solder bonding. Flip chip interconnects formed by solder bonding are called solder-bumped flip chip interconnects, and they consist of essentially three basic elements. These include the chip, the solder bump, and the substrate. An underfill is often used to improve the reliability of the interconnect system. Figure 2 shows a brief schematic diagram showing the flip chip packaging structure with a ball grid array bonding. Each of these elements and the processes used to assemble them together affect the performance and cost of the interconnect system. Therefore, the performance and cost must be compared on the

Underfill

Solder bumps

Flip chip

BGA solder ball

Substrates • BGA: Ball grid array

Fig. 2 Package structure for flip chip ball grid array

50

Electrical Industry

1453

basis of the interconnect system as a whole, and not merely on the basis of any single element of the interconnect assembly. Eutectic Sn-Pb alloys are widely used as interconnects in solder-bumped flip chip packages. However, as indicated above, there are increasing concerns about the use of Sn-Pb alloy solders, because Pb, a major component in the solder, has long been recognized as a health threat to human beings. With the phasing out of Pb-bearing solders, various ECAs have been identified as environmentally friendly alternatives to Sn-Pb solders in flip chip applications. There are three leading ECA flip chip methodologies: isotropic conductive adhesives (ICAs), anisotropic conductive adhesives (ACAs), and non-conductive adhesives (NCAs). ICAs consist of an insulating adhesive polymer matrix with a high loading of conductive particles. The conduction mechanism is based on the percolation theory of electrical conduction afforded by the physical contact of the conductive particles. ACAs are similar to ICAs, except that they have a lower conductive filler content and thus provide unidirectional conductivity in the vertical or z-direction (perpendicular to the plane of the substrate). This means that the controlled conductive filler loading provides a higher resistance interconnect through the thickness of the adhesive than those of the ICA or solder-bumped interconnects. NCAs, on the other hand, do not consist of conductive particles but rather maintain a pure mechanical contact between the bumps and pads due to the application of a compressive force, and it is this that ensures the electrical connectivity between them. Nowadays, NCAs are being considered as an alternative to ACAs, because they can achieve a finer pitch of input/outputs (I/Os) due to the absence of any dispersion of conductive particles between I/Os and their better electrical performances. However, some issues concerning their performance and long-term reliability have recently been raised, so that a technical review of each material is essential. Figure 3 provides a comparison of the three ECA technologies, and a more detailed description of them will be provided in the next chapter.

50.2

Classification of Electrically Conductive Adhesives

Adhesives are basically organic materials containing a small amount of additive materials such as conductive or dielectric materials. The organic materials in adhesives can be separated into two different types depending on their structure and thermal behavior: thermoplastics and duromers (Aschenbrenner et al. 1997; Habenicht 1990). While thermoplastics are hot melting and can be plasticized and reshaped at temperatures above their glass transition temperature (Tg), duromers are thermosetting and can be modeled during the curing process only (Aschenbrenner et al. 1997). Today, most adhesives are thermosetting ones based on epoxy resins, which provide high heat resistance and excellent adhesion to a variety of substrates. However, they are not thermally reversible and cannot be dissolved in common organic solvents, because they are crossed-linked after bonding. In addition, thermosetting adhesives provide a low interconnection resistance, because they are easy to exclude between the bonding electrodes during the bonding process. There are various kinds of conductive particles that can be used to endow the adhesives with

1454

a

K.-S. Kim et al.

b

Underfill

ICA

ACA Chip

Chip

Pad

Substrate

Isotropic conductive adhesive (ICA)

c

Substrate

Anisotropic conductive adhesive (ACA)

NCA Chip

Substrate

Non-conductive adhesive (NCA)

Fig. 3 Schematic descriptions of three ECA technologies: (a) isotropic conductive adhesive (ICA), (b) anisotropic conductive adhesive (ACA), (c) non-conductive adhesive (NCA)

electrical conductivity. Ag is by far the most common conductive material for ECAs, but most of the recent ACAs employ polymer particles coated with various metals. Polymer particles are more elastic and, therefore, have the potential to compensate for the strains within the adhesive interlayer induced by thermal expansion (Aschenbrenner et al. 1997). Here, we will briefly introduce the various adhesives, starting with ICAs.

50.2.1 Isotropic Conductive Adhesives ICAs are generally fabricated using adhesive and metallic conductive fillers that impart the desired adhesive and electrical properties, respectively. Typically, silver flakes and powders are utilized, because of their high electrical conductivity, chemical stability, and low cost. Processed Ag fillers must have suitable properties in order to fabricate compatible ICAs. For example, several micron-sized Ag flakes with a loading of 25–30% by volume in ICAs can give good conductive and adhesive properties (Markley et al. 1999). The particle size, particle size distribution, and other important characteristics can be controlled by the manufacturer by manipulating the various processing conditions, e.g., the lubricant concentration used during the formation of the neutral Ag particles. Only slight variations in the concentration of the lubricant, such as oleic acid, can cause significant changes in the application properties, such as the adhesive conductivity and stability (Markley et al. 1999). Therefore, it is important to understand how Ag fillers impact the properties of ICAs.

50

Electrical Industry

1455

(1)

Chip (2)

ICA

Bond pad

Substrate (3)

(4)

Syringe with underfiller

Temperature

underfiller

(5)

Temperature

Fig. 4 Bonding and packaging procedure with ICAs

The general form of ICAs is a viscous paste statistically filled with a sufficient content of metal flakes to provide conductivity in all directions (up to 30 vol.%). The conductivity between the particles is based on their touching each other. The process flow can be divided into five steps, as shown in Fig. 4 (Aschenbrenner et al. 1997): – Application of the isotropic conductive paste (ICP) precisely onto the points to be electrically connected – Alignment of the bumps to the electrodes on the substrate – Bonding the chip by curing the adhesive typically at over 150
C for several minutes – Applying an underfill material to compensate for the thermal mismatch – Curing of the underfill material Generally, the electrical conductivity of ICAs increases with increasing filler content, which means that they are transformed from an insulator to a conductor. This can be explained by percolation theory. Percolation theory was developed to mathematically deal with disordered media, in which the disorder is defined by a random variation in the degree of connectivity (Zallen 1983). The main concept of the percolation theory is the existence of a percolation threshold, which is a critical

1456

K.-S. Kim et al.

filler concentration at which the resistivity drops dramatically. At the percolation threshold, it is believed that all of the conductive filler particles come into contact with each other and form a three-dimensional network. The resistivity decreases only slightly as the filler concentration is further increased (Lau et al. 2003). Figure 5 schematically describes the decrease in the resistivity with increasing filler content based on percolation theory. The ICAs used in flip chip technologies serve not only to provide electrical continuity but also to support the mechanical load experienced during chip operation. Therefore, the mechanical strength of ICA joints has been another major reliability concern in electronic assembly technology. The electrical conductivity is proportional to the filler content; however, a too high filler concentration can deteriorate the mechanical integrity. The electrical continuity is achieved by the contact of the metal fillers, but the mechanical integrity is provided by the curing of the polymer resin. Therefore, the challenge in an ICA formulation is to use as much conductive filler as possible to achieve high electrical conductivity without affecting the mechanical properties adversely (Lau et al. 2003).

50.2.2 Anisotropic Conductive Adhesives From the viewpoint of their conduction and mechanical joining, ACAs are similar to ICAs, except that they have lower concentrations of conductive particles. This lower concentration provides unidirectional conductivity in the vertical or z-direction (perpendicular to the plane of the substrate), which is why they are called anisotropic conductive adhesives. In the same way, ACA materials are prepared by dispersing electrically conductive particles in an adhesive matrix at a concentration far below the percolation threshold. The concentration of particles is controlled, so that sufficient particles are present to ensure reliable electrical conductivity between the assembled parts in the z-direction, while insufficient particles are present to achieve percolation conduction in the X-Y plane (Kim Fig. 5 Resistivity with volume fraction of a conductive filler according to the percolation theory

50

Electrical Industry

1457

Fig. 6 Schematic description of flip chip structure with ACA

Si chip Insulation Conduct Substrate

a

b Si chip Electrode-1

Solder

Electrode-2

Si chip Electrode-1

Electrode-2 Substrate

Substrate

Fig. 7 Structure comparison between solder and ACA joints: (a) flip chip with solder bumps, (b) flip chip with ACAs

et al. 2008b). Figure 6 shows a schematic description of an ACA interconnect, showing the electrical conductivity in the joint. The volume loading of conductive filler in ACAs is 5–20 vol.%, which is much lower than that of ICAs. This low volume loading is insufficient for interparticle contact and prevents conductivity in the X-Y plane of the adhesives. The Z-axis adhesive, in film or paste form, is interposed between the surfaces to be connected (Li and Wong 2006). However, the controlled conductive filler loading still provides a higher resistance interconnect through the thickness of the adhesive than that of the solder-bumped interconnect. A comparison between the solder-bumped and ACA joint structures is shown in Fig. 7. From Fig. 7, it can also be seen that the area of the pad used for joining with the solder should be much larger than that used for ACAs. When we employ solder for the interconnect, the smallest pitch, which means the distance between the center point of a solder bump and that of the next one, is more than 50 μm, but it can be decreased to less than 30 μm by using ACAs. This limitation of solders based on ECAs is mainly caused by the difficulty involved in the bumping technology of solder, e.g., the screen printing of the solder paste, the evaporation of the solder materials, and the electroplating process. Using these processes, aligning the mask opening before the deposition of the solder is essential and requires high accuracy, so

1458

K.-S. Kim et al.

the extent to which the pitch and joining area can be reduced is limited. However, ultrafine pitch interconnections can be achieved when employing ACAs for joining. The fine pitch capability of ACAs is limited by the particle size of the conductive filler, which is only a few microns or a few nanometers in diameter. Therefore, a finer pitch is easily achieved these days by using smaller particles. Heat and pressure are simultaneously applied to this stack-up until the particles bridge the two conductor surfaces. ACAs utilize both a thermosetting polymer and a thermoplastic polymer for their adhesive matrix. The thermosetting type is hardened by the thermally initiated chemical reaction during heating, while the thermoplastic matrix is hardened by cooling after heating. Many recently commercialized ACAs employ a thermosetting polymer based on epoxy resin, which provides a high heat resistance and excellent adhesion to a variety of substrates, although they are not thermally reversible and cannot be dissolved in common organic solvents, because they are crossed-linked after bonding. In addition, thermosetting adhesives provide a low interconnection resistance, because they are easy to exclude between the bonding electrodes during the bonding process. The main advantage of thermoplastic adhesives is the relative ease with which the interconnection can be disassembled for repair operations, because they are rigid materials at temperatures below the Tg of the polymer, but exhibit flow characteristics above their Tg (Li and Wong 2006). Conductive fillers are generally metal microparticles or polymer cores coated with some high conductivity metals. Their basic particle shape is spherical. Ag is the most frequently used material for metal particles, due to its moderate cost, high electrical conductivity, high current-carrying capacity, and low chemical reactivity. Au is another widely employed filler material, but its cost prohibits large volume applications. An alternative to these noble metal fillers is polymer particles coated with conductive metals, such as Au, Ag, and Cu. The size of the polymer cores generally ranges from 1 to 5 μm, and they are coated with a Cu, Ni, or Au layer a few nanometers in thickness. Polymers generally have viscoplastic properties, so that they allow the external compressive stresses applied to the package to be easily released. Figure 8 shows the two different deformation behaviors of the particles when external forces are applied to the ACA joint. However, most importantly, they are much cheaper to make than metal-only fillers, so that they are the most commonly used conductive fillers in industry.

a

b Si die Electrode-1

Si die Electrode-1

Electrode-2

Electrode-2

PCB

PCB

Fig. 8 Schematic description of two deformation behaviors in ACA joints: (a) soft conductive fillers, (b) rigid conductive fillers

50

Electrical Industry

1459

50.2.3 Non-conductive Adhesives In contrast to ICAs and ACAs, NCAs do not consist of conductive particles but rather maintain a pure mechanical contact between the bumps and pads due to the application of a compressive force, and it is this that ensures the electrical connectivity between them. The quality of the electrical contact depends solely on the magnitude of the mechanical contact pressure. A high mechanical contact pressure is critical in providing a low electrical connection resistance, which is induced by the cross-linking shrinkage and thermal strain during the curing and cooling processes. The formation of contact spots depends on the surface roughness of the contact partners. Bringing the two surfaces close to each other enables a small number of contact spots to form, which allows the electric current to flow. When the parts are pressed together during the sealing process, the number and area of these single contact spots are increased according to the macroscopic elasticity or flexibility of the parts and the microhardness and plasticity of the surfaces, respectively (Li and Wong 2006). Nowadays, NCAs are being considered as an alternative to ACAs, because they can achieve a finer pitch of input/outputs (I/Os), due to the absence of any dispersion of conductive particles between the I/Os and better electrical properties. NCA joints avoid short-circuiting and are not limited by the particle size or percolation phenomena, leading to a reduction of the connector pitch. Another important advantage of NCAs is their cost-effectiveness, which is achieved by the exclusion of conductive fillers. Figure 9 compares the shape of ACA and NCA joints enlarging electrical contact area.

50.3

Conductive Filler Materials for ECAs

The conductive filler in the conventional ECAs is metallic particles or flakes. These filler materials have some critical limitations in terms of lower electrical conductivity and inferior impact performance induced by unstable network formation of particle fillers. Considering these challenges of traditional filler materials, there are three

a

b Si chip

Si chip

Electrode-1 Electrode-1 Coated conductive ball Electrode-2 Electrode-2 Substrate

Substrate

Fig. 9 Comparison of contact areas between ACA and NCA joints: (a) ACA joint, (b) NCA joint

1460

K.-S. Kim et al.

major approaches to enhance the electrical conductivity of ECAs, which include (i) hybrid metallic fillers, (ii) carbon-based nanomaterials, and (iii) hybrid nonmetallic fillers: conductive polymers and poly(3,4-ethylenedioxythiophene)-poly (styrenesulfonate) (PEDOT:PSS).

50.3.1 Hybrid Metallic Fillers Metallic fillers are widely used in the field of printed electronics due to their high conductivity and excellent compatibility with metal electrodes. Printed components using ink or paste with the metallic fillers have been used in fabrications of various applications such as flexible radio frequency identification (RFID) tags, flexible heaters, and organic light-emitting diodes. In general, printed components are thermally treated in high temperature in order to decrease contact resistance formed among metal fillers or induce interparticle necking by atomic interdiffusion. Recently, highly conductive interconnects can also be established via some other technical approaches, e.g., plasma, microwave, and intense pulsed light (IPL) treatment, instead of thermal heating (Kim et al. 2013, 2015; Song et al. 2015). Figure 10 shows a schematic of IPL sintering and its commercial system which provide the selective sintering at lower temperatures within several microseconds. Hybrid metallic fillers mean the incorporation of one-dimensional metallic nanowires into metal particle or flake fillers to improve electrical conductivity through an increase of metallurgical connections between each individual metallic filler. The metallic nanowires prefer to establish conductive networks rather than metallic particle or flake fillers because one-dimensional metallic structures are capable of forming a continuous network at lower loading level of the filler than spherical or platelet fillers (Grujicic et al. 2004). In particular, at lower temperatures, sintering process is controlled by surface diffusion due to mass flow to minimized Gibbs free energy (Mayo 1996), resulting in weak contacts between conventional metallic fillers. Since the

Fig. 10 (a) A schematic of IPL sintering and (b) its system (PSFL-4021-PC, www.pstek.co.kr)

50

Electrical Industry

1461

metallic nanowires are also likely to sinter at a low temperature, the large longitudinal size of nanowires makes it easy for fillers contacted weakly to sinter each other and bond. As shown in Fig. 11, the increased number of conductive pathways and the better quality of conductive linkage are obtained by mixing metallic nanowires with metal particles or flakes. Since the nanowires act as bridges among the spherical metal fillers, electrical conductivity of ECAs is dramatically improved (Chen et al. 2007). The dimension of metal fillers in ECAs critically determines their electrical conductivity. As explained above, the conventional filler is metal particles (0 dimensional), and different dimensional fillers such as nanowires and nanorods (1 dimensional) are regarded as promising conductive fillers, since they can provide electrical percolation at relatively low concentrations (Wu et al. 2007). Figure 12 shows the examples of different dimensional silver fillers. Chen et al. reported the incorporation of silver nanowires into ECAs (Chen et al. 2010). It was found that the electrical conductivity of the system composed with silver nanowires and nanoparticles is higher than that with either one of the two co-fillers (Chen et al. 2010). Zhang et al. reported the sintering characteristics of ECAs filled with surface-modified silver nanowires (Zhang et al. 2011). The sintering of silver nanowires can form high electrical conductivity of the ECAs although the high temperature of heat treatment is not suitable for most epoxy adhesives (Zhang et al. 2011). Yang et al. described a novel formulation method for a high-performance ECA which consisted of silver nanorods, nanoparticles, and modified epoxy matrix (Yang et al. 2012). The ECA film was able to achieve low resistivity of 3–4  105 Ω cm.

50.3.2 Carbon-Based Nanomaterials To meet the requirements for high-performance electronic devices, ECAs are required to achieve high electrical conductivity at lower filler contents. Recently,

Fig. 11 Illustration of conductive paths for ECAs filled with (a) silver nanoparticles and (b) silver nanowires

1462

K.-S. Kim et al.

Fig. 12 Field-emission scanning electron microscope images of (a) silver nanoparticles and (b) silver nanowires

Fig. 13 Transmission electron microscopy images of (a) the graphene flakes and (b) multi-walled carbon nanotubes

many efforts have been devoted to obtaining high conductivity of ECAs with low silver loading by introducing the carbon-based nanomaterials such as carbon nanotubes (CNTs) and graphene into the conventional formulation of ECAs. Figure 13 shows examples of representative carbon-based nanomaterials. CNTs are one of the ideal reinforcing carbon-based nanomaterials for enhancement of polymer composites because of their intrinsic marvelous physical, electrical, and thermal properties. Addition of CNTs can improve the mechanical properties as well as the electrical conductivity. However, the introduction of CNTs into ECAs has significant challenges in processing, dispersion control, and their high cost (Yim et al. 2016). Graphene is also one of the promising conductive fillers which can provide more stable electrical networks for ECAs’ application. It is confirmed that graphene is much more effective than other carbon-based filler materials. However, the agglomeration problem of carbon-based filler materials including CNTs and

50

Electrical Industry

1463

graphene leads to other problems such as increase in viscosity and poor dispersion, which cause the outstanding properties to be suppressed, resulting in the ECAs with minimal performance improvement (Amoli et al. 2015). To overcome these problems, (1) surface modification of graphene using organic materials (to exfoliate graphene in which the interaction occurs via either covalent bonding or via π-π stacking) (Kim et al. 2012) and (2) surface decoration of graphene with inorganic nanoparticles (incorporation of silver nanoparticle-decorated graphene to reduce the tunneling resistance) (Amoli et al. 2015) have been suggested.

50.3.3 Hybrid Nonmetallic Fillers (Conductive Polymers and PEDOT: PSS) This approach to improve the electrical conductivity of ECAs is mixing with conventional ECAs with conductive polymers such as polypyrrole (PPy) (Mir and Kumar 2010), polyaniline (PANI) (Desvergne et al. 2011), and polypoly (3,4-ethylenedioxythiolphene)-poly(styrenesulfonate) (PEDOT:PSS) (Chueh et al. 2015). The electrical conductivity of conductive polymers are determined by the loading, dispersion, and geometry (shape and diameter) of conductive fillers, degree of resin curing, and cross-link density, and so forth (Lilei et al. 1999; Kim and Shi 2001; Lee et al. 2005). Chemical structures of the representative electrically conductive polymers are shown in Fig. 14. PPy is one of the promising conductive polymers for ECAs due to ease of synthesis, stability, and low cost (Mir and Kumar 2010). PPy has been modified with other polymers to form composites and improve its processing properties which

Fig. 14 Schematic of the representative electrically conductive polymers: PPy, PEDOT, PSS, and PANI

1464

K.-S. Kim et al.

enable to provide better electrical performance such as smooth percolation curves. Mir et al. reported the composite manufactured by blending PPy with epoxy was utilized as ICAs, and their electrical conductivity, impact properties, and thermal and morphological characteristics were demonstrated with various concentrations of PPy in the epoxy matrix (Mir and Kumar 2010). PANI is also a suitable conductive polymer for improving the properties of the conventional ECAs owing to its high polymerization yield, good environmental stability combined with moderate electrical conductivity, and low cost. Mir et al. presented the composite of PANI in an epoxy/anhydride matrix, which is suitable as ICAs with enhanced impact properties and conductivity (Mir and Kumar 2010). Gumfekar et al. reported the reduced percolation threshold and enhanced conductivity of the silver-PANY-epoxy system via improved dispersion of silver fillers (Gumfekar et al. 2011). PEDOT:PSS has recently been studied and is emerging as a major attraction for next commercial ECAs. Chueh et al. reported ECAs comprising of the silver-plated graphite powder, PEDOT:PSS, and organic additives improved the electrical conductivity by at least two orders of magnitude compared with those filled with graphite powders only (Chueh et al. 2015).

50.4

Specific Requirements

There are at least three important requirements for the application of conductive adhesives in the electrical industry. The first one is the electrical resistance. Because of the typical low electrical conductivity of the polymer materials, a high electrical conductivity or low resistivity is definitely the most important factor for the conductive adhesive. Secondly, we should carefully consider the thermal properties of the adhesive. The bonding process with the adhesives is optimized by their thermal properties, e.g., their Tg and curing behavior. Finally, we would like to discuss in this chapter the thermomechanical properties of the adhesives. The thermomechanical properties determine the long-term reliability of the adhesive interconnect, so it is necessary to review how the thermomechanical properties can be measured and the results analyzed.

50.4.1 Electrical Conductivity When two metallic conductive pads are electrically interconnected by an adhesive, there is a large difference in their electrical conductivity in comparison with solderbumped interconnects. The solder-bumped joint is based on the melting of the solder and chemical reaction between the solder and metallic pad. In this case, some chemical compounds such as intermetallic compound layers form at the interface, which makes it possible to fill the interface without any large gap remaining.

50

Electrical Industry

1465

Therefore, the electrical resistance of the solder joint largely depends on the bulk resistance of the constituent components, e.g., the metal pad used for bonding, intermetallic compounds, and solder. However, in the case of adhesive interconnects, the joining process is essentially based on the mere physical contact between the pad and pad or pad and conductive fillers included in the electrically conductive adhesives. When two conductors are connected by a mere physical contact, the interfacial resistance, i.e., the contact resistance, is much larger than the bulk resistance when a current is flowing through the joint pathway. Therefore, the calculation or exact measurement of the contact resistance is a key point to evaluate the electrical conductivity of each adhesive interconnect. There are three main types of electrically conductive adhesive interconnects, all of which include bulk resistance and contact resistance components. Here, we will discuss the electrical conductivity of the adhesive joint with a simple derivation of the total resistance of the ACA interconnect. When a spherical particle meets a metallic pad, there are three separate contributions to the resistance: the intrinsic conductive resistance of each particle, the constriction resistance at the contacts, and the tunneling resistance at the contacts (Holm 1979; Ruschau et al. 1992). Therefore, the connection resistance through the conductive particles, Rc, is expressed by RC ¼ Ri þ RCR þ Rf

(1)

where Ri is the intrinsic resistance of an individual particle, RCR is the constriction resistance caused by the bending of the current flow paths, and Rf is the tunneling or film resistance associated with any insulating film which may be covering each metallic surface. At the point of contact between the conducting particles and the metallic bumps or pads, the majority of the current will flow through the areas of metal-to-metal contact. Thus, the current is constricted to flow through these areas, and the resistance will therefore be higher than if the current were passing directly through a straight path with no constriction. The resistance increase caused by this structure is called the constriction resistance. Some examples of insulating films in which tunneling behavior is observed are oxidation layers, residues of resin in the adhesives, and contaminants on the metal films. The electrical conduction must then proceed across these gaps, which represent electrical potential barriers to conduction. Therefore, tunneling is needed to allow conduction between the electrodes. Tunneling is a process by which electrons on one side of the barrier jump to the other side without any energy input, in which case the resistance increase caused by the potential barriers is called the tunneling resistance. Among the contributions to the connection resistance described above, the intrinsic resistance of an individual conductive particle can be estimated with respect to its deformation. Since the metal-coated conductive particles are deformed plastically during the bonding process of an ACF, the numerical model for the calculation of the connection resistance should consider the plastic deformation of the conductive particles. Yim and Paik (1998) derived the connection resistance of an ACF joint

1466

K.-S. Kim et al.

a

r a

b R h

Fig. 15 Deformation model of the conductive particle: (a) before deformation and (b) after deformation

from Holm’s electrical contact theory, and the derived model described the connection resistance of the ACF in terms of the sample geometry, particle size, applied pressure, and the intrinsic properties of the conductive particles. However, this model is valid only for the elastic deformation of the conductive particles; therefore, it cannot reflect their plastic deformation. For a more exact prediction of the connection resistance, the plastic deformation of the conductive particles should be considered. Maattanen (2003) derived the connection resistance of metal-coated polymer particles using the degree of deformation during the bonding process of an ACF. He described the degree of deformation as
Degree of deformation ¼ 1 

h 2R

 (2)

where h is the distance between the electrodes and R is the radius of the undeformed conductive particles which are shown in Fig. 15. The area of the spherical cap forms the metal contact area, and, therefore, the metal contact area, A, and the length of the conductor, l, are given as
A ¼ 2πR

2

l¼

h 1 2R

πh 2

 (3)

(4)

However, since the current would flow only through the thin metal layer of the conductive particles, the average area of the metal film should be calculated. The average area can be derived as follows:

50

Electrical Industry

1467

dS
 Ð πh=4 0 1 2πtR cos α ¼ Ð πh=4R A dS 0

Ð πh=4R

dα 2πtR cos α ¼ Ð πh=4R dα 0 ð πh=4R 4R 1 dα ¼ πh 2πtR 0 cos α π αi 2 πh=4R h ¼ 2 0 þ ln tan π ht 
4 2 2 π h 1þ ¼ 2 ln tan π ht 4 2R 0

(5)

where r = R cos (α), S = Rα, dS = Rdα, and t is the thickness of the metal layer on the particle. For α, see Fig. 15. Since there is spherical symmetry, an approximation can be made, and the problem is solved using Ohm’s law, and the formula for the intrinsic resistance of the deformed particle, Ri, is written as follows: Ri ¼ ρ


 
 l πh 1 ρ π h ¼ρ 1þ ¼ ln tan A 2 A πt 4 2R

(6)

where ρ is the resistivity of the metal coating. In electrical interconnects made using ACAs, electrical conduction is provided by the conducting particles trapped within the interconnection area. At the point of contact between the conducting particles and the metallic bumps or pads, the majority of the current will flow through the areas of metal-to-metal contact. Thus, the current is constricted to flow through these areas, and the resistance will, therefore, be higher than if the current were passing directly along a straight path with no constriction. This resistance is called the constriction resistance. According to Holm (1979), the constriction resistance, RCR, can be written as RCR ¼

ð ρ1 þ ρ2 Þ 2d

(7)

where ρ1 and ρ2 are the intrinsic resistivities of the conductive particles and substrate, respectively, and d is the diameter of the metal contact area. Looking closer at the diameter of the contact area in different compression situations, it is possible to derive a formula for it in the case of metal-coated polymer particles. The area is a spherical cap of the conductive particle where the size is a variable depending on the degree of deformation. The formula for the metal contact area is

1468

K.-S. Kim et al.

π


2
 d h ¼ 2πR2 1  2 2R

(8)

Therefore, sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 h d ¼R 8 1 2R

(9)

Substituting Eq. 9 into (7) yields RCR ¼

ð ρ1 þ ρ2 Þ sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 h 2R 8 1  2R

(10)

For contacting materials that form metal-to-metal contacts and that do not oxidize, the tunneling resistance, Rf, is zero. This oxidation problem is considered a special case, and, therefore, we can neglect the oxidation term in this simple derivation. Finally, according to the schematic of an ACF interconnection, the connection resistance, Rc, considering the plastic deformation of a metal-coated polymer sphere is described by Rc ¼ RCR1 þ RCR2 þ Ri 
 ðρ1 þ ρ2 Þ ð ρ1 þ ρ3 Þ ρ1 π h ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi s s ¼
þ
 þ πt ln tan 4 1 þ 2R h h 2R 8 1  2R 8 1  2R 2R

(11)

where RCR-1 and RCR-2 are the constriction resistances generated between the conductive particle and the two electrodes, ρ1 is the resistivity of the conductive particles, and ρ2 and ρ3 are the resistivities of the electrodes. Using this derivation, the connection resistance of the ACA interconnect can be easily calculated by the simple investigation of the physical shape of the connected pathway. This derivation is based solely on the physical contact of the two unmelted metals, so it could be applied to the other two types of adhesives, i.e., ICAs and NCAs. In the case of ICAs, the shape of the current flowing path is more complex, which means that the contact resistance might also be more complex than that in the ACA case. However, the resistance components of an ICA interconnect are completely identical to those of an ACA interconnect, so the derivation and calculation of the components are also easy to perform. In the case of an NCA, there are no conductive fillers and, therefore, RCR and Ri can be excluded from the derivation. This means that the connection resistance for an NCA joint consists only of Rf, so that the resistance should also be lower than that of an ACA joint. In summary, the electrical conductivity of adhesive joints is dominated by the physical contact between the two unmelted conductive materials, and, therefore, the

50

Electrical Industry

1469

resistance of the joint seems to be nearly equal to the contact resistance. Among the various adhesives, ACAs are considered to be the least conductive materials, and the resistance is controlled by the joint structure and constituent materials.

50.4.2 Thermomechanical Properties According to various previous studies concerning the bonding process for the joining of adhesive electronic components, it was reported that the electrical properties of the ACF flip chip package are highly affected by the bonding temperature and pressure. Especially, the contact resistance of the adhesive joints is highly correlated with the deformation of the metal bump or conductive fillers induced by the pressure applied during the bonding process. This is because if the conductive fillers are too spread out between adjacent bumps or pads, due to too much pressure being applied, they may end up contacting each other, creating the same effect as short-circuiting; whereas if the bonding pressure is too low, the particles may not be able to make contact between the connecting bumps and pads. After the complete bonding process, the adhesive joints are established by the compressive forces generated by the shrinkage stress in the adhesive matrix. The shrinkage stresses originate from the thermomechanical properties of the adhesives, such as the thermal shrinkage and thermal stiffness generated during or after the thermal reaction (Kwon and Paik 2004; Yu et al. 2013). This shrinkage stress, influenced by the thermal shrinkage and buildup stiffness, is responsible for the contact integrity in an adhesive joint, and, therefore, the evaluation of the shrinkage behavior is very important for the optimization of the bonding process. Generally, the shrinkage stress is generated by the curing reaction of the adhesives, which means that the thermal and thermomechanical behavior of the adhesives during heating and cooling should be evaluated before the application of the adhesives in electronics. The first parameter we should examine is the degree of curing of the adhesives containing Tg and the reaction ratio during heating and cooling. Kwon and Paik (2004) described the evaluation of the degree of curing of an ACF joint with a differential scanning calorimeter (DSC). They defined the term, degree of curing, as follows: Degree of cure ¼

ΔH ðuncured adhesivesÞ  ΔH ðheated adhesivesÞ  100 (12) ΔHðuncured adhesivesÞ

Figures 16 and 17 present an example of the DSC measurement curves for an ACF and NCF. From Fig. 16, we can obtain the Tg information of the ACF and NCF employed, while the degree of curing is calculated using Fig. 17. Generally, for electronic applications, faster curing with a lower Tg is favorable for various reasons, such as the impact of a high temperature on other nearby components, the deviation of the process temperature from the conventional reflow profile for solder joining, and the manufacturing efficiency. Especially, from the viewpoint of the impact on nearby components, the Tg of the adhesive is highly correlated with its applicable

1470

K.-S. Kim et al. 5,000 ACA 0

NCA

Heat flow (mW)

–5,000 –10,000 –15,000 –20,000 –25,000 50

100

150

200

250

Temperature ( C)

Fig. 16 Comparison of heat flow with temperature between ACF and NCF

ACA

100 Degree of curing reaction (%)

NCA 80

60

40

20

0 0.0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

4.5

Time (min)

Fig. 17 Comparison of curing degree with time between ACF and NCF

areas in electronics. That is, the bonding temperature is determined by the thermal stability of the nearby components, e.g., the substrate. For example, in display applications, a glass panel is used as a substrate, which means that the bonding temperature can be as high as 400
C. However, polymer-based plastic substrates are typically used in printed circuit board (PCB) applications, where the bonding

50

Electrical Industry

1471

temperature must be less than the Tg of the substrate, which is generally less than 200
C. Therefore, adhesives with a lower Tg would be more favorable due to their wider range of applications. As shown in Fig. 17, the degree of curing might also be different between adhesives consisting of different materials. The degree of curing determines the bonding time for adhesive joining, so that the manufacturing efficiency is highly related to this factor. For flip chip joining with adhesives, a chip with numerous metal bumps should be aligned on a substrate, e.g., glass, printed circuit board (PCB), or flexible polyimidebased one. After aligning the chip, a certain bonding force is applied to the chip to generate a compressive force in the adhesive area. At this time, the temperature is raised to that required for curing, which is typically 30–50
C higher than the Tg. At this stage, the adhesive resin starts to undergo the curing reaction, which increases its viscosity and modulus, due to the development of a cross-linking structure. Both the bonding force and temperature are continuously applied in order for the adhesive to be fully cured. After the adhesive resin is fully cured, the bonding force is removed, and the assembly is cooled to room temperature. During the cooling process, shrinkage stress can develop, due to the thermal shrinkage and buildup of stiffness of the ACFs. Although the external bonding force is removed after the bonding process, the electrical conduction of the ACF joint is mechanically maintained by the compressive force due to the thermal shrinkage of the ACF (Kwon and Paik 2004). Referring to Fig. 17, it can be seen that a shorter time for full curing is favorable for the efficient joining of electronic components. Next, we should examine the thermomechanical properties through the measurement of the coefficient of thermal expansion (CTE) of the adhesives. The CTE values can be obtained by the measurement of the length deformation with temperature (Noh et al. 2008). The length deformation can be measured by using a thermomechanical analyzer (TMA). Generally, the smaller the length deformation with temperature, the higher the CTE value. It is anticipated that the CTE value of an adhesive will determine the thermomechanical reliability of the electronic interconnect with the adhesive, that is, the similarity of the CTE values of the adhesive and the substrate and other nearby components is favorable for its long-term reliability. If the CTE of the adhesive is significantly different from those of the nearby components, the resulting mismatch of the CTE values may give rise to certain kinds of reliability problems, such as the thermomechanical fatigue cracking of the material, delamination of a layer, and unstable electrical contact between the metals. Therefore, for the design of a new adhesive material, the CTE values of its constituent materials should be carefully considered, and reliability testing with a fully fabricated sample or a computer simulation is essential to verify its acceptability for the application.

50.5

Reliability

The short-term quality and long-term reliability are always the most important considerations to determine the value of materials. Typically, it is considered that the best method of evaluating the reliability of an electronic component is an actual

1472

K.-S. Kim et al.

application (Lau 1995). However, this is not always practical, especially during process development and optimization. Therefore, the reliability behavior of an interconnection is characterized by performing accelerated environmental tests, such as temperature cycling, thermal shock, high and low temperature storage, and high temperature and humidity testing. According to some previous studies, the main degradation mechanisms include the oxidation of the metal surfaces in the joint, stress concentration at a certain area during temperature fluctuations, and so on (Li and Wong 2006). Here, we will review the various reliability testing methods for adhesive interconnects, beginning with an introduction to test vehicle design for adhesive interconnects.

50.5.1 Test Vehicle Design Basically, the test vehicle for the accelerated testing should resemble the real components or interconnect systems. That is, the test vehicle design must be such that it can be processed using all the technologies of interest. In addition, the ability to change the wafer layout, such that the user has flexibility in selecting an appropriate chip size for the application, is sometimes desirable to evaluate its technological capability (Lau 1995). Adhesives are used to join various systems in electronic manufacturing, e.g., Si die bonding to rigid printed circuit board (flip chip application), Si die bonding to flexible substrate (COF, chip on flexible substrate), Si die bonding to glass substrate (COG, chip on glass), rigid printed circuit board to flexible substrate, and so on. Therefore, proper test vehicle design is possible only after the consideration of the real application of the adhesives developed. After considering the application, the materials used for the chip, substrate, metal bump for joining, the number of bumps on a die, pitch of bump array, and bump array specification can be easily determined in the case where the target application is definite. Then, we should design conductive metal circuits of the test vehicle, in order to measure the total resistance of the circuits or a single bump interconnect. Figure 18 shows an example of the test vehicle used for the reliability testing of the adhesives. As can be seen in the figure, the test vehicle is fabricated by joining an Si die to a flexible substrate that is typically made from polyimide. An adhesive film, ACF, was used for the joining process. This design is called the COF, and it is widely used in the manufacture of display modules such as liquid crystal displays (LCDs). Enlarged views of the test circuits are shown in Fig. 19. In the design of conductive circuits for electrical evaluation, it is important that we measure or calculate the resistance of only the adhesive interconnects, which consist of a metal bump on a die, adhesive, and metal pad on a substrate, excluding the resistance of any conductive traces on the Si die or substrate. This is possible if we apply the four-point electrical resistance measurement system with a proper design using the so-called Kelvin structure. Figure 20 shows a schematic description of the Kelvin structure for the selective electrical characterization. This structure contains four metal bumps and pads for adhesive joining. The traces used as the voltage sense and current sense can be interchanged. The principle of operation of this structure is fully described in the

50

Electrical Industry

1473

Fig. 18 Test vehicle sample for reliability testing of adhesive interconnects

Fig. 19 Scanning electron microscope image for surface of flip chip test vehicle

literature (Lau 1995) and is briefly introduced here as follows. Let us assume that the current flows through trace A, through bump 1, over the trace on the die, down bump 2, and out of trace B. It should be noted that traces C and D have no current flowing through them. Furthermore, trace C and bump 3 are at the same potential as the bottom of bump 2. Also, trace D and bump 4 are at the same potential as the top of bump 2. Therefore, any potential difference measured across traces C and D would represent the potential drop across bump 2. Given the current flowing through bump 2 and the measured potential difference across traces C and D, the bump resistance may be readily estimated.

1474

K.-S. Kim et al.

Fig. 20 Schematic description of four-point measurement for Kelvin structure design

50.5.2 Storage Test in Harsh Environments The past few years have witnessed explosive growth in the research and development efforts devoted to the improvement of adhesive joint reliability for mobile electronics. In the evaluation of the reliability, it is emphasized that the following must be noted in order to obtain a full understanding of the adhesive technology. To investigate the reliability of adhesives in an electronic component, the stability of the connection resistance of a single interconnect in the specific electrical system under environmental stress should be evaluated. Therefore, the design and fabrication of the test vehicle are essential prerequisites. Then, the environmental conditions to use for the reliability testing should be selected. Various testing methods are generally selected for the evaluation of adhesive interconnects, ranging from the reflow test to the pressure cooker test (PCT). Here, we will review the various environmental test methods and their failure analysis techniques. The reflow test is typically conducted to evaluate the thermal stability of adhesive interconnects. Reflow is the process of melting and cooling of the solder placed on a substrate for joining various component electrodes to metal pads. Today, adhesive interconnection technology is also applied to join ICs to a PCB, and, therefore, these kinds of products are packaged with adhesive technologies together with the conventional reflow process. In these soldering processes, the package needs to be heated up to above 200
C, which is much higher than the final bonding temperature of the adhesive interconnection. Therefore, the development of adhesives which can withstand multiple solder reflows was a major milestone in the process of its acceptance within the semiconductor packaging industry (Kim et al. 2008a). Figure 21 shows a reflow machine for soldering, and Fig. 22 shows a typical temperature profile for the reflow of Sn-based Pb-free solders. Thermal shock and temperature cycling tests are typically employed to evaluate the ability of the adhesive interconnect to resist the temperature variation during the operation of the electrical system. The adhesives in an electronic package serve as a mechanical mechanism for resisting the thermal deformation induced by the CTE mismatch between the upper chip and lower substrate. The thermally induced

50

Electrical Industry

1475

Fig. 21 Reflow machine (RF-430-N2, Japan Pulse Laboratory Ltd. Co., Japan)

Fig. 22 Reflow temperature profile for Sn-based Pb-free solders Temperature ( C)

250 200 150 100 50 0

60

120

180

240

300

360

Time (s)

fatigue cracking through a constituent material and the delamination of the adhesive from the attached area, resulting from the mismatch in the CTE of the chip and substrate, are the major wear-out failure modes in adhesive joints. The thermal fatigue process involves microcrack initiation and coalescence followed by crack propagation. Figure 23 shows an example of a catastrophic failure mode of adhesive interconnect after thermal shock testing. The thermal shock test is an accelerated test form of the temperature cycling test. The thermal shock test controls only the highest and lowest temperatures in the testing zones, and then moves the position of the test samples from the first zone to the other, which means that the samples experience a very abrupt temperature variation during the testing. Meanwhile, the temperature cycling test controls the total temperature profile, so that the test samples can experience milder temperature variations. The detailed test methods for the test of electronic components are provided by the Joint Electron Device Engineering Council (JEDEC) standards. Figure 24 shows an example of a thermal shock testing chamber.

1476

K.-S. Kim et al.

Fig. 23 Failure mode for ACA interconnect after thermal shock testing Fig. 24 Thermal shock testing chamber (TSA-101S, Espec, Japan)

The high and low temperature storage test is also typically used to evaluate the ability of the adhesives to resist a harsh environment. In this case, the temperature variation is not included in the testing profile, so the test evaluates the effects of a fixed environmental temperature on the stability of the electrical system. Conventionally, this test was widely conducted to investigate the effects of a certain

50

Electrical Industry

1477

temperature on the growth behavior of an intermetallic compound layer, which is a reaction layer formed between a solder and a metal substrate during a reflow process. However, it is also employed to evaluate the impact of long-term storage at a rigorously stabilized temperature on the electrical quality of the adhesive interconnect. The high temperature and humidity test is considered to be one of the most important test methods for the evaluation of adhesive joints, because of their high absorption of humidity. When a sample with an adhesive joint is placed in an environment having high humidity, the swelling of the polymeric conductive fillers in the ACA and z-directional expansion of the adhesive matrix occur. Figure 25 shows this behavior for the example of an ACA joint before and after the high temperature and humidity test (85
C and 85% relative humidity for 500 h). After the test, no significant defects such as delamination or cracking of the adhesive matrix were evident, but initially the deformed conductive particles swelled upward due to the release of the residual z-directional shrinkage stress and the expansion of the polymeric cores of the conductive particles and adhesive matrix. After the joining process with the adhesives at room temperature, the z-directional stress in the adhesive areas was mainly compressive stress, but it decreased with increasing temperature and humidity. A small compressive stress induces the swelling of the polymer particles and adhesive matrix layer, and ultimately electric disconnection may occur in the worst case. The swollen conductive particles decrease the contact area, and, thus, the electrical connection resistance should increase during the testing. For the failure analysis of the test samples after the harsh environmental tests, various analysis tools are employed, e.g., scanning electron microscopy (SEM) with a cross-sectional analyzing technique; X-ray microscopy; A-, B-, and C-mode scanning acoustic microscopy (A-SAM, B-SAM, and C-SAM, respectively); and

a

b

Bonding stage

Testing stage

Heat, force, and cooling

Clamping force

Conductive ball

Adhesion

Clamping force

Electrode-01 Force of expansion

Electrode-01

Force of expansion

Electrode-02 Adhesion Electrode-02

Adhesion

Fig. 25 Schematics of stress field around the conductive particle: (a) after bonding, (b) during testing at high temperature and humidity

1478

K.-S. Kim et al.

tomographic acoustic microimaging (TAMI), to analyze the failure samples. The continuous measurement of the electrical resistance should consequently be conducted during all of the environmental tests mentioned above.

50.5.3 Mechanical Reliability The adhesives used in the electronic industry serve not only to provide electronic continuity but also to support the mechanical load experienced during service. Therefore, the mechanical stability of the adhesive joints has been one of the major reliability concerns in electronic assembly technology. The mechanical peeling and shear test of the adhesive joints are typically performed to characterize the bonding strength at the adhesive-substrate interface. These two methods are very simple to implement. That is, the test procedure consists of the detachment of a component bonded with an adhesive on a substrate. During the detachment process, mechanical stress such as shearing or peeling is applied to the component, while the loading applied to a stress sensor is recorded for the characterization of the mechanical stability. Owing to the increasing demand for portable electronic devices, which are affected by low-frequency and random vibrations, the reliability of electronic packages subjected to repetitive mechanical loads has recently become an important issue in the electronic industry (Noh et al. 2009). In this case, assembly warpage has been recognized as an important factor for determining the package reliability, as it can lead to IC failure, due to the delamination of a layer and cracking through a specific material in the subsequent process caused by dimensional instability and non-coplanarity. In particular, the cyclic bending reliability testing of various packages under different test conditions is widely conducted at the board or system level (Noh et al. 2009). The three-point or four-point bending test is one of the most frequently employed test methods, due to its suitability for testing large package samples under similar loading conditions (bending moments) within the inner load span. Here, we will review the four-point bending test for the evaluation of the mechanical reliability of adhesive joints. Figure 26 shows a four-point bending tester used for testing a sample with an ACF. The tip of each support consists of a roller to ensure line contact with the test board, as shown in the figure. In this case, the in situ electrical resistance should also be measured, in order to identify the failure of the package during bending testing. The bottom supports are fixed, while the top supports move downward repetitively with the displacement profile prescribed by the sinusoidal movement of the actuator. The top and bottom support spans are fixed at a certain distance apart. For the measurement of the electrical resistance, a current is flowed through the metal circuits and adhesive joints. An event detector is generally used to determine when the overall electrical resistance exceeds a specific value. The experiments are typically performed at room temperature, but could also be done at higher ones. The bending fatigue test can be conducted using the load-control or displacementcontrol method.

50

Electrical Industry

1479

Fig. 26 Four-point bending tester for adhesive joint

50.6

Conclusions

This chapter describes the adhesive technologies used in the electrical industry with an introduction to the conduction mechanisms in adhesive joints, the specific requirements for the application of adhesives in electronics, and reliability evaluation methods. We will summarize this chapter with the following conclusions: 1. There are three types of electrically conductive adhesives typically used in the electrical industry, e.g., isotropic conductive, anisotropic conductive, and non-conductive adhesives. 2. Isotropic conductive adhesives have a large amount of conductive fillers (25–30 vol.%) with the electrical conductivity provided by the contact between the conductive fillers. 3. Anisotropic conductive adhesives are similar to isotropic conductive adhesives, except that they have a lower conductive filler content (5–20 vol.%) and thus provide unidirectional conductivity in the vertical or z-direction (perpendicular to the plane of the substrate). 4. Non-conductive adhesives have no conductive fillers but rather maintain a pure mechanical contact between the bumps and pads, due to the application of a compressive force, and it is this that ensures the electrical connectivity between them. 5. Considering the drawbacks of conductive filler materials in conventional ECAs, there are three major approaches to enhance the electrical conductivity and thermomechanical properties, which include (i) hybrid metallic fillers, (ii) carbon-based nanomaterials, and (iii) hybrid nonmetallic fillers: conductive polymers and PEDOT:PSS. 6. The conductivity of electrically conductive adhesives is described by percolation theory, which explains that the resistivity of an adhesive drops dramatically at a certain vol.% of conductive fillers.

1480

K.-S. Kim et al.

7. The connection resistance at the interface between the adhesives and metal bumps is composed of three resistance components, i.e., the intrinsic resistance of the constituent materials, constriction resistance, and tunneling resistance. 8. Shrinkage stress is generated by the curing reaction of the adhesives, which means that the thermal and thermomechanical behavior of the adhesives during the bonding process should be evaluated before their use in electronics. 9. For the reliability testing of adhesives in electronics, a test vehicle for accelerated testing should be fabricated, which resembles the real components or interconnect systems. This means that the design of the test vehicle must be such that it can be processed using all of the technologies of interest. 10. Various testing methods are employed to evaluate the reliability of adhesives in electronics, such as temperature cycling, thermal shock, high and low temperature storage, and high temperature and humidity testing. 11. The mechanical peeling and shear tests of the adhesive joints are typically performed to characterize the bonding strength at the adhesive-substrate interface.

References Amoli BM, Trinidad J, Hu A, Zhou YN et al (2015) Highly electrically conductive adhesives using silver nanoparticle(Ag NP)-decorated graphene: the effect of NPs sintering on the electrical conductivity improvement. J Mater Sci: Mater Electron 26:590–600 Aschenbrenner R et al (1997) Adhesive flip chip bonding on flexible substrates. In: Proceedings of 1st IEEE international symposium on polymeric electronics packaging, IEEE, p 86 Chen C, Wang L, Li R et al (2007) Effect of silver nanowires on electrical conductance of system composed of silver particles. J Mater Sci 42:3172–3176 Chen D, Qiao X, Qiu X et al (2010) Effect of silver nanostructures on the resistivity of electrically conductive adhesives composed of silver flakes. J Mater Sci: Mater Electron 21:486–490 Chueh TC, Hu CH, Yen SC (2015) Electrically conductive adhesives with low Ag content prepared by Ag self-activated plating and PEDOT:PSS. J Electrochem Soc 162:D56–D61 Desvergne S, Gasse A, Pron A (2011) Electrical characterization of polyaniline-based adhesive blends. J Appl Polym Sci 120:1965–1973 Grujicic M, Cao G, Roy WN (2004) A computational analysis of the percolation threshold and the electrical conductivity of carbon nanotubes filled polymeric materials. J Mater Sci 39:4441–4449 Gumfekar SP et al (2011) Silver-polyaniline-epoxy electrical conductivity adhesives – a percolation threshold analysis. In: IEEE 13th electronics packaging technology conference, IEEE, p 180 Habenicht G (1990) Kleben. Springer-Verlag, Berlin Holm R (1979) Electric contacts. Springer-Verlag, Berlin Kim HK, Shi FG (2001) Electrical reliability of electrically conductive adhesive joints: dependence on curing condition and current density. Microelectron J 32:315–321 Kim JW, Kim DG, Hong WS et al (2005) Evaluation of solder joint reliability in flip-chip packages during accelerated testing. J Electron Mater 34:1550–1557 Kim JW, Kim DG, Lee YC et al (2008a) Analysis of failure mechanism in anisotropic conductive and non-conductive film interconnections. IEEE Trans Compon Packag Tech 31:65–73

50

Electrical Industry

1481

Kim JW, Lee YC, Jung SB (2008b) Reliability of conductive adhesives as a Pb-free alternative in flip-chip application. J Electron Mater 37:9–16 Kim J, Yim BS, Kim JM, Kim J (2012) The effects of functionalized graphene nanosheets on the thermal and mechanical properties of epoxy composites for anisotropic conductive adhesives (ACAs). Microelectron Reliab 52:595–602 Kim KS, Bang JO, Choa YH et al (2013) The characteristics of Cu nanopaste sintered by atmospheric-pressure plasma. Microelectron Eng 107:121–124 Kim KS, Park BG, Jung KH et al (2015) Microwave sintering of silver nanoink for radio frequency applications. J Nanosci Nanotechnol 15:2333–2337 Kwon WS, Paik KW (2004) Fundamental understanding of ACF conduction establishment with emphasis on the thermal and mechanical analysis. Int J Adhes Adhes 24:135–142 Lau JH (1995) Flip chip technologies. McGraw-Hill, New York Lau JH, Wong CP, Lee NC, Lee SWR (2003) Electronics manufacturing with lead-free, halogenfree & conductive-adhesive materials. McGraw-Hill, New York Lee HH, Chou KS, Shih ZW (2005) Effect of nano-sized silver particles on the resistivity of polymeric conductive adhesives. Int J Adh Adh 25:437–441 Li Y, Wong CP (2006) Recent advances of conductive adhesives as a lead-free alternative in electronic packaging: materials, processing, reliability and applications. Mater Sci Eng R 51:1–35 Lilei Y, Zonghe L, Johan L et al (1999) Effect of Ag particle size on electrical conductivity of isotropically conductive adhesives. In: IEEE transactions on electronics packaging manufacturing, IEEE, p 299 Liu YH, Lin KL (2005) Damages and microstructural variation of high-lead and eutectic SnPb composite flip chip solder bumps induced by electromigration. J Mater Res 20:2184–2193 Maattanen J (2003) Contact resistance of metal-coated polymer particles used in anisotropically conductive adhesives. Solder Surf Mount Tech 15:12–15 Markley DL, Tong QK, Magliocca DJ et al (1999) Characterization of silver flakes utilized for isotropic conductive adhesives. In: Proceedings of international symposium on advanced packaging materials, IEEE, p 16 Mayo MJ (1996) Processing of nanocrystalline ceramics from ultrafine particles. Int Mater Rev 41:85–115 Mir IM, Kumar D (2010) Development of polypyrrole/epoxy composites as isotropically conductive adhesives. J Adhes 86:447–462 Noh BI, Lee BY, Jung SB (2008) Thermal fatigue performance of Sn-Ag-Cu chip-scale package with underfill. Mater Sci Eng A 483–484:464–468 Noh BI, Yoon JW, Kim JW et al (2009) Reliability of Au bump flip chip packages with adhesive materials using four-point bending test. Int J Adhes Adhes 29:650–655 Plumbridge WJ (2004) Long term mechanical reliability with lead-free solders. Solder Surf Mt Tech 16:13–20 Ruschau GR, Yoshikawa S, Newnham RE (1992) Resistivities of conductive composites. J Appl Phys 72:953–959 Song CH, Ok KH, Lee CJ et al (2015) Intense-pulsed-light irradiation of Ag nanowire–based transparent electrodes for use in flexible organic light emitting diodes. Org Electron 17:208–215 Wu H, Wu X, Ge M et al (2007) Properties investigation on isotropical conductive adhesives filled with silver coated carbon nanotubes. Comps Sci Technol 67:1182–1186 Yang X, He W, Wang S et al (2012) Preparation and properties of a novel electrically conductive adhesive using a composite of silver nanorods silver nanoparticles and modified epoxy resin. J Mater Sci: Mater Electron 23:108–114 Yim MJ, Paik KW (1998) Design and understanding of anisotropic conductive films (ACF’s) for LCD packaging. IEEE T Comp Pack Manu Tech 21:226–224 Yim BS, Kim K, Kim J et al (2016) Influence of functionalized graphene on the electrical, mechanical, and thermal properties of solderable isotropic conductive adhesives. J Mater Sci: Mater Electron 27:4516–4525

1482

K.-S. Kim et al.

Yin CY, Lu H, Bailey C et al (2006) Analyzing the performance of flexible substrates for lead-free applications. In: Proceedings of 7th international conference on thermal, mechanical and multiphysics simulation and experiments in micro-electronics and micro-systems, IEEE, p 1 Yu H, Adams RD, da Silva LFM (2013) Development of a dilatometer and measurement of the shrinkage behaviour of adhesives during the cure process. Int J Adhes Adhes 47:26–34 Zallen R (1983) The physics of amorphous solids. Wiley, New York Zhang YL, Shi DXQ, Zhou W (2006) Reliability study of underfill/chip interface under accelerated temperature cycling (ATC) loading. Microelectron Reliab 46:409–420 Zhang ZX, Chen XY, Xiao F (2011) The sintering behavior of electrically conductive adhesives filled with surface modified silver nanowires. J Adhes Sci Technol 25:1465–1480

Shoe Industry

51

José Miguel Martín-Martínez

Contents 51.1 51.2 51.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Overview of the Shoe Bonding Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface Preparation of Upper and Sole Materials for Bonding . . . . . . . . . . . . . . . . . . . . . . . . 51.3.1 Upper Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.2 Sole Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.3 Adhesives in Upper-to-Sole Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.4 Polyurethane Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.3.5 Polychloroprene Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.4 Testing of Adhesive Joints in Shoe Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1484 1486 1487 1488 1490 1502 1503 1517 1523 1526 1527

Abstract

This chapter constitutes one of the very few reviews in the existing literature on shoe bonding, and it gives an updated overview of the upper-to-sole bonding by using adhesives. The surface preparation of rubber soles and the formulations of the solvent-borne and waterborne polyurethane and polychloroprene adhesives are described in more detail. The preparation of the adhesive joints and the specific adhesion tests in shoe bonding are also revised. Finally, the most recent developments dealing with shoe bonding are described.

J. M. Martín-Martínez (*) Adhesion and Adhesives Laboratory, University of Alicante, Alicante, Spain e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_51

1483

1484

51.1

J. M. Martín-Martínez

Introduction

The attachment of the upper to the sole by means of adhesives in shoe manufacturing is one of the most exigent bonding technologies. Several different types of materials are used in shoe manufacturing most of them varying in nature and composition from year to year because of fashion requirements. Furthermore, the geometrical shape and the design of the upper and the sole change often in male and female shoes, and several models and kinds of shoes can be manufactured (casual shoes, sport shoes, boots, safety shoes, sanitary and orthopedic shoes). For each shoe, bonding must be specifically designed, and this situation happens twice a year (spring and fall seasons in shoe industry). This is one of the reasons for the difficulty in establishing unified methodologies for shoe bonding as there is no time for shoe manufacturers to react once debonding appears, i.e., the upper and the sole separate during use (Fig. 1). Another particular feature in shoe industry is the lack of quality control of the materials and the adhesives used for bonding. Shoe manufacturers have extensive knowledge on the technology of making shoes, but their knowledge on bonding comes mainly from the experience and the advice of the adhesive manufacturers. Certainly, the bonding of the shoe parts is more of an art than a technology, and the shoe manufacturers are always worried and even scared when they have to attach the upper to the sole in the new models of shoes in the season. In fact, it is a common practice in several famous shoe brands to sew the upper to the sole simply because they do not trust in the quality of the adhesive bonding! The above given arguments justify the writing of a review on the bonding of shoe industry based mainly on the experience of the author in the field acquired during more than 25 years. The bonding with adhesives has been introduced in the footwear industry as an alternative to sewing or the use of nails, staples, or tacks to bond several parts of the Fig. 1 Upper-to-sole detachment in a shoe

Leather upper

Adhesive

Rubber sole

51

Shoe Industry

1485

shoes, the most critical part being the upper-to-sole bonding. The introduction of adhesive technology in shoe manufacturing provided several advantages: (i) More flexible and homogeneous joints were obtained; (ii) the stresses during use were similarly distributed all along the joints; (iii) improved aesthetic properties were obtained, and new design and fashion shoes were possible; and (iv) automation was feasible. However, as a limitation, the bonding with adhesives in shoe industry needs a severe control of all steps involved in the formation of the joints to avoid adhesion problems, mainly the separation of the sole from the upper. Although several parts of a shoe are bonded with adhesives (mounting, heel covering, box toe bonding, sticking of the soft lining, lift attachment, shank and cushion bonding, etc.), this chapter will be devoted only to the upper-to-sole attachment because it is the most critical bonding operation in shoe manufacturing. Depending on the type of shoe, different bonding performance is required. Due to fashion issues, the bonding of the casual and the free-time shoes is relatively low exigent as the requested durability is not too high (3 years maximum). However, the sport shoes are extremely exigent in bonding as low weight, high performance under impact and flexural stresses, high comfort, and high durability are mandatory requirements. Safety shoes are the most exigent in terms of bonding because they have to work in the presence of solvents, high or low temperature, high-impact resistance, and all these should last for a long time. There are not many reviews in the existing literature dealing with the adhesive bonding in shoe industry. Some Spanish books (Diputación de Alicante 1974; AmatAmer 1999) devoted to shoe industry include some basic technical chapters dealing with the use of the adhesives for bonding operation, but they have not been translated to English. In the 1960s and 1970s, the Prüf- und Forschungsinstitut Pirmasens (PFI) in Pirmasens (Germany) (Fisher and Meuser 1964; Fisher 1971) and the Shoe Allied Trade Research Association (SATRA) in Kettering (England) (Pettit and Carter 1973) published excellent reports dealing with several specific aspects of the adhesive joints in shoe industry. More extensive literature dealing with the solvent- and waterborne polyurethane adhesives used in shoe industry has been published (some to be cited are Pettit and Carter 1973; Schollenberger 1977; Dollhausen 1985; Kozakiewicz 1991; Sultan Nasar et al. 1998; Frisch 2002), but minor attention has been paid to polychloroprene adhesives (some to be cited are Whitehouse 1986; Kelly and McDonald 1963; Harrington 1990; Guggenberger 1990; Lyons and Christell 1997; Martín-Martínez 2002). On the other hand, the surface preparation procedures of the upper and sole materials for shoe bonding have been widely considered in the literature (some to be cited are Oldfield and Symes 1983; Extrand and Gent 1987; Hace et al. 1990; Fernández-García et al. 1991; Lawson et al. 1996; Cepeda-Jiménez et al. 2002). Most of the books devoted to footwear consider only a chapter devoted to adhesives or surface preparation of materials before bond formation, but rarely both aspects are described in the existing literature, and the understanding of the complex nature of the bonding process in shoe manufacturing is almost absent. A previous chapter written by the author dealing with shoe bonding was published in 2005 (Adams 2005), and the present chapter is an update of that chapter in which the

1486

J. M. Martín-Martínez

most recent scientific contributions have been addressed. Recently, a review dealing with the use of adhesives in the footwear industry has been published (Paiva et al. 2016). Obviously, the reader will find some similarities between the two chapters wroten by the author, but as a comprehensive updated view on shoe bonding is given, the reading of this chapter will certainly compensate for it. This chapter is divided into four sections. The first section provides an overview of the bonding in the shoe industry. Conventional surface preparation of different sole and upper materials is described in section two, with special attention to the chlorination of rubber soles by using organic solvent solutions of trichloroisocyanuric acid. Both solvent-borne and waterborne polyurethane and polychloroprene adhesives are described in detail in section three with special reference to their formulations. Finally, the testing, the quality control, and the durability of the adhesive joints in shoe bonding are considered in section four.

51.2

Overview of the Shoe Bonding Process

Although each shoe may be constituted by materials of different nature with different shapes and the adhesives commonly used are polyurethane or polychloroprene adhesives, to produce an adequate bonding, the different operations/steps given in Fig. 2 must be followed. The first operation consists in the surface preparation of the materials to be bonded (upper and sole). Irrespective of the inherent difficulty of bonding these materials, solvent cleaning and roughening are generally carried out. For difficult-tobond materials, more aggressive chemical surface treatments and/or primer application are mandatory. Once the materials are ready for bonding, the adhesive is Surface activation (cleaning, surface treatment)

Material (upper, sole)

Adhesive application (on both material surfaces) Adhesive (polyurethane, polychloroprene)

Preparation of solution (viscosity, solids content)

Joint formation

Testing (72 h after joint formation)

Reactivation process (heat + pressure)

Drying of adhesive

Ageing (50ºC / 95% RH / 1 week)

Fig. 2 Schematic overview of the bonding process in shoe manufacturing (upper-to-sole attachment)

51

Shoe Industry

1487

prepared by adjusting its viscosity and by adding isocyanate hardener, mainly for improving the durability of the adhesive joints. The application of the adhesive is a critical step to reach good bonding as the method of application (brush, roller, etc.), and the amount of adhesive determine the success of the joint; the adhesive is usually applied to both materials to be bonded, and, therefore, contact adhesives are used (reactive polyurethane hot melts are an exception because of they are applied on the sole only). The next step in the joint formation consists in the evaporation of the most solvent of the adhesive on the surfaces to be bonded, and because of the absence of tack, reactivation (i.e., sudden heating with infrared or halogen lamps) of the adhesive layers is carried out followed by immediate joining of the melted adhesive films applied on the two materials to be bonded followed by the application of pressure for a given time. Finally, the joints are cooling down before placing the shoe in the box. Interestingly, green adhesion is a critical issue for shoe manufacturers. It is a general practice to place the thumb nail at the shoe front and try to separate the adhesive joint. If the joint resists the press of the thumb nail, the joint is considered acceptable. If it is not the case, sewing is a common practice! In terms of the mechanisms of adhesion (see ▶ Chap. 2, “Theories of Fundamental Adhesion”) involved in upper-to-sole bonding, mechanical interlocking is important. Considering that most of the upper materials are porous and the adhesives used in shoe bonding are applied as liquids, an acceptable penetration of the adhesive into the upper is expected, and then the mechanical adhesion is generally favored. However, because several synthetic sole materials are nonporous and have in general a relatively low surface energy, both mechanical (e.g., roughening) and chemical (e.g., halogenation) surface preparations are necessary, and enhanced chemical adhesion is required. Therefore, the main mechanisms of adhesion involved in shoe bonding are mechanical interlocking and chemical. Furthermore, in the bonding of some rubber soles, the weak boundary layers produced by surface contaminants and/or by migration of antiadherend moieties to the interface during joint formation must be removed before joint formation. In the following sections, the materials and their surface preparation prior to be bonded, the adhesives used, and the manufacture and testing of the upper-to-sole joints will be separately discussed. Each section will consider an overview of the current state of the art in shoe industry, and future trends will be highlighted too.

51.3

Surface Preparation of Upper and Sole Materials for Bonding

Depending on fashion, each year different materials have been and are currently used in the manufacturing of shoes, including the rubber soles (vulcanized styrene–butadiene rubber (SBR), thermoplastic rubber, ethylene propylene diene monomer (M-class) rubber (EPDM)) and different polymeric materials (leather; polyurethanes; ethylene–vinyl acetate copolymers, EVA; polyvinyl chloride, PVC; polyethylene, Phylon). In order to produce adequate adhesive joints, the surface

1488

J. M. Martín-Martínez

preparation of those materials is required (see part related to Surface Treatments). Surface preparation procedures for these materials must be quickly developed, and the validity of these treatments is generally too short because of fashion constrictions. Several procedures have been established to optimize the upper-to-sole bonding; the most of them are based in the use of organic solvents. Due to environmental and health issues, the solvents should be removed from the surface preparation procedure, and several environmental-friendly procedures for the surface preparation of different materials have been proposed. Most of the upper and sole materials used in shoe industry cannot be directly joined by using the current adhesives (polyurethane and polychloroprene adhesives) due to their intrinsic low surface energy, the presence of contaminants and release agents, and antiadherend moieties on the surface. Therefore, the surface preparation of upper and sole must remove contaminants and weak boundary layers, and roughness and chemical functionalities able to produce adequate bond strength should be created. A review on the formulation and characteristics of the main materials used in shoe manufacturing (except adhesives) was published by Guidetti et al. 1992. To the author’s best knowledge, no additional reviews on the subject have been published later. In this section, a short review of the main materials used in shoe bonding as well as their current surface preparation procedures will be considered. In the last part of this section, the current and alternative surface treatments for upper and sole materials are described.

51.3.1 Upper Materials Leather Full-chrome, semi-chrome, and vegetable-tanned leathers are still the main source and most common upper material in shoe manufacturing. In general, the porous nature of the leather facilitates its bonding with adhesives in solution, mainly solvent-borne polychloroprene adhesives. To obtain good adhesion, the weak grain layer of the leather must be removed by using rotating wire brushes (roughening) to expose the corium (the part of the leather with higher cohesion of the collagen fibers) to the adhesive (SATRA 1963). In general, deeper roughening of upper leather produces a stronger bond than light roughening. Furthermore, two consecutive adhesive layers are generally applied on the roughened leather. First, a low-viscosity adhesive solution (incorrectly called primer) is applied to fill the pore entrances of the corium fibers and to facilitate the wettability of the leather surface. At least 5 min later (or alternatively when the adhesive layer becomes dried), a second high-viscosity adhesive solution of the same nature than the previous one is applied; diffusion of the polymer chains between the two adhesive layers is produced, and a homogeneous adhesive layer with high cohesion is created on the upper leather surface.

51

Shoe Industry

1489

The main issues that make difficult the bonding of upper leather are the existence of finishing on the surface to be bonded and the presence of high grease content. Deep or slight scouring or roughening is sufficient to remove the finishing and the weak grain leather. However, in thin leathers (such as buffalo or calf leathers), mild surface scouring followed by application of a primer is necessary to avoid leather tearing; the primer must increase the cohesion of the collagen fibers, and also it has to be compatible with the adhesive. Several primers for thin leathers have been proposed (Ferrandiz-Gómez et al. 1994), but they generally require the application of moderate temperature for being effective. Furthermore, different low-viscosity reactive solvent-based polyurethane primers containing about 10 wt% free isocyanate groups have been developed for improving the adhesion of difficult-to-bond leather (Vélez-Pagés 2003). Depending on the amount (more than 11 wt% is critical) and type of grease (mainly unsaturated fatty acids) in the leather, poor adhesive bond can be produced. The fat may diffuse through to the surface of the polyurethane adhesive film (possibly while the adhesive still wet, with the solvent acting as the vehicle) and interfere with its subsequent coalescence with the adhesive or the sole. When polychloroprene adhesive is used, the reaction of the fatty acids with some components in the adhesive formulation (zinc oxide, magnesium oxide) forms insoluble metal soaps (salts of fatty acids) at the interface that impart antiadherend properties. Polyurethane adhesives are more tolerant to greasy leather than polychloroprene adhesives. The roughening of the leather is always necessary, and the addition of 5 wt% isocyanate hardener to the adhesive just before application helps to reach good adhesive joints. Solvent-borne polychloroprene adhesives are excellent for bonding leather because of their high wettability and permanent tack (heat activation is not needed). However, these adhesives have poor performance with most of the rubber soles and with several synthetic upper materials. In these cases, polyurethane adhesives provide better performance.

Synthetic Upper Materials Several synthetic upper materials such as canvas, textiles, nylon, PVC on woven or nonwoven, and poromerics are used in shoe manufacturing. As for leather, most of these materials need roughening (to remove finishing) and/or solvent wiping, followed by application of two consecutive adhesive layers, to produce good adhesive joints. PVC uppers may cause lack of adhesion due to the migration of plasticizer into any material with which the PVC comes in contact. Light roughening is necessary, or alternatively methyl ethyl ketone (MEK) wiping can be applied followed by the application of polyurethane adhesive containing 5 wt% isocyanate hardener. Alternatively, the application of a primer based on nitrile rubber followed by application of two-component polychloroprene adhesive may perform adequately. PVC-coated fabrics may have acrylic or urethane coatings on top. Acrylic coatings are readily soluble by MEK wiping, whereas the urethane coatings need tetrahydrofuran wiping.

1490

J. M. Martín-Martínez

Polyurethane-coated fabrics can be bonded by using polyurethane adhesives. Mild scouring using a fine rotary wire brush must be used to completely remove the coating before applying the adhesive. Poromerics (nonwoven fabric supporting a microporous urethane layer) also need to be roughened with extreme care. When nylon, polyester, or cellulose interlayers are present, the application of isocyanate-based primers promotes the adhesion. Nylon fabric uppers show serious adhesion problems. The best bonding is obtained by mild roughening followed by applying 2 wt% isocyanate-based primer; the application of two-component polyurethane adhesive is also necessary.

51.3.2 Sole Materials Several kinds of sole materials are currently used in shoe manufacturing, the rubber soles are the most common. Few soles are made of natural products, such as leather or cork. The leather soles are not difficult to bond because they are porous and easily wetted by adhesive solutions. A primer is required to produce good bond with solvent-borne polychloroprene adhesives. The use of polyurethane adhesives is also feasible but not so effective, and there is the need of the previous application of low-viscosity polyurethane primer to which 2.5 wt% isocyanate hardener must be added. On the other hand, the cork soles are widely used for sandals, and they are joined with polychloroprene adhesives. In some cases, a thin EVA midsole is attached to the cork to improve its abrasion resistance and comfortability by using two-component polychloroprene adhesive (Martin 1971).

Rubber Sole Materials Rubber soles are by far the most common in shoe industry. Vulcanized or unvulcanized (thermoplastic) rubber soles are generally used. In general, the bonding is produced with one- or two-component polyurethane adhesives, and a surface treatment is always required to produce good adhesive joints. Thermoplastic Rubber (TR) Soles Styrene–butadiene–styrene (SBS) block copolymers are adequate raw materials to produce thermoplastic rubber (TR) soles. SBS block copolymers contain butadiene domains – soft and elastic – and styrene domains – hard and tough. Because the styrene domains act as virtual cross-linking in the SBS structure, the vulcanization is not necessary to provide dimensional stability. The TR soles generally contain polystyrene (to impart hardness), plasticizers, fillers, and antioxidants, and processing oils can also be added. The TR soles have low surface energy, so a surface modification is needed to reach proper adhesion to polyurethane adhesive. Special adhesives for the TR soles have been developed to avoid surface preparation, but they have poor creep resistance. Due to their softness, the TR soles cannot be roughened. The TR soles tend to swell by application of adequate solvents, and the mechanical interlocking of the adhesive is favored; however, vigorous rubbing must be avoided to maintain the mechanical integrity of the sole. Chemical treatments are necessary to improve the

51

Shoe Industry

1491

adhesion of the TR to polyurethane adhesive, mainly cyclization (treatment with sulfuric acid) and halogenation. The TR can be treated with concentrated sulfuric acid to yield a cyclized layer on the surface (Cepeda-Jiménez et al. 2001). This layer is quite brittle, and when it is flexed and stretched, microcracks are developed which help in the subsequent bonding by favoring the mechanical interlocking of the polyurethane adhesive with the TR. The thickness of the cyclized layer depends on the length of the treatment with sulfuric acid. On the other hand, the treatment with sulfuric acid mainly produces the creation of highly conjugated C–C bonds and the sulfonation of the butadiene units of the TR material. The immersion of TR for less than 1 min, the neutralization with ammonium hydroxide (it extracts the hydrogen from the sulfonic acid and leaves stabilized SO3
NH4+ ion pair), and the high concentration of the sulfuric acid (95 wt%) are essential to produce an adequate surface treatment. The treatment with H2SO4 increases the T-peel strength of surface-treated TR/polyurethane adhesive joints (Fig. 3). On the other hand, the styrene content in the TR determines the effectiveness of the treatment with sulfuric acid. The lower is the styrene content in the TR, the more noticeable are the modifications produced on the surface. Furthermore, the styrene content in the TR affects the extent but not the nature of the surface modifications produced by treatment with sulfuric acid. Halogenation is the most common surface treatment for TR. The use of the chlorination surface treatment in improving the adhesion of several types and formulations of rubbers has been extensively demonstrated, and it is cheap and easy to apply. Furthermore, the chlorination surface treatment removes contaminants and antiadherend moieties from the rubber surface and imparts improved durability and aging resistance to the upper-to-sole joints, and the treated surfaces remain 8

T-peel strength (kN/m)

7 6 5 4 3

15 min 72 h

2 1 0 0

1

2

3

4

5

6

ti (min)

Fig. 3 T-peel strength values (15 min and 72 h after joint formation) of sulfuric acid-treated TR rubber/polyurethane adhesive joints as a function of the immersion time (Cepeda-Jiménez et al. 2001)

1492 Fig. 4 Chemical structures of DCI (sodium dichloroisocyanurate) and TCI (trichloroisocyanuric acid)

J. M. Martín-Martínez CI O C

N

N CI

C

CI O C

O C

N

N CI

C C

N CI

O

NaO DCI

Fig. 5 Chemical structure of chloramine T

O

N

TCI

R

R = H, CH3 X = H, Cl, Br

Cl (Br) SO2N X

reactive with the polyurethane adhesive for at least 3 months after surface preparation. Immersion in aqueous chlorine or in sodium hypochlorite aqueous solution is a very effective surface preparation for TR soles although fast evolution of chlorine is produced. Therefore, the organic chlorine donors for the treatment of the TR have been proposed alternatively for allowing moderate evolution of chlorinated species during surface treatment. The most commonly used chlorinating agent is an organic solvent (ketone or ester) solution of trichloroisocyanuric acid (TCI) – 1,3,5-trichloro-1,3,5-triazin-2,4,6-trione – (Fig. 4) which is very cheap and common (it is generally used as chlorinating agent in swimming pools). The organic solvent in the TCI solutions determines the degree of rubber wetting, and furthermore, the actual chlorinating species are produced by reaction of the TCI with the organic solvent. The use of the TCI solutions is quite satisfactory for the TR surface preparation although it should be applied with a soft brush and left to dry during moderate time to avoid degradation of the TR surface by the solvent (Carter 1971). For avoiding the use of organic solvents in the chlorination agents, acidified sodium hypochlorite aqueous solutions containing 1-octyl-2-pyrrolidone as wetting agent have been successfully used for the treatment of the TR soles (Cepeda-Jiménez et al. 2003a). The treatment is restricted to about 1 μm depth surface of the TR, and the enhanced adhesion is due to the improved wettability and the creation of chlorine moieties on the rubber surface. Furthermore, aqueous solutions of sodium dichloroisocyanurate (DCI) (Fig. 4) have been used to increase the adhesion of the TR soles (Cepeda-Jiménez et al. 2002). The chemical structure of DCI is somewhat similar to that of TCI, and relatively concentrated DCI water solutions are necessary to obtain good peel strength values. On the other hand, successful surface chlorination treatment of the TR has been obtained by using aqueous N-chloro-p-toluenesulfonamide solutions (obtained by acidifying chloramine T solutions) (Cepeda-Jiménez et al. 2003a) – Fig. 5. The actual chlorinating species is SO2-C6H4-NCl-Cl. The surface treatment is carried out by immersion of the TR in

51

Shoe Industry

1493

the aqueous chlorinating solution at 20–80  C for about 1 min. The effectiveness of this treatment was ascribed to the introduction of chlorine and oxygen moieties on the rubber surface. Electrochemical treatments have also been successfully proposed to improve the adhesion of the TR (Brewis et al. 1997). The treatment consists in the immersion of the TR in an electrochemical cell of silver nitrate in 3.25 M nitric acid. After treatment, the rubbers have to be washed with nitric acid and then with bidistilled deionized water until neutral pH is reached. Several environmental-friendly surface preparations of the TR sole materials by means of radiations have been more recently proposed. These treatments are clean (no chemical or reaction by-products are produced) and fast, and furthermore online bonding at shoe factory can be produced, so the future trend in the surface modification of the rubber soles will be likely directed to the industrial application of those treatments. Corona discharge, low-pressure radio-frequency (RF) gas plasma, and ultraviolet (UV) enriched in ozone – UV-ozone – treatments have been successfully used at laboratory scale to improve the adhesion of several TR sole materials. Corona discharge has been used to improve the adhesion of the TR to polyurethane adhesives (Romero-Sánchez et al. 2003a). Although the treatment with corona discharge improved the wettability of the TR due to the formation of polar moieties on the surface, and surface cleaning and removal of contaminants (mainly demolding silicone moieties) were produced, the peel strength values moderately increased. Low-pressure RF gas plasma has been shown effective to enhance the adhesion of the TR to different polyurethane adhesives (Ortiz-Magán et al. 2001). Different gases (oxygen, nitrogen, and oxygen–nitrogen mixtures) were used to generate the RF plasma. The treatment produced the partial removal of hydrocarbon moieties from the TR surface, and the formation of oxygen moieties and surface roughness was also produced. On the other hand, the treatment of SBS rubber with low-pressure RF plasma in CCl4 has also been proposed for improving its adhesion to polyurethane adhesive, and a drastic increase in the peel strength was observed after only a few seconds of treatment which was ascribed to the chemical bonding between the carbon–oxygen moiety species on the SBS surface and the isocyanate groups of the polyurethane adhesive (Tyczkowski et al. 2009). Water molecules strongly attached to the SBS surface were also found, and they reduced the adhesion by partial blocking of the functional groups against the action of isocyanates. On the other hand, because no correlation between the peel strength and the surface free energy was observed, the thermodynamic adhesion was unimportant, and the dominant role of the chemical adhesion was confirmed. The surface treatment of the TR with UV–ozone produced with low-pressure mercury vapor lamp enriched in 175 nm wavelength has been shown to successfully increase its adhesion to polyurethane adhesive (Romero-Sánchez et al. 2003b). The UV–ozone treatment produced important surface modifications (improved wettability, roughness) incorporating oxygen and nitrogen moieties at the TR surface. The peel strength values increased after UV–ozone treatment of the TR, in a greater extent by increasing the treatment time (Fig. 6). More recently, several SBS rubbers

1494

J. M. Martín-Martínez 12

T-peel strength (kN/m)

10 8 6 Before ageing 4

After ageing

2 0 0

5

10

15

20

25

30

Treatment time (min)

Fig. 6 T-peel strength values of UV-ozone treated TR/polyurethane adhesive joints as a function of the UV treatment time (Romero-Sánchez et al. 2003)

containing different amounts of calcium carbonate and/or silica fillers were surface treated with UV–ozone to improve their adhesion to polyurethane adhesive (RomeroSánchez and Martín-Martínez 2008). The nature and content of fillers determined the extent of surface modification and adhesion of SBS rubber treated with UV–ozone. The UV–ozone treatment improved the wettability of all rubber surfaces, and chemical (oxidation) and morphological modifications (roughness, ablation, surface melting) were produced. The increase of the UV–ozone treatment to 30 min led to surface cleaning due to ablation and/or melting of the most external rubber layers, and also incorporation of more oxidized moieties was produced. As a consequence of these surface modifications, the UV–ozone treatment increased the adhesive strength in all joints made with polyurethane adhesive, more noticeably in the joints made with the SBS rubbers containing silica filler. On the other hand, although the chemical modifications were produced earlier in unfilled rubber for short time of treatment with UV–ozone, they were more noticeable in silica-filled rubbers with low silica content for extended length of treatment. However, the oxidation process seemed to be inhibited for the SBS rubbers containing calcium carbonate filler. Vulcanized NR and SBR Soles Although natural rubber (crepe) and nitrile rubber soles (common in the manufacturing of safety shoes due to its chemical inertness) are used, sulfur-vulcanized natural rubbers and styrene–butadiene rubbers (SBR) are the most common sole materials. The vulcanization system contains sulfur and activators (N-cyclohexyl-2benzothiazole sulfenamide, dibenzothiazyl disulfide, hexamethylenetetramine, zinc oxide), and fillers (silica and/or carbon black, calcium carbonate) are added to control hardness and abrasion resistance. Antioxidants (zinc stearate, phenolic

51

Shoe Industry

1495

antioxidant) and antiozonant (microcrystalline paraffin wax) are also necessary to avoid degradation processes and early aging. In general, vulcanization is carried out in a metal mold placed in hot plate press at about 180  C for allowing the reaction of the zinc oxide and the stearic acid to produce zinc stearate, which is combined with hexamethylenetetramine to form an unstable complex. After vulcanization, the contact with moisture or different solvents apparently causes the breakdown of this complex with the appearance of the zinc stearate that migrates to the vulcanized NR or SBR sole which acts as an antiadherend material to polyurethane adhesive (Pettit and Carter 1964). On the other hand, during cooling after vulcanization, the paraffin wax migrates to the rubber surface also producing adhesion problems. The causes of poor adhesion in vulcanized NR and SBR soles are quite diverse. Over-vulcanized external rubber layer is too stiff to support peel strength stresses. The presence of silicone release agents may reduce the bond strength of the vulcanized SBR sole joints. The presence of antiadherend moieties on their surfaces, their low surface energy, and the migration of low-molecular moieties (antiozonant wax, processing oils) to the interface once the adhesive joint is produced are other causes for the poor adhesion in vulcanized NR and SBR. The paraffin wax and zinc stearate are concentrated in a layer covering the vulcanized rubber surface with a thickness of about 2.0 μm. The factors favoring or inhibiting the migration of antiozonant wax to the vulcanized NR and SBR surfaces have not been fully understood. The temperature is one key factor that determines the migration of the antiozonant paraffin wax to the rubber surface. The antiozonant migration with the temperature in vulcanized rubber has been recently studied (Torregrosa-Coque et al. 2012) by heating at 40–90  C for 15 h in an oven. The heating of vulcanized SBR at different temperatures caused a partial removal of paraffin wax from the surface, and at a temperature close to the paraffin wax melting point, the crystals of paraffin wax on the vulcanized SBR surface were melted, causing a decrease in the thickness of the paraffin wax film on the surface. Furthermore, the migration of zinc stearate to the paraffin wax layer on the vulcanized SBR surface occurred by heating at 90  C, and even after heating at 90  C, a thin film of paraffin wax always remained on the vulcanized SBR surface. Time is another key factor that determines the migration of the antiozonant paraffin wax to the rubber surface causing the formation of a weak boundary layer of nonrubber contaminants which is deleterious for adhesion of rubber to polyurethane adhesives (Torregrosa-Coque et al. 2011a). Figure 7 shows the ATR-IR spectra of polyurethane coating on MEK surface cleaned vulcanized SBR just after application and after 2 weeks. The increase of the time favored the migration of paraffin wax which is noticed by the appearance of the methylene bands in the ATR-IR spectra of the vulcanized SBR at 2848, 2915, and 2950 cm
1 on the polyurethane coating. One of the key steps in the manufacturing of rubber/adhesive joints is reactivation, i.e., sudden heating of the thin adhesive layers on the substrates to be joined under infrared (IR) radiation to 80–90  C for a few seconds to allow diffusion of the polymeric chains under pressure. This reactivation may cause the migration of lowmolecular-weight additives to the rubber surface causing a lack of adhesion. The influence of the reactivation temperature (40–90  C) on the surface properties of sulfur-vulcanized styrene–butadiene rubber on the extent of the diffusion of the

1496

J. M. Martín-Martínez

Fig. 7 ATR-IR spectra obtained with ZnSe prism of polyurethane coating on MEK surface cleaned vulcanized SBR for different time after polyurethane coating deposition. 3100–2700 cm
1 region (Torregrosa-Coque et al. 2011b)

paraffin wax and zinc stearate to the rubber surface has been studied (TorregrosaCoque et al. 2011a). The reactivation of the rubber at different temperatures produced changes in the morphology and thickness of the paraffin wax layer on the surface. By increasing the reactivation temperature, a partial removal of paraffin wax was produced, and the thickness of the paraffin wax film on the rubber surface was reduced. On the other hand, the increase in the reactivation temperature increased the surface area of the melted paraffin wax layer that prevented migration from the rubber bulk, but a critical reactivation temperature at 90–100  C existed at which the migration of zinc stearate to the paraffin wax layer on the rubber surface was favored. Furthermore, it has been demonstrated that the paraffin wax migrated from the vulcanized rubber bulk to the polyurethane–rubber interface, and, then, it diffused across the polyurethane coating and migrated later to the polyurethane coating surface (TorregrosaCoque et al. 2011b). On the other hand, the extent of migration of the paraffin wax depended on the time after polyurethane adhesive deposition, and the faster migration was produced 1 h after applying the polyurethane adhesive. The application of surface treatments to NR and SBR soles should produce improved wettability; creation of polar moieties able to react with the polyurethane adhesive, cracks, and heterogeneities should be formed to facilitate the mechanical

51

Shoe Industry

1497

interlocking with the adhesive; and an efficient removal of antiadherend moieties (zinc stearate, paraffin wax, processing oils) has to be reached. Several types of surface preparation involving solvent wiping, mechanical and chemical treatments, and primers have been proposed to improve the adhesion of vulcanized NR and SBR soles. Chlorination with solutions of trichloroisocyanuric acid (TCI) in different organic solvents is by far the most common surface preparation for vulcanized NR and SBR. Roughening of the SBR soles removes the over-vulcanized layer and surface contaminants (oils, molding release agents, antiadherend moieties, etc.) and creates roughness. After roughening, residues of the treatment must be removed by solvent wiping and/or the application of pressurized air. Although the removal of weak layers and the creation of surface heterogeneities may produce improved adhesion of the vulcanized NR and SBR soles, an additional chemical treatment is mandatory for achieving sufficient adhesive strength. The soles can be chemically treated with concentrated sulfuric acid to improve their adhesion to polyurethane (Cepeda-Jiménez et al. 2000). This treatment is not restricted to the surface but also produces a bulk modification of the rubber because a decrease in the tensile strength and the elongation at break are caused. If the SBR contains paraffin wax, the treatment with sulfuric acid promotes its migration to the surface, and solvent wiping with petroleum ether is mandatory for obtaining good adhesion. Several chlorinating agents have been used for SBR sole bonding. Acidified sodium hypochlorite aqueous solutions have been used successfully (Table 1), and the mechanism of chlorination has been established (Vukov 1984) that consists in the formation of H2ClO+ species which produces the evolution of chlorine; then, the chlorine reacts with the double C=C bond of the rubber producing chlorinated hydrocarbon moieties on the surface and some cross-linking. The most commonly used chlorinating agent in the chemical treatment of the SBR sole is ethyl acetate or MEK (butan-2-one) solution of trichloroisocyanuric acid (TCI). The organic solvent in the TCI solution determines the degree of rubber wetting, and the actual chlorinating species are α-chloro ketones in TCI/MEK solutions, whereas acid chlorides are the reactive moieties in TCI/ethyl acetate solutions (Pastor-Blas et al. 2000). The reaction of these chlorinating species with the double C=C bonds of the rubber produces C–Cl species, and the reaction is not restricted to the surface, but it is penetrating about 100 μm depth. On the other hand,

Table 1 T-peel strength values (kN/m) of surface-chlorinated rubber/polyurethane adhesive/ leather joints Rubber SBR 1 SBR 2 SBR 3

As-received 1.8 (A) 0.2 (A) 0.9 (A)

A: Adhesion failure between the rubber and the adhesive R: Cohesive rupture of the rubber

NaClO + HCl 5.6 (R) 14.0 (R) 7.8 (R)

1498

J. M. Martín-Martínez

the improved adhesion of the SBR soles treated with TCI solutions is due to the contribution of several mechanisms of adhesion: mechanical adhesion (cracks and pits are produced on the surface), chemical adhesion (chlorinated hydrocarbon, C=O moieties), removal of weak boundary layers (zinc stearate, paraffin wax), and thermodynamic adhesion (a decrease in contact angle is produced, i.e., improved wettability) (see ▶ Chap. 6, “Thermodynamics of Adhesion”). Furthermore, unreacted solid prismatic TCI crystals deposit during the application of the chlorination treatment on the SBR surface which can be dissolved in contact with the organic solvent in the polyurethane adhesive facilitating the creation of in situ reactive species. The effectiveness of the chlorination treatment of the SBR soles with TCI solutions strongly depends on several experimental variables of the treatment (Martín-Martínez 2008). The most relevant ones are the application procedure of the TCI solution (brushing providing the best performance), the time between the application of the TCI solution and the polyurethane adhesive (a minimum of 10 min is necessary), the TCI concentration in the solution (less than 3 wt% TCI is recommended), and the roughening or solvent wiping before chlorination (more extensive treatment is produced). The active chlorine content of the ethyl acetate solutions containing 3 wt% TCI determined the extent of the surface treatment of vulcanized SBR. The active chlorine concentration in the TCI solutions was not the only parameter determining the adhesion of vulcanized SBR, as the highest adhesive strength was not achieved by treating the rubber with the TCI solution with higher active chlorine content (García-Martín et al. 2010). Furthermore, the chlorine concentration in the TCI solutions was not stable in the course of time; the increase in time between TCI solutions preparation and SBR treatment allowed an increase in adhesive strength, the highest value corresponded to the joint produced with the vulcanized SBR treated with the TCI solution prepared for 60 days. The influence of the temperature on the effectiveness of the chlorination treatment of vulcanized SBR with TCI solutions in organic solvents has been addressed. Vulcanized SBR was treated with 1 and 2 wt% TCI solutions in ethyl acetate and then thermally treated at 23  C, 50  C, or 65  C for different time (Yin et al. 2012). Although this study was not intended for shoe industry, the results obtained revealed that the increase of the chlorination temperature up to 50  C was very effective for SBR surface modification by TCI, leading to enhanced surface wettability, creation of chlorinated hydrocarbon moieties, and increased shear strength (from 2.0 to 9.7 MPa). However, later study (Yañez-Pacios et al. 2013) has shown that the application of temperature during chlorination of high antiozonant content vulcanized NR with TCI solutions in MEK did not increase the adhesion strength of the upper-to-sole joints. However, changes in wettability, surface chemistry, and roughness of the rubber were produced, all being irrelevant for improving adhesion likely due to the migration of antiozonants to the polyurethane adhesive-vulcanized NR interface with time after joint formation. Due to organic solvent restrictions in the industry, aqueous chlorinating solutions have been tested as an alternative to TCI solutions in ethyl acetate or MEK. Aqueous

51

Shoe Industry

1499

solutions of sodium dichloroisocyanurate (DCI) have been recently demonstrated to increase the adhesion of vulcanized SBR soles (Cepeda-Jiménez et al. 2002). The surface treatment with aqueous DCI solutions creates C–Cl moieties, removes antiadherend moieties from the SBR surface, and creates roughness. The use of a low DCI concentration in water is sufficient to obtain good peel strength values in SBR/waterborne polyurethane joints. Furthermore, vulcanized NR has been treated with aqueous sodium hypochlorite solution for various chlorination times up to 30 min (Radabutra et al. 2009). The degree of chlorination increased with the chlorination time up to 10 min then leveled off, but the roughness and stiffness increased gradually with chlorination time. The maximum peel strength was found at 1 min of chlorination time and decreased afterward because of the increase in the surface stiffness that acted as weak boundary layer. As for TR, successful treatments of vulcanized NR and SBR soles with aqueous N-chloro-p-toluenesulfonamide solutions and electrochemical treatments have been reported at laboratory scale. Similarly, the treatment of SBR sole with oxidant inorganic salts (acidified potassium dichromate, potassium permanganate, Fenton’s reactive) has been shown to be partially successful (Brewis and Dahm 2003). Primers have also been used to improve the adhesion of SBR soles. Due to its high reactivity, isocyanate wipe has been proposed, but the most common treatment uses organic acid solutions. Solutions of lactic acid applied with stiff-bristled brush and with a scrubbing action are adequate to remove metal soap contamination (i.e., zinc stearate) on the surface of SBR sole and even replace roughening (Carter 1969). Solutions of different carboxylic acids (fumaric acid, maleic acid, acrylic acid, succinic acid, malonic acid) in ethanol have also been used as primers of SBR sole (Pastor-Sempere et al. 1995). The nature of the carboxylic acid determines the rate of diffusion into the polyurethane adhesive and the extent of rubber-adhesive interfacial interaction. Finally, mixtures of TCI and fumaric acid solutions have been tested in improving the adhesion of difficult-to-bond SBR sole, but it was demonstrated that the effectiveness of this treatment was mainly due to chlorination by TCI (RomeroSánchez and Martín-Martínez 2003). Treatment with supercritical fluids was also tested to remove zinc stearate and waxes from vulcanized SBR soles (Garelik-Rosen S, Torregrosa-Maciá R, MartínMartínez JM (2003), unpublished results). The relative effectiveness of this treatment was ascribed to the dissolution of antiadherend moieties by penetration of the fluid into the rubber surface. An alternative to water-based treatment is the use of low-pressure RF gas plasmas in the enhancement of the adhesion of vulcanized rubbers. The treatment in oxygen plasma for 1 min is enough to noticeably increase the adhesion of vulcanized SBR sole to polyurethane adhesive (Pastor-Blas et al. 1998). Later, three different configurations (direct, etching, and secondary downstream) of low-pressure RF oxygen plasmas for length of treatment between 1 and 10 min were used to modify the surface of vulcanized SBR (Torregrosa-Coque and Martín-Martínez 2011c). The direct oxygen plasma was the most aggressive treatment and the secondary downstream plasma the less one, and the oxygen plasma treatment caused surface oxidation and ablation on the SBR surface, as well as an increase of the temperature

1500

J. M. Martín-Martínez

that also determined the extent of paraffin wax migration which was produced for at least 24 h after oxygen plasma treatment; as a consequence, poor adhesion to polyurethane adhesive was obtained, due to the creation of a weak boundary layer of paraffin wax at the rubber–polyurethane interface. Vulcanized ethylene propylene diene polymethylene (EPDM) rubber surface was treated with low-pressure RF argon–oxygen plasma to improve adhesion with compounded natural rubber (NR) during co-vulcanization (Basaka et al. 2011). The plasma treatment changed both surface composition (formation of C–O and –C=O functional groups) and roughness of EPDM rubber, and consequently increased peel strength was obtained. In a different study, vulcanized NR containing intentionally noticeable excess of processing oils in its formulation was treated with argon–oxygen (Ar–O2) (2:1, vol/vol) low-pressure plasma for achieving a satisfactory level of adhesion to waterborne polyurethane adhesive (Cantos-Delegido and Martín-Martínez 2015). The effectiveness of the Ar–O2 low-pressure plasma treatment depended on both the configuration of the plasma chamber shelves and the treatment time, the direct configuration provided the most effective surface modification of the vulcanized NR. However, even important surface modifications were produced by Ar–O2 low-pressure plasma treatment, adhesion was not improved due to the creation of weak boundary layer at the polyurethane–rubber interface after joint formation. Heating at 80  C for 12 h prior to Ar–O2 low-pressure plasma treatment enhanced the extent of the surface modifications, and improved adhesion was obtained for treatment times higher than 600 s. Atmospheric plasma torch treatment of vulcanized NR with different excess of paraffin wax and processing oils was carried out for improving adhesion to waterborne polyurethane adhesive (Kotrade et al. 2011). Similar chemical modifications (new C=O and C–N functional groups) and roughness were produced by plasma torch treatment in all rubbers, and the best results were obtained at a rubber surface nozzle of the plasma torch distance of 5 mm and a platform speed of 1 m/min. However, although the rubber surfaces were effectively modified by plasma torch, the adhesion to waterborne polyurethane adhesive was not improved. In a different study (Moreno-Couramjou et al. 2009), vulcanized NR was treated by atmospheric dielectric barrier discharges to improve its adhesion to silicone adhesive. The atmospheric plasma treatment with the use of allyl alcohol improved noticeably the adhesion by a factor of 10 due to the preferential formation of C–O bonds. More recently the treatment with UV radiation combined with ozone (UV–ozone) in improving the adhesion of difficult-to-bond vulcanized NR (it contained an excess of processing oils) to waterborne polyurethane adhesive in footwear has been carried out (Moyano and Martín-Martínez 2014). Both the length of the treatment and the distance between the UV radiation source to the rubber surface were varied, and the effects of the treatment on the surface chemistry, wettability and surface energy, and topography of the rubber were analyzed. The treatment of the vulcanized rubber with UV–ozone removed hydrocarbon moieties and zinc stearate from the surface, surface oxidation (C–O, C=O, and COO- groups formation) occurred, and improved wettability and increased surface energy (mainly due to the polar component) were obtained. The UV–ozone surface treatment also caused heating of the surface and

51

Shoe Industry

1501

increased the adhesion of the vulcanized rubber to waterborne polyurethane adhesive; the highest adhesion was obtained in the joints made with UV–ozone-treated rubber for 3 and 6 min at a UV radiation source-rubber surface distance of 5 cm.

Polymeric Soles Ethylene–vinyl acetate (EVA) block copolymer soles are common in sport and casual shoes due to their low density and easy coloring. EVA soles are difficult to bond with the polyurethane and polychloroprene adhesives commonly used in shoe industry, mainly due to their low surface energy. The higher the vinyl acetate content, the less difficult to bond the EVA sole is. The lightweight microcellular EVA soles can usually be adequately bonded after scouring using polyurethane and polychloroprene adhesives. The injection-molded EVA soles are more difficult to bond although roughening followed by application of two-component polychloroprene adhesive can give moderate bond strength (Martínez-García et al. 2003a). Corona discharge has been shown to be useful to improve the peel strength of injectionmolded EVA soles containing 12 and 20 wt% vinyl acetate/polychloroprene adhesive joints. The improvement is ascribed to the creation of C=O and R-COO
moieties. Treatment with sulfuric acid also provided improved adhesion of injection-molded EVAs with vinyl acetate contents between 9 and 20 wt%, because this treatment produces sulfonation and creation of oxygen moieties on EVA surface (Martínez-García et al. 2003b). Oxygen and other gases low-pressure plasmas are very effective in improving the peel strength of joints produced with EVA soles with different vinyl acetate contents and two-component polyurethane adhesive (CepedaJiménez et al. 2003b). The treatment is more effective with non-oxidizing plasmas, giving rise to a high roughness. Finally, the treatment with UV–ozone has been developed to enhance the bond strength between EVA sole and polyurethane adhesive (Landete-Ruiz and Martín-Martínez 2005). The surface modifications produced by UV–ozone treatment of two ethylene–vinyl acetate (EVA) copolymers containing 12 (EVA12) and 20 wt% (EVA20) vinyl acetate have been studied (Landete-Ruiz and Martín-Martínez 2015). The treatment with UV–ozone improved the wettability of both EVAs due to the creation of new carbon–oxygen moieties, and the extent of these modifications increased with increasing the time of treatment; the modifications produced in EVA20 were produced for shorter lengths of treatment. The UV–ozone treatment also created roughness and heterogeneities, and whereas roughness formation prevailed on the UV–ozone-treated EVA12, important ablation was dominant on the treated EVA20. T-peel strength values in joints made with polychloroprene adhesive increased. Short length of UV–ozone treatment (1 min) produced higher T-peel strength in joints made with EVA20, whereas higher T-peel strength values in joints made with EVA12 were obtained after treatment for 5–7.5 min in which a cohesive failure into a weak boundary layer on the treated EVA surface was found. Finally, the aging resistance of the treated EVA/polychloroprene adhesive joints was good, and the surface modifications on the UV–ozone-treated EVAs lasted for 24 h after treatment at least. Ethylene–vinyl acetate (EVA) copolymers intended for sport sole manufacturing may contain noticeable amounts of low-density polyethylene (LDPE) – Phylon-type

1502

J. M. Martín-Martínez

soles – for improving abrasion resistance and decreasing cost; however, the EVA–PE blend had low polarity and showed poor adhesion. Good results have been obtained by applying extensive organic solvent wiping, followed by application of an UV-activated primer and by using two-component polyurethane adhesive. More recently, an effective environmental-friendly and fast surface treatment based on UV–ozone has been used to increase the wettability, polarity, roughness, and adhesion of EVA–PE material to leather with waterborne polyurethane adhesive (Jofre-Reche and Martín-Martínez 2013). The more extended length of treatment and the shorter UV source–substrate distance increased the wettability and created new carbonyl groups mainly, and the amounts of hydroxyl and carboxylic groups were increased. The UV–ozone treatment produced ablation and etching of the EVA–PE material surface, mainly in the vinyl acetate of the EVA; this topography was also caused by heating during UV–ozone treatment. For low length of UV treatment or high UV source–material distance, the modifications of the EVA–PE material were mainly produced in the ethylene causing the selective removal of vinyl acetate, whereas more aggressive conditions produced strong oxidation in the EVA–PE material. Finally, adhesive strength was noticeably increased in the UV–ozone-treated EVA–PE/polyurethane adhesive joints, and a cohesive failure in the leather was obtained. Only polyurethane adhesives should be used to bond PVC soles. The adhesion problems of PVC derive from the presence of plasticizers and stabilizers (stearate type) able to migrate to the surface impeding the contact with the adhesive (creation of weak boundary layers). A solvent wiping on the PVC surface is usually effective in improving adhesion, and solvents such as MEK are adequate. On the other hand, treatment with 10 wt% aqueous solutions of sodium hydroxide has been successfully applied to increase the adhesion performance of plasticized PVC soles (Abbott et al. 2003). Polyurethane soles usually contain silicone mold release agents which prevent adhesion. To remove them, mild roughening following by solvent wiping is necessary. The application of an isocyanate primer is very useful before polyurethane adhesive is applied. Solvent-free treatments have been proposed for polyurethane soles, mainly cryoblasting (Abbott et al. 2003). Cryoblasting consists in the bombardment of the PVC soles with particles of solid carbon dioxide at pressures about 0.4 MN/m2. The impact of the solid particles successfully removed mold release agents from the polyurethane surface improving their adhesion to polyurethane adhesive.

51.3.3 Adhesives in Upper-to-Sole Bonding Contact adhesives based in one- and two-component polychloroprene (neoprene) and mainly polyurethane adhesives are the most commonly used in shoe industry to bond upper to sole (see ▶ Chaps. 14, “Adhesive Families,” and ▶ 12, “Classification of Adhesive and Sealant Materials”). These adhesives are bonded by autoadhesion which implies the application of the adhesive to both surfaces to be joined; diffusion of polymer chains must be achieved across the interface between the two adhesive

51

Shoe Industry

1503

films on the substrates to be joined to produce intimate adhesion at molecular level. To achieve an optimum diffusion of the polymer chains, both high wettability and adequate viscosity and rheology of the adhesive should be achieved. In the past, polychloroprene adhesives were more extensively used in upper-tosole bonding, but nowadays polyurethane adhesives are preferred. Polychloroprene adhesives have better tack and improved wettability than polyurethane adhesives, but the polychloroprene adhesives are not compatible with the surface-treated rubber soles and cannot be used to joint PVC soles. Therefore, polyurethane adhesives show higher versatility on broader range of substrates and also have lower oxidative degradation in time. However, they require always the surface preparation of the materials to be bonded.

51.3.4 Polyurethane Adhesives In 1960s, the solvent-borne polyurethane adhesives replaced the polychloroprene adhesives in the bonding of PVC soles for sport shoes. Today, the use of solventborne polyurethane adhesives is in decline due to environmental regulations and safety issues. Shoe industry was initially reluctant to the use of waterborne polyurethanes due to higher cost and reduced green strength. However, the use of waterborne polyurethanes is more common nowadays because their solid content is high and small amount of adhesive is needed to produce a similar joint than with solvent-borne adhesives. On the other hand, during the last decade, solvent-free polyurethane adhesives have been tested in shoe bonding. Although these adhesives have been widely used and effectively tested in the automotive industry, the advent in the shoe manufacturing is not as successful, likely because different machinery and bonding concepts need to be implanted. Shoe industry is, in general, quite reluctant to change bonding technologies, particularly when it implies a cost increase! In this section, the solvent-borne, the waterborne, and the solvent-free polyurethane adhesives will be considered separately.

Solvent-borne Polyurethane Adhesives Elastomeric thermoplastic polyurethanes are the main components in solvent solution adhesives for shoe industry. These polyurethanes are generally prepared in the form of pellets or chips by reacting an isocyanate (such as MDI – 4,40 -diphenylmethane diisocyanate), a long-chain diol (polyester or ε-caprolactone type), and a chain extender (glycol) to produce a linear polymer with negligible chain branching and relatively low-molecular-weight (Mw = 200.000–350.000 Da). The isocyanate to hydroxyl equivalent ratio (NCO/OH ratio) is usually kept near 1, thus producing a polymer with terminal hydroxyl groups which are able to form hydrogen bonds with the urethane moieties. From a polymer physics point of view, the configuration of the elastomeric polyurethanes corresponds to a segmented structure (block copolymer of (AB)n type) consisting of soft and hard segments (Fig. 8). Typically the soft segments are

1504

J. M. Martín-Martínez

Fig. 8 Scheme of the segmented structure of thermoplastic polyurethane

Hard segments

Soft segments

composed of a rubbery polymer (mainly the polyol), the glass transition temperature (Tg) which is located well below ambient temperature. The hard segments are generally produced by the reaction of the isocyanate and the short chain glycol (chain extender) and have a rigid and crystalline structure. The nonpolar low-melting soft segments are incompatible with the polar high-melting hard segments. As a result, phase separation (segregation) occurs in the polymer network. Typical elastomeric polyurethanes used as adhesives in upper-to-sole bonding have a relatively low content of hard segments, and their properties are mainly determined by the soft segments. Therefore, these polyurethanes will be elastic in the range of temperature between the glass transition temperature (generally located between
30 and
40  C) and the softening temperature of the elastomeric domains (50–80  C). The low melting point permits the elastomeric polyurethane to be softened at relatively low temperature, with sufficient thermoplasticity (i.e., loss of cohesion at moderate temperature) and surface tack to ensure correct bond. Solvent-borne polyurethane adhesives are generally prepared by dissolving the solid elastomeric polyurethane pellets or chips in a solvent mixture. Because polyurethanes have a linear molecular structure, the solid polymer does not need mastication prior to dissolving. The solubility of the elastomeric polyurethanes in ketone solvents is mainly governed by the degree of crystallinity in the polyester soft segments. The crystallinity can be varied by selecting the reactants and by controlling the molecular weight of the polyester. Generally, the elastomeric polyurethanes with low crystallization rates have long open times but poor peel strength and heat resistance. Those polyurethanes with high rates of crystallization have short open times but high peel strength and improved heat resistance. For this reason, mixtures of low and high crystallization rate polyurethanes are used to balance the open time and adhesion of the solvent-borne adhesives. The solvents determine the viscosity and solubility of the elastomeric polyurethane, its storage stability, its wetting properties, and its evaporation rate when

51

Shoe Industry

1505

applied on a substrate. The most common solvents are aromatic (toluene, xylene), tetrahydrofuran, dioxane, cyclohexanone, esters (ethyl acetate, butyl acetate), and ketones (acetone, methyl ethyl ketone). Commonly, two or more solvents are employed, a low-boiling and a high-boiling solvent. The low-boiling solvent assures rapid flash-off of the majority of the solvent after applying the adhesive solution on the substrate. The higher-boiling solvent helps to control the crystallization kinetics of the soft segments, thereby helping to extent the open time of the adhesive. In fact, once crystallization of the soft segments occurs, the open time expires. Other function of the high-boiling solvent is to keep the viscosity of the adhesive low (this is the reason for being called diluents), allowing the adhesive solution to wet the substrate by facilitating the polyurethane to penetrate the substrate surface, thus improving the mechanical interlocking. On the other hand, the addition of toluene avoids gel formation of the adhesive solutions during storage. The formulation of the solvent-borne polyurethane adhesives may include several additives such as tack and heat resistance modifiers, plasticizers, fillers, tackifiers, antihydrolysis agents, and cross-linkers. Carboxylic acid as an adhesion promoter can also be added. A typical formulation of a solvent-borne polyurethane adhesive contains less than 20 wt% solid content only. The spotting tack and/or the heat resistance of the elastomeric polyurethane adhesives may be extended by adding resins (alkyl phenolic, epoxide, terpene phenolic, coumarone) or polymers (low-crystallizing polyurethane, acrylic, nitrile rubber, chlorinated rubber, acetyl cellulose) (Penczek and Nachtkamp 1987) having low miscibility with the polyurethane. To improve adhesion together with heat resistance, reactive alkyl phenolic resins, chlorinated rubber, or other chlorinecontaining polymers can be added. The main additives used in the formulation of solvent-borne polyurethane adhesives are described in more detail below. Tackifiers. One of the limitations of the elastomeric polyurethane adhesives is the lack of high immediate (green) adhesion to rubbers and polymers. One way to increase the green strength of the elastomeric polyurethanes is the addition of tackifiers mainly rosins or hydrocarbon resins (Arán-Aís et al. 2002). Addition of rosin as internal tackifier (i.e., incorporated during the synthesis of the polyurethane) favors the miscibility between the tackifier and the polyurethane. The hard segment content of the elastomeric polyurethanes containing rosin is higher than in regular polyurethanes because two kinds of hard segments are produced, urethane and urea–amide (this is produced by reaction of the isocyanate groups with the carboxylic acid of the rosin). Addition of rosin increases the molecular weight of the polyurethane and retards its kinetic of crystallization, and the immediate adhesion to vulcanized SBR sole is also increased. A model system (phenyl isocyanate and acetic acid) has been studied to understand the reaction between the isocyanate group and the rosin (IrustaMaritxalar ML, Fernández-Berridi MJ (2002), personal communication). The expected by-products of the reaction of the isocyanate and the carboxylic acid moieties are given in Eq. 1:

1506

J. M. Martín-Martínez

R-NCO þ R0 -COOH!½R-NHCOOCO-R0 ðIÞ!R -NHCO-R0 ðIVÞ þ CO2 # "Heating or several days at room temperature 0

(1)

0

R-NHCONH-RðIIÞ þ R -COOCO-R ðIIIÞ

An unstable intermediate carbo-anhydride is formed (compound I), which under heating decomposes to produce urea (compound II) and anhydride (compound III) derivatives. These two compounds react by heating or by standing several days at room temperate to produce an amide compound (compound IV). Fillers. Fillers should be highly dispersible in solution and should not settle down during the storage period of the adhesive. Whiting, talc, barite, calcium carbonate, attapulgite, and quartz flour have been suggested as fillers to lower the cost of the solvent-borne polyurethane adhesive, improve joint filling, and reduce the loss of adhesive during setting. In general, the fillers are incorporated during the preparation of the adhesive solutions by adding a small amount of solvent under high stirring speed for a short time, followed by addition of the other ingredients of the adhesive and all solvent under lower stirring rate and longer time (Maciá-Agulló et al. 1992). Pyrogenic (fumed) silicas are the most common fillers of the polyurethane adhesives in shoe industry. When bonding highly porous materials (leather, textiles), fumed silicas are added to prevent undesirable penetration of the adhesive. The addition of the fumed silica adjusts the viscosity and rheology (thixotropy, pseudoplasticity) for application of the polyurethane adhesive solutions. The rheological properties of the polyurethanes containing fumed silica are due to the creation of hydrogen bonds between the silanol groups on the fumed silica particles and the urethane and carbonyl groups of the polymer, which increase the mechanical and rheological properties of the adhesives (Jaúregui-Beloqui et al. 1999; Vega-Baudrit et al. 2009; Bahattab et al. 2011, 2012; Donate-Robles et al. 2014). Consequently, the cohesion properties of the polyurethane are increased, and improved initial peel strength is obtained, in a greater extent by increasing the amount of fumed silica (Bahattab et al. 2012). Several parameters of the fumed silica determine its performance as rheological additive of solvent-borne polyurethane adhesives. Low particle size and degree of agglomeration, specific surface areas between 200 and 300 m2/g, and the hydrophilic nature of fumed silicas increase the rheological, mechanical, and adhesion properties of the solvent-borne polyurethane adhesives (Pérez-Limiñana et al. 2003; Bahattab et al. 2011). Calcium carbonate fillers have also been added to solvent-borne polyurethane adhesives for improving their mechanical properties and initial adhesion (DonateRobles and Martín-Martínez 2011a, b). The addition of precipitated calcium carbonate (PCC) filler produced a moderate increase in the rheological and viscoelastic properties of the polyurethane due to the poor dispersion of filler and the weak interactions between the PCC nanoparticles and the polymer chains. Furthermore, the first glass transition temperature of the polyurethane decreased by adding PCC filler and the crystallinity of the soft segments as well, because of the disruption of the degree of phase separation in the polyurethane. The initial adhesive strength in

51

Shoe Industry

1507

PVC/polyurethane adhesive/PVC joints increased noticeably by adding PCC filler; the greater the amount of filler, the greater the initial adhesive strength, and the highest final adhesive strength (72 h after joint formation) was obtained in the joint produced with the solvent-borne polyurethane adhesive containing 10 wt% PCC. The thermoplastic polyurethane–calcium carbonate interactions have been studied by using flow microcalorimetry and diffuse reflectance Fourier transform infrared spectroscopy (Donate-Robles et al. 2014). Three calcium carbonates (coated and uncoated precipitated calcium carbonate and natural ultramicronized uncoated calcium carbonate) were added to solvent-borne polyurethane adhesives, and stronger polyurethane-filler interaction was shown in the uncoated precipitated calcium carbonate due to the fact that more of the surface was available for interaction. In fact, the polyurethane strongly adsorbed onto untreated calcium carbonate via the ester groups, and the stearate treatment of calcium carbonate greatly reduced the strength of interaction with the polyurethane due to blockage of the surface adsorption sites. Nanometric carbon black (CB) filler has been added to solvent-borne polyurethane adhesives for improving adhesion. The CB and the polyurethane pellets were adequately dispersed in methyl ethyl ketone matrix, and an increase in the number and size of the carbon black aggregates in the polyurethane matrix was obtained by increasing the carbon black loading (Alvarez-García and Martín-Martínez 2015). The addition of the carbon black improved the rheological and viscoelastic properties the adhesive, and the addition of higher amounts of CB changed the viscoelastic behavior of the polyurethane which became mainly elastic. On the other hand, the addition of CB loadings up to 12 wt% increased the thermal stability of the polyurethanes and increased their elongation at break without noticeable reduction in tensile strength. However, the polyurethane adsorbs less carbon black than fumed silica and calcium carbonate due to less accessible surface groups due to the presence of relatively important amount of micropores (DonateRobles et al. 2014). Carboxylic acids. It has been shown that improved adhesion can be obtained in joints produced with solvent-borne polyurethane adhesives containing small amount of different carboxylic acids. These carboxylic acids can be aliphatic or aromatic, can be unsaturated or saturated, may contain one or more carboxylic acid functionalities, and have hydroxyl, carbonyl, or halogen groups. Fumaric and maleic acids are the most common carboxylic acids added to the solvent-borne polyurethane adhesives for the joining of vulcanized SBR sole. The effectiveness of the carboxylic acid as adhesion promoter is governed by its compatibility with the polyurethane and the solubility in the organic solvent selected to dissolve the polyurethane. It has been established (Pastor-Sempere et al. 1995) that the mechanism by which adhesion of SBR is increased is due to the carboxylic acid migration to the polyurethane adhesive surface once the adhesive joint is formed, removing zinc stearate and paraffin wax from the rubber surface. On the other hand, the carboxylic acid reacts with the polyurethane in the presence of a tiny amount of water which results in chain cleavage and disruption of polyurethane crystallinity (Eq. 2).

1508

J. M. Martín-Martínez

NH-C-O

O

=

=

O

C-O

+

HOOC-CH=CH-COOH

H2O

ð2Þ =

NH-C-O

O

O

=

=

O

C-OH

+

HOOC-CH=CH-C

The addition of carboxylic acid decreases the molecular weight, tensile strength, and glass transition temperature of the polyurethane more markedly upon increasing the time after the adhesive is prepared and more noticeably for high amounts of carboxylic acid. As a consequence, lower viscosity of the polyurethane adhesive solutions and smaller peel strength values are obtained. The increased amount of the carboxylic acid in the solvent-borne polyurethane adhesive enhances its bond strength to SBR until a maximum value is reached. On the other hand, the nature of the polyurethane determines the degree of effectiveness of the carboxylic acid on adhesion of vulcanized rubber. Cross-linkers. Polyisocyanates with functionality greater than two (such as p, p”, p”’ – triisocyanate triphenylmethane, thiophosphoric acid tris(p-isocyanatophenyl) ester) can be added to improve specific adhesion and heat resistance of the polyurethane adhesives (especially those with slower crystallization rate). Most of them have 20–30 wt% solids and 5.4–7 wt% free NCO, and the most common solvent is ethyl acetate. About 5–10 wt% polyisocyanate is generally added in the adhesive solution just before application on the substrate, and it acts as cross-linking agent. Since all isocyanates react with the residual hydroxyl groups on the polyurethane, in solution they yield adhesive systems with a limited pot life (1–2 h to several days). The addition of the polyisocyanate improves the spotting tack and the heat activation of the freshly dried adhesive films, but after several days they lose their capacity to be reactivated. In finished bonds, the elastomeric polyurethanes show a relatively low cross-link density and then are exposed to outstanding hydrolytic degradation under high humidity and temperature that can lead to a partial or complete loss of strength during shoe use. The hydrolytic degradation of the polyurethane can be inhibited by adding 2–4 wt% carbodiimide to the adhesive solution because the carbodiimide forms acyl ureas by reaction of the carboxyl groups arising from the hydrolytic degradation of the polyurethane (Dollhausen 1988).

Waterborne Polyurethane Adhesives Organic solvents are, in general, volatile, flammable, and toxic, in some degree. Further, organic solvent may react with other airborne contaminants contributing to smog formation and workplace exposure. Although solvent recovery systems and afterburners can be effectively attached to ventilation equipments, many shoe factories are switching to the use of waterborne adhesives. Waterborne polyurethane adhesives are an environmental-friendly alternative to the solvent-borne polyurethane adhesives. Those are the more logical choice to replace the solvent-borne adhesives because they can be processed on the same

51

Shoe Industry

1509

machines, their performance in many applications is just as good as that of solvent systems, and they can be used economically despite their higher raw material and processing costs. However, the use of the waterborne polyurethane adhesives needs some minor changes in the current existing technology, essentially the additional heat required to remove water before joint formation. Furthermore, the waterborne polyurethane adhesives show additional limitations: (i) lack of tackiness at room temperature (heat activation is needed), (ii) poor wettability of some substrates (particularly greased leather), and (iii) polyurethane dispersions are thermodynamically unstable, and therefore they show relatively poor storage stability (dispersion collapses in the presence of metallic contaminants, at low temperature (generally below 5  C), or by applying high stresses). The polyurethane dispersions are constituted by urethane polymer particles wholly dispersed in water. The dispersion of the polymeric particles is feasible because of the presence of ionic hydrophilic groups chemically bonded to the main polymer chains which are oriented to the surface of the particles (Fig. 9). These dispersions are based on crystalline, hydrophobic polyester polyols (such as hexamethylene polyadipate), and aliphatic isocyanates (such as H12MDI – methylene bis(cyclohexyl isocyanate) – or IPDI, isophorone diisocyanate). There are several routes to produce polyurethane dispersions, the two most common are the so-called acetone and prepolymer methods. In the acetone method, the polyol, the anionic internal emulsifiers (such as dimethylolpropionic acid or

Fig. 9 Scheme of the structure of waterborne polyurethane adhesive

–

–

–

–

–

–

–

–

–

Polymer –

–

– – –

Polymer –

–

–

– –

–

Hydrophilic anionic layer – –

–

– –

–

– –

Polymer –

–

–

– Polymer –

– –

– –

–

–

–

1510

J. M. Martín-Martínez

diethylamine sodium sulfonate), and the isocyanate are reacted to obtain a prepolymer with hydrophilic groups (García-Pacios et al. 2010). The prepolymer is dissolved in acetone for decreasing viscosity, and the ionic groups are neutralized by addition of an amine (trimethylamine is generally used to neutralize the carboxylic acid moieties). The chain extension of the prepolymer is carried out by addition of or diol; both hydrazine and ethylene diamine are commonly used as chain extenders because of their faster reaction with isocyanates than with water. Then, water is added for obtaining the polyurethane micelles by phase inversion, and the acetone is removed by distillation at reduced pressure. In the prepolymer method, two basic steps are needed in order to produce a waterborne polyurethane dispersion: a prepolymer step and a chain extension step. The prepolymer obtained by reacting the polyol and the aliphatic isocyanate is modified through an internal emulsifier containing hydrophilic groups, thereby eliminating the use of external emulsifiers. The prepolymer containing the internal emulsifier is reacted with a tertiary amine (such as triethylamine) to increase the length of the polymer chain and increase the molecular weight of the polyurethane. The amine reacts with the pendant carboxylic acid groups, forming a salt that under adequate stirring allows the dispersion of the prepolymer in water. The chain extension step takes place in the water phase. The waterborne polyurethane dispersions obtained by the acetone and the prepolymer methods contain high-molecular-weight polyurethane chains, and they have polyurethane and polyurea linkages. The method for preparing the waterborne polyurethane dispersions determines their properties, and it has been scarcely studied. In general, the acetone method produces more homogeneous anionic dispersions and well-controlled polyurethane structure. It has been shown that the method of synthesis determined the mean particle size, the crystallinity, and the viscosity of the waterborne polyurethane dispersions (Barni and Levi 2003). In later study it has been shown that the acetone method produced dispersions with narrower particle size distributions and higher crystallinity than the prepolymer method, and the adhesive strength was similar in the joints made with waterborne polyurethane dispersions prepared with both methods (Pérez-Limiñana et al. 2006). The waterborne polyurethane dispersions generally have a pH between 6 and 9 and higher solid content (35–50 wt%) and lower viscosities (about 100 mPa.s) than the solvent-borne polyurethane adhesives. They are high-molecular-weight linear polyester ureas or urethanes constituted by small rounded spherical particles of about 0.1–0.2 μm diameter. The adhesive characteristics of the waterborne polyurethanes are mainly defined by the melting point and the crystallization kinetics of the polymer backbone. It is highly desirable to activate the adhesive at room temperature, but most of the waterborne polyurethane adhesives have melting point above 55  C and need reactivation. The crystallization kinetics defines the open time of the adhesive. On the other hand, unlike solvent-borne adhesives, the viscosity of waterborne polyurethane adhesives is not dependent on the molar mass of the polymer but on the solids content, average particle size of the dispersion, and the existence of additives in the formulation.

51

Shoe Industry

1511

The adhesion properties of the waterborne polyurethane adhesives are greatly determined by the polymer, and the formulations are generally simple including only small amounts of a thickener and an emulsifier. When formulating the waterborne polyurethane adhesives, the application conditions (heat activation, contact bond time, pressure) and substrate type must be taken into account. Because the properties of the waterborne polyurethane adhesives are mainly determined by the structure of the polyurethane, several studies analyzing the influence of different raw materials and the experimental variables of the synthesis procedure have been published. It has been concluded that the content of ionic groups, the hard-to-soft segment ratio, the nature and molecular weight of the polyol, the nature of the chain extender, the degree of neutralization of the prepolymer, and the nature of the counterion determined the properties of the waterborne polyurethane dispersions (Kim and Lee 1996; Jang et al. 2002). Several studies have considered the influence of the hard-to-soft segments or NCO/OH ratio on the properties of the waterborne polyurethane dispersions. Most studies have shown that the increase in the NCO/OH ratio increased the mean particle size and enhanced the viscoelastic properties of the polyurethane dispersions (Pérez-Limiñana et al. 2007; García-Pacios et al. 2010; Vicent and Natarajan 2014). The variation of the viscosity of the waterborne polyurethane dispersion with the NCO/OH ratio is controversial. García-Pacios et al. (2011a) found that the increase in the NCO/OH ratio in waterborne polyurethane dispersions prepared with polycarbonate diol decreased the mean particle size and increased the Brookfield viscosity of the dispersions; however, the opposite trend was described previously (Madbouly et al. 2005). On the other hand, the increase in the NCO/OH ratio increases the urea and urethane hard-segment content in the polyurethane, and, therefore, the mechanical properties (i.e., Young’s modulus and tensile strength) and the glass transition temperature increased by increasing the NCO/OH ratio of the waterborne polyurethane dispersion. On the other hand, the NCO/OH ratio affected the T-peel strength and single-lap shear adhesive strength of the waterborne polyurethane adhesives, and a maximum value is obtained for an NCO/OH ratio of 1.5 (García-Pacios et al. 2011a). The influence of the nature, the structure, and the amount of the diisocyanate on the properties of waterborne polyurethane adhesives has been extensively studied. The cycloaliphatic diisocyanates (methylene bis(cyclohexyl isocyanate) – H12MDI – isophorone diisocyanate, IPDI) are the most commonly used because of the controlled reactivity during prepolymer formation. It has been shown that the waterborne polyurethane dispersions prepared with IPDI had lower mean particle size but lower mechanical properties than the ones prepared with H12MDI (Kim et al. 1994). Aromatic diisocyanates such as MDI and TDI have been used for preparing waterborne polyurethane dispersions, and they showed higher thermal and mechanical properties than the ones prepared with cycloaliphatic diisocyanates (Yang et al. 1999). Mixtures of aliphatic and aromatic diisocyanates have been used for preparing waterborne polyurethane dispersions, and the mechanical properties and adhesive strength were higher in the polyurethane made with MDI and lower with the one prepared with H12MDI (Rahman et al. 2014). However, after accelerated aging in

1512

J. M. Martín-Martínez

salt water, the mechanical strength and adhesive strength were better in the polyurethanes prepared with mixture of H12MDI and MDI diisocyanate. The nature (diol, diamine), chain length, and amount of the chain extender determined the properties of the waterborne polyurethane dispersions. The diol chain extender produced polyurethanes with high degree of phase separation, higher crystallinity, and good adhesive strength; this later is higher by increasing the chain length of the diol (Orgilés-Calpena et al. 2016a). On the other hand, the diamine chain extenders produced polyurea–urethanes that show lower crystallinity, lower degree of phase separation, and lower peel strength than the ones prepared with diol chain extenders (Rahman 2013). With respect to other raw materials, the polyol determines in greater extent the properties of the waterborne polyurethane adhesives. The increase in the molecular weight of the polyol (i.e., soft segments) produces dispersions with lower mean particle size and increased toughness and elongation at break (Mumtaz et al. 2013), but both the hardness and tensile strength decrease (Wang et al. 2015). On the other hand, the peel strength decreases by increasing the molecular weight of the polyol (García-Pacios et al. 2013a). The nature of the polyol (polyester, polyether, polycarbonate diol, polycaprolactone) determines the degree of phase separation in the polyurethane and affects noticeably the properties of the waterborne polyurethane dispersions. Because of the existence of carbonate groups, the use of polycarbonate diol produces lower degree of phase separation and increased shear adhesion with respect to the use of polyester and polyether (García-Pacios et al. 2013b). More recently the polyols derived from natural oils have been used for the synthesis of waterborne polyurethane dispersions (Yang et al. 2014; Hu et al. 2015). The waterborne polyurethane dispersions prepared with natural dimer fatty acid-based polyester polyols exhibited excellent water resistance, outstanding hydrolytic resistance and superior thermal stability than the ones prepared with nonnatural-derived polyester polyols (Liu et al. 2011). The properties of the waterborne dispersions are also affected by the nature and amount of the ionic groups. The increase in the content of dimethyl propionic acid (DMPA) decreases the mean particle size and increases the viscosity of the waterborne polyurethane dispersions (Huang and Chen 2007). Increased DMPA amounts result in the higher hard-segment contents and the increase in the mechanical properties of the polyurethanes (Cakić et al. 2013). Furthermore, the increase of the DMPA content decreased the adhesion of the polyurethane dispersions (Rahman and Kim 2006). The influence of the solids content on the properties of the polyurethane dispersions has been studied, and contradictory findings with respect to their influence on the mean particle size have been found. It has been stated that the increase of the solids content caused an increase in the mean particle size that was related to the decrease in the viscosity of the polyurethane dispersions (Lee and Kim 2009). Another study concluded that the mean particle size of the polyurethane dispersion decreased by increasing its solids content, and an increase in the viscosity was found, and this was related to the decreased particle–particle interactions and the higher degree of phase separation (García-Pacios et al. 2011b). On the other hand, the increase in the solids content does not affect the peel strength and decreases the single-lap shear strength of

51

Shoe Industry

1513

the waterborne polyurethane dispersions (García-Pacios et al. 2011b). A recent study analyzed the structure and properties of high solids content waterborne polyurethane dispersions synthesized by combining ionic and nonionic monomers (Lijie et al. 2015). The adequate combination of the molecular weight of the polyether and the amount of DMPA improved the stability of the dispersion and increased the solids content of the polyurethane dispersion up to 52 wt%. Additionally, the glass transition temperature of the soft segments and the crystallinity of the polyurethane decreased, although the mechanical properties were improved. Thickeners (polyvinyl alcohol, polyurethanes, polyacrylates, cellulose derivatives) are added to increase the low viscosity of the waterborne polyurethane adhesives and to avoid excessive penetration in porous substrates (Orgilés-Calpena et al. 2009a). The mechanism of thickening of the waterborne polyurethane adhesives by adding urethanebased thickener has been analyzed by confocal laser microscopy (Fig. 10). The addition of the thickener produces an entanglement in the polyurethane adhesive (seen as denser and whiter images in Fig. 10) which can be ascribed to the existence of interactions between the urethane-based thickener and the polyurethane. Several interactions take part in the thickening mechanism of the polyurethane adhesive dispersions by adding urethane-based thickener: (i) interactions between the terminal hydrophobic groups of the thickener with the polyurethane micelles, (ii) interactions by ionic adsorption of the polyester chains of the thickener with the ionic groups on the surface of the polyurethane particles, and (iii) interactions by hydrogen bonds between the urethane groups of the polyurethane dispersion and the urethane-based thickener (Orgilés-Calpena et al. 2009a, b). Figure 11 shows a schematic diagram of the thickening mechanism of waterborne polyurethane adhesive containing urethane-based thickener in which the formation of 3D micelle network is seen. On the other hand, the hard-to-soft segment ratios and the

14D6/5e

30 µm

14D6

30 µm

Fig. 10 Micrographs of confocal laser microscopy corresponding to the waterborne polyurethane adhesives without (14D6) and with (14D6/5e) 5 wt% urethane-based thickener

1514

J. M. Martín-Martínez

HEUR thickener

Hydrophobic groups

Polyurethane dispersion

Protonated POE

Urethane groups

Ionic group

3. Hydrogen bonds 1. Micelles

2. Ionic adsorption

Fig. 11 Thickening mechanism of waterborne polyurethane adhesive containing urethane-based thickener through the formation of 3D micelle network (Orgilés-Calpena et al. 2009b)

content of ionic groups determine the properties of the waterborne polyurethane adhesives. Thus, the addition of 5 wt% urethane-based thickener markedly increases the viscosity of the waterborne polyurethane adhesive more noticeably by decreasing the hard-to-soft segments ratio and by increasing the ionic group contents (Orgilés-Calpena et al. 2009a). On the other hand, as the hard-segment content of the thickened polyurethane adhesive decreases, the kinetics of crystallization is favored as a result of stronger polyurethane–thickener interactions. Emulsifiers (surface active agents) assist to stabilize the pH of the waterborne polyurethane dispersion and to decrease its surface tension to obtain improved wettability. The emulsifiers are always oriented on the interface between the polymer particles and the aqueous phase. Small amounts must be added to avoid loss in bonding characteristics. On the other hand, to achieve particular performance, dispersions of several polymers (vinyl acetate, acrylic ester) and resins (hydrocarbon resin, rosin esters) can be added to waterborne polyurethane adhesives. Special attention must be paid to the compatibility with the polyurethane to avoid dispersion collapse. In general, fillers are not added to waterborne polyurethane adhesives although dispersions of silicas can be added for adjusting the viscosity and the rheology. To produce adequate bond, the waterborne adhesive solution is applied to the two substrates to be joined, and the water is evaporated at room temperature for about 30 min or by heating with hot air or infrared lamps at 50–60  C for a few minutes. When the water comes out, the polyurethane viscosity rises, and a continuous film is formed

51

Shoe Industry

1515

on the substrates. Then, the dry films are heat activated, and the melting of the crystalline soft segment is produced imparting tackiness to the films. The heat activation process of the waterborne polyurethanes is affected very strongly by the substrates to be bonded and by the process parameters such as temperature, pressure, and contact bond time. Increasing the adhesive film temperature, raising the pressure, and extending the time of contact have all similar effect, i.e., an increase of the actual contact area is produced, and better bond is obtained. After heat activation, the decrystallized polyurethane film is an amorphous viscoelastic melt with good flow properties. The decrease in viscosity after heat reactivation allows the adhesive film to wet the substrate, and the joining to the adhesive film on a second substrate under pressure is produced. Once the adhesive joint is formed, the bulk viscosity and the modulus of the adhesive increase, initially by cooling and afterward by the recrystallization of the polyurethane. If an isocyanate is added as cross-linker of the waterborne polyurethane adhesive, the bond formation occurs much slower than with only the polyurethane, but a further increase in viscosity with time is obtained. The cross-linking reaction does not initially bring any increase in the peel adhesive strength at room temperature but increases the heat resistance of the bond. The most common cross-linkers are polyisocyanates (adequately modified with surface active agents to become emulsifiable in water), although azyridines, polycarbodiimides, and epoxies can also be used. The cross-linker is normally added to the dispersion and stirred in before application on the substrates. Once the cross-linker is added, the drying off of the water is important because the reaction with water reduces its effectiveness. Depending on the formulation, the adhesives can be applied over a period of 4–10 h without affecting the adhesion properties. The isocyanate cross-linker generally takes several days before the cross-linking reaction is completed. Unlike solvent-borne adhesives, the pot life of the adhesive dispersions is not associated with an increase in viscosity. In general, the durability of the polyurethane adhesives is acceptable although they are water sensitive, particularly the polyether-based polyurethane. Protection against water degradation can be reduced by adding carbodiimides and/or polyisocyanates during curing. The bond strength of the waterborne polyurethane adhesives is particularly dependent on the substrate. Good results are obtained with surface-chlorinated vulcanized rubbers and PVC, but adhesion to TR may be poor or variable (Abbott 1992). Adhesion to leather is sometimes insufficient, and roughening must be always carried out; adhesion to greasy or waterproof leathers and unroughened polyurethane coatings may also be difficult.

Solvent-free Polyurethane Adhesives Currently, the waterborne adhesives are being introduced into the shoe industry. Their performance is quite similar to that of the solvent-borne adhesives, so it can be estimated that for several years, they will be used in shoe industry. However, the future seems to be directed through the use of moisture-curing holt-melt urethane and thermoplastic urethane adhesives as they are 100% solids systems and evaporation of solvents is not necessary. Although the hot-melt urethanes could replace

1516

J. M. Martín-Martínez

waterborne adhesives, this could take longer to occur because of the vastly different equipment requirements and the change in the bonding concept by the shoe manufacturers. Some literature has been published dealing with moisture-reactive hot-melt polyurethane adhesives (Frisch 2002). Most moisture-curing hot-melt adhesives are obtained by reacting a crystallizable polyol such as poly(hexamethylene adipate) and monomeric MDI (NCO/OH ratios = 1.5–2.2). A catalyst such as dimorpholinediethyl ether is also necessary. The polyester polyol plays a key role because the open time, the viscosity, and the glass transition temperature can be adequately tailored. A mixture of hydroxyl-terminated polyesters having different characteristics allows control of the adhesion and hardening time of the adhesive. Recently, reactive polyurethane hot-melt adhesives containing polycarbonate polyols derived from CO2 and 4,40 -diphenylmethane diisocyanate (MDI) have been developed from improving the T-peel strength of leather/polyurethane adhesive/vulcanized SBR joints, and they show good green and final strength before and after high temperature/humidity aging and hydrolysis tests (Orgilés-Calpena et al. 2016a, b). The moisture-curing hot-melt adhesives must be stored in nitrogen-sealed containers, and moisture must be fully excluded. These adhesives are crystalline solids at room temperature, and for application, heating is necessary (about 125  C) to become low-viscosity liquids, and they are only applied on one of the substrates to be joined; the second substrate is immediately joined with minimal pressure. Simultaneously to the application of the adhesive, water can be sprayed onto the adhesives/ materials in the amount necessary for facilitating the curing process. Although after cooling down the crystallinity is reached, about 24 h is necessary to develop the strength of a structural adhesive. Initial strength can be improved by adding polyester, thermoplastics, or polymerized acrylates. Addition of ketone–formaldehyde and terpene phenolic resins also increases adhesive performance. Limitations of these adhesives derive from the special equipments necessary for application and from the substrate nature (not all substrates can be bonded). However, some pre-reacted hot-melt adhesive types are liquids and may be applied by hand, by spraying, or even better by nozzles. When the polyurethane hot-melt adhesive comes into contact with moisture, the irreversible curing of the adhesive is initiated. First, the water adds to the isocyanate group producing unstable carbamic acid, which in turn is transformed into a primary amine group. These amine groups generate linear urea bridges by reaction with additional isocyanate; a three-dimensional network is produced by cross-linking with more isocyanate groups, providing the structural bonding in moisture-cured hot-melt polyurethane adhesives. The thermoplastic polyurethane adhesives are another future alternative for bonding in shoe industry. These adhesives have good holding strength after crystallization, but their cost is higher than the majority of the adhesives used in shoe industry. Most thermoplastic polyurethane adhesives are based in fast crystallization polyesters. They are prepared using a NCO/OH ratio near 2, and they have linear structure. They consist in two components (polyol and isocyanate) which have to be mixed at about 60  C in the presence of a catalyst such as dibutyltin dilaurate; after reaction, the

51

Shoe Industry

1517

resulting polymer is pelletized and packed in nitrogen- and moisture-free atmosphere containers. Application temperatures are higher than 170  C, and to produce bonding the same procedure as for moisture-cured hot-melt urethane is used.

51.3.5 Polychloroprene Adhesives Although polychloroprene adhesives (more often called neoprene adhesives because of the trade name given by Du Pont to the first commercial product) were extensively used in the past in the sole-to-upper bonding, nowadays they have been practically substituted by polyurethane adhesives. Therefore, there is no sense in providing an extensive description of the polychloroprene adhesives in this chapter which the prime aim is to provide an actual overview. A detailed description of the polychloroprene adhesives for shoe bonding can be found in Chap. 14 of the book entitled “Adhesive Bonding. Science, Technology and Applications” (Adams 2005). Therefore, in this chapter, the main features of these adhesives will be only considered, and the reader can refer to that chapter for more detailed information. Solvent-borne polychloroprene adhesives are less used in shoe industry because of the presence of aromatic organic solvents in their formulation which are currently unacceptable. However, the most recent developed waterborne polychloroprene adhesives can compete with waterborne polyurethane adhesives because of their good performance in the upper-to-sole bonding. Furthermore, they are extremely useful as nonpermanent adhesives in several mounting operations (previously to the sole-to-upper attachment) involving porous substrates in the shoe manufacturing.

Solvent-Borne Polychloroprene Adhesives Polychloroprene adhesives are contact adhesives (see ▶ Chap. 14, “Adhesive Families”). The diffusion process in polychloroprene rubber adhesives is mainly affected by the solvent mixture of the adhesive (which determines the degree of uncoiling of rubber chains) and by the ingredients in the formulation (mainly the amount and nature of tackifier). The chemical nature and molecular weight of the polychloroprene greatly determine its adhesive properties. The polychloroprene adhesives show high peel strength; high green strength and noticeable resistance to moisture, chemicals, and oils; excellent aging properties; and excellent temperature resistance. These characteristics are important in shoe industry. Polychloroprene elastomers are produced by free radical emulsion polymerization of 2-chloro-1,3-butadiene monomer. The emulsion polymerization of chloroprene involves the dispersing of monomer droplets in an aqueous phase by means of suitable surface active agents, generally at a pH of 10–12. Polymerization is initiated by addition of free radical catalyst at 20–50  C. To obtain a high conversion in the polymerization reaction and processable polymer, the addition of sulfur, thiuram disulfide, or mercaptans is necessary (Whitehouse 1986). The crystallinity in the polychloroprene is essential in sole-to-upper bonding in shoe industry. Crystallinity is produced by noticeable trans-1,4 addition (more than 90%) during polymerization. As a result of crystallization, the cohesive strength of

1518

J. M. Martín-Martínez

the polychloroprene is much greater than that of the amorphous polymer. Crystallization is reversible under temperature or dynamic stresses. Thus, for a temperature higher than 50  C, uncured polychloroprene adhesives lose their crystallinity, and upon cooling the film recrystallizes, and cohesive strength is regained. The increase in the crystallinity improves the modulus, hardness, and cohesive strength of the polychloroprene adhesives but decreases their flexibility. A typical composition of solvent-borne polychloroprene adhesive for upper-tosole attachment is given in Table 2. The role of the different components in solvent-borne polychloroprene adhesive is described below in more detail. The sulfur-modified polychloroprenes with little or no branching are the most commonly used in shoe adhesives as they are soluble in aromatic solvents, and for difficult-to-bond substrates, methacrylic graft polymers show better performance. The polymer type influences several properties of the solvent-borne polychloroprene adhesives, mainly the molecular weight and the rate of crystallization. The increase in the molecular weight of the polychloroprene imparts higher solution viscosity, adhesive strength, and heat resistance. On the other hand, the increase in the crystallization rate of the polychloroprene improves the rate of development of the bond strength and the ultimate strength at high temperature, but a reduction in open time (tack) is obtained. However, the tack can be varied by changing the pressure during joint formation and by an adequate selection of the solvent; the lower the volatility of the solvent, the longer is the open time of the solvent-borne polychloroprene adhesive. Metal oxides provide several functions in solvent-borne polychloroprene adhesives. The main function of the metal oxides in the polychloroprene adhesive formulations is acting as acid acceptor. Upon aging, small amounts of hydrochloric acid are released which may cause discoloration and substrate degradation. Magnesium oxide (4 phr) and zinc oxide (5 phr) act synergically in the stabilization of the solvent-borne polychloroprene adhesives against dehydrochlorination. On the other hand, the magnesium oxide retards scorch during mill processing of the polychloroprene adhesives and also reacts in solution with the t-butyl phenolic resin to produce an infusible resinate (a small amount of water is also necessary) which provides improved heat resistance. Furthermore, the zinc oxide produces a room temperature cure of the solvent-borne polychloroprene adhesives, giving increased strength and improved aging resistance.

Table 2 Typical composition of a solvent-borne polychloroprene adhesive. phr means parts per hundred parts of rubber

Polychloroprene Tackifying resin Magnesium oxide Zinc oxide Water Antioxidant Solvent mixture

100 phr 30 phr 4 phr 5 phr 1 phr 2 phr 500 phr

51

Shoe Industry

1519

Addition of resins to solvent-borne polychloroprene adhesives serves to improve its specific adhesion, increases tack retention, and increases hot cohesive strength. Para-tertiary butyl phenolic resins are the most common, and amounts between 35 and 50 phr are generally added. In general, the tack decreases by increasing the phenolic resin content in the polychloroprene adhesive, and bond strength reaches a maximum at about 40 phr, decreasing for higher amounts of phenolic resin. Solutions of polychloroprene adhesives containing metal oxides and t-butyl phenolic resin may show phasing (e.g., clear upper layer and flocculated lower layer of metal oxides) on standing upon days or months. To recover the full utility of the metal oxides, agitation before use is sufficient. Phasing is ascribed to the wide molecular weight distribution of the t-butyl phenolic resins. Terpene phenolic resins can also be added to solvent-borne polychloroprene adhesives to increase open tack time and to provide a softer glue line than the t-butyl phenolic resins. To provide adequate hot-bond strength, these resins are used in combination with a polyisocyanate-curing agent. An antioxidant (about 2 phr) should be added to the polychloroprene adhesives to avoid oxidative degradation and acid tendering of the substrates. Derivatives of diphenylamine (octylated diphenylamine, styrenated diphenylamine) provide good performance, but staining is produced. To avoid staining, hindered phenols or bisphenols can be added. Solvent affects the adhesive viscosity, the bond strength development, the open time, and the ultimate strength of the polychloroprene adhesives. Blends of three solvents (aromatic, aliphatic, oxygenate – e.g., ketones, esters) are generally added. Typical solvent blends for polychloroprene adhesives are given in Table 3. The open tack time of the solvent-borne polychloroprene adhesives partially depends on the evaporation rate of the solvent blend. If a solvent evaporates slowly, the solventborne polychloroprene adhesive will retain tack longer, whereas if the solvent evaporates quickly, the cohesive strength will develop more rapidly.

Table 3 Influence of the solvent on the viscosity and the tack of polychloroprene adhesives. Formulation: 100 phr polychloroprene elastomer; 4 phr MgO; 5 phr ZnO; 2 phr hindered phenolic antioxidant; 500 phr solvent mixture (Whitehouse 1986) Solvent Toluene Toluene/n-hexane/MEK n-Hexane/MEK Cyclohexane/acetone MEK/acetone Toluene/n-hexane/ethyl acetate Toluene/MEK/acetone n-Hexane/acetone Dichloromethane 1,1,1-Trichloroethane

Mixing ratio – 35/5/15 55/45 80/20 75/25 34/33/33 34/33/33 50/50 – –

Solution viscosity 20  C (mPa.s) 4460 3500 1520 3700 2100 3600 3400 1140 4670 4600

Tack time (min) 35 25 15 18 15 30 22 7 30 38

1520

J. M. Martín-Martínez

Isocyanates can be added to solvent-borne polychloroprene adhesive solutions as two-part adhesive systems. The reaction of the isocyanate with polychloroprene that leads to improved heat resistance property has not been fully explained as there are no active hydrogen atoms in the polychloroprene to allow reaction with isocyanate group. The two-part adhesive system is less effective with rubber substrates containing high styrene content and with butadiene–styrene block (thermoplastic rubber) copolymers. To improve the specific adhesion to those materials, addition of poly-alpha-methylstyrene resin to solvent-borne polychloroprene adhesives is quite effective (Tanno and Shibuya 1967).

Waterborne Polychloroprene Adhesives In recent years, the use of solvent-borne polychloroprene adhesives has been seriously restricted, and waterborne adhesives have been developed. Polychloroprene latex differs from its solid elastomer counterpart (used in solvent-borne polychloroprene adhesive) in that it is a gel polymer (e.g., insoluble in organic solvents). Latex systems derive their bond strength characteristics from the gel structure rather than crystallinity as in solvent solution systems. Higher gel content leads to the same properties than polymers with higher crystallinity (Lyons and Christell 1997). Polymers with higher gel content exhibit higher cohesive strength, modulus, and heat resistance, but tack, open time, and elongation at break are reduced. The initial formulations of the polychloroprene latex adhesives contain essentially the same components than the solvent-borne adhesives, except that waterbased ingredients have to be used, surfactant, thickener and antifoaming agent must be added, and the compounding is particularly exigent. A typical composition of initial waterborne polychloroprene adhesive is given in Table 4. In the last years, more simple and efficient formulations of waterborne polychloroprene adhesives for shoe industry have been developed (Kueker et al. 2016). The typical formulation of the current waterborne polychloroprene adhesives is given in Table 5 and consists in polychloroprene, acrylic or silica sol dispersion, zinc oxide, surfactant, antifoaming agent, and antioxidant. The preparation of these waterborne polychloroprene adhesives starts with the addition of the antioxidant to the polychloroprene dispersion, followed by slow addition of zinc oxide and the final incorporation of the acrylic or silica sol dispersion. The final adhesive strengths obtained with the current waterborne polychloroprene adhesives are due to the cross-linking between the polychloroprene

Table 4 Typical composition of the initial waterborne polychloroprene adhesive formulation

Polychloroprene dispersion Surfactant Antifoam Tackifying resin Thickener Zinc oxide Antioxidant

100 phr As required As required 50 phr As required 5 phr 2 phr

51

Shoe Industry

Table 5 Typical composition of the current waterborne polychloroprene adhesive formulation

1521 Polychloroprene dispersion Surfactant Antifoam Silica sol dispersion Zinc oxide Antioxidant

100 phr As required As required 10 phr 1 phr 2 phr

and the silica or acrylic micelles, and similar adhesive strengths to the ones obtained with solvent-borne adhesives are obtained, although the open time is generally longer. The use of silica sol dispersions with low mean particle size (i.e., about 5 nm) improves more the initial adhesive strength or wet bonding to leather than the use of silica sol dispersions with higher mean particle size, and the increase in the amount of silica causes softening and a decrease in the mechanical properties of the polychloroprene. Finally, the heat resistance of the polychloroprene adhesives can be noticeably increased by using hydroxylated polychloroprene dispersion due to the hydrogen bond formation with the silica sol particles; in fact when this hydroxylated polychloroprene dispersion is used, there is no need of adding polyisocyanate cross-linker for increasing the aging resistance, and this is even better than by using solvent-borne polychloroprene adhesives (Kueker et al. 2016). Recently, the magnetic conditioning for 3 h of aqueous-based polychloroprene contact adhesive has been proposed for increasing its adhesion (Souza et al. 2016). To promote adhesion, a nanometric zinc oxide, carbon dioxide as catalyst, hydroxylamine, and a magnetic conditioning process before the application of the adhesive were carried out. The magnetic conditioning was performed in magnetic cell comprising of containers for circulation of the adhesive, two submerged centrifugal hoses for transfer of the adhesive, a set of magnets iron–boron–neodymium with magnetic field of 2120 gauss, gaussmeter, and agitators. Noticeable increase in shear adhesion was obtained. The polychloroprene latex determines the initial tack and open time, the bond strength and the hot-bond strength, the application properties, and the adhesives viscosity. Because most of the latexes have low viscosities, most of the waterborne polychloroprene rubber adhesives are sprayable. Thickeners such as fumed silicas can be added to some formulations for increasing viscosity and thixotropy. Grafting of methyl methacrylate (MMA) onto polychloroprene rubber latex (CRL) has been carried out by emulsion polymerization using a redox initiator, and the surface diffusion-controlled process model was proposed to describe the grafting mechanism (Zhang et al. 2012). The graft copolymer was blended with tackifier and filler for waterborne contact adhesive applications, and the 180 peel strength to canvas to other decoration materials showed similar performance than solvent-based contact adhesive and superior to the one obtained with polychloroprene and polymethyl methacrylate. Based on the synergistic effect of polychloroprene latex and styrene–acrylate emulsion, waterborne contact adhesive containing 40 wt% styrene–acrylate emulsion

1522

J. M. Martín-Martínez

and 1.25 wt% boric acid has been developed (Zhang et al. 2009). The blend had a good shelf stability, its set time was 5 min, its tensile strength was reasonable, and it had very high elongation at break. The performance of the waterborne contact adhesives was found to be comparable to the solvent-based contact adhesives because they offered good wetting ability to substrate, high initial tack was obtained, it possessed excellent crystallization ability, and it enhanced the cohesive strength of the adhesive. Although noncompounded polychloroprene latex has good mechanical and storage stability, surfactants are added commonly for stabilization. They function by strengthening the interfacial film by maintaining or increasing the degree of solvation or by increasing the charge density on the latex particle. More precisely, surfactants are added to improve storage stability, substrate wetting, and attain improved freeze resistance. However, incorporation of surfactants has an adverse effect on cohesive properties and should be kept to a minimum. Water resistance and tack may also be affected. Excessive stabilization of the adhesive mixture may negatively affect coagulation (which is desirable in the wet bonding process). Anionic emulsifiers (alkali salts of long-chain fatty acids and alkyl/aryl sulfonic acids) or nonionic emulsifiers (condensation products of long-chain alcohols, phenols, or fatty acids with ethylene oxide) can be used. Zinc oxide is the most effective metal oxide and plays three main functions: (i) promote cure; (ii) improve aging, heat, and weathering resistance; and (iii) acid acceptor. In general, 2–5 phr zinc oxide is added in latex formulations. Resins influence the adhesion, open time, tack, and heat resistance of the waterborne polychloroprene adhesives. In some formulations, 30–60 phr is added, and attention should be paid to the pH and compatibility with the surfactant. The glass transition temperature, the softening point, the polarity, and the compatibility of the resin with the polymer determine the adhesive properties. Thus, the hot-bond performance is generally proportional to the softening point of the resin. t-Butyl phenolic resins cannot be used in waterborne polychloroprene adhesives because of colloidal incompatibility. Instead, terpene phenolic resins can be added to polychloroprene latex without great reduction in hot strength as the resin content is increased; however, the contact ability is reduced, and an adhesion failure is obtained, even at the 50 phr level. Furthermore, the terpene phenolic resins have relatively poor tack but impart good resistance to elevated temperatures to the polychloroprene latexes and require either heat activation or pressure to achieve adequate bond strength. Rosin ester resin emulsions are also effective in latex adhesives as they extend the tack life of the polychloroprene latexes, but they do not have the reinforcing characteristics of the terpene phenolic or alkyl phenolic resins. Hence, the cohesive strength and heat resistance are sacrificed to obtain surface tack. New nitrogen-containing compounds have been proposed as highly efficient adhesion promoters for adhesive compositions based on polychloroprene (Keibal et al. 2011). On the other hand, similar antioxidants than for solvent-borne polychloroprene adhesives can be used in the formulation of waterborne polychloroprene adhesives. Addition of thickeners increases the viscosity of the polychloroprene latex adhesives. Amounts up to 1 wt% of polyacrylates, methyl cellulose, alginates, and

51

Shoe Industry

1523

polyurethane thickeners can be used. Particular attention should be paid to fluctuations in pH when thickener is added in the formulations. For low pH (7–10) formulations, fumed silica or some silicates can be used. Curing agents have little effect on the performance of latexes with the highest gel content, but they are sometimes used with low-gel polymers to improve hot-bond strength while maintaining good contactability. Suitable curing agents are thiocarbanilide either alone or in combination with diphenylguanidine, zinc dibutyldithiocarbamate, and hexamethylenetetramine. Two-part systems have been developed using more active materials such as aqueous suspension isocyanates and hexamethoxy melamine. These agents produce a cross-linking reaction at room temperature and give fast bond development but exhibit a finite pot life. The mostcommon use of the curing agents is with carboxylic latices. Isocyanates and melamines can be used, but zinc oxide is the most common curing agent. Zinc oxide cross-links carboxylated latices and improves bond strength by ionomer formation (Cuervo and Maldonado 1984). Carboxylated polychloroprene reacts slowly with zinc oxide in dispersed form, causing a gradual increase in adhesive gel content that can lead to restricted adhesive shelf life. Resin acid sites compete with the polymer acid sites for ZnII; the more resin acid sites, the more stable the adhesive is.

51.4

Testing of Adhesive Joints in Shoe Bonding

In order to produce an optimum adhesive bonding of upper to sole, apart from the adequate surface preparation of the sole and upper materials and the adequate choice of the adhesive, the procedure to produce the joint should be carefully controlled and optimized. To achieve adequate adhesion of upper to sole, the following issues should be considered. 1. Selection of the upper and sole materials. In general, formulators of upper and sole materials do not consider that they have to be bonded, but they pay more attention to match the hardness and mechanical properties and aesthetic of the materials. Furthermore, depending on fashion, different difficult-to-bond materials are used to produce shoes which make problematic to standardize bonding. In general, the formulation of the shoe materials must elude the presence of additives able to decrease the adhesion to polyurethane or polychloroprene adhesives (plasticizers, excessive amounts of processing oils, inadequate selection of antiozonants, and antioxidants), and especially the use of too greased leather upper must be avoided. On the other hand, the formulations must be repetitive, and first-class raw materials must be used. As a general rule, the use of adequately formulated uppers and soles avoid more than 90% of the bonding problems in shoe industry. 2. Selection of the adhesive. The adhesive must be selected considering the performance required for each joint. Adhesive must properly wet the upper and sole surfaces, and in this aspect, solvent-borne adhesives are excellent. One of the problems in the use of waterborne adhesives is their poor wettability, although an

1524

J. M. Martín-Martínez

adequate formulation may solve quite satisfactorily this limitation. In the selection of an adhesive, the following aspects should be particularly considered: nature and formulation of the materials to be bonded; stresses produced during shoe use; environment of use (solvents, acid, or alkali media), only for safety shoes; adhesive application restrictions (viscosity, rheological properties); specific requirements of the adhesive (pot life, working scheme in the shoe factory): and safety regulations. 3. Design of the joint. Fashion dictates the shape and geometry of the shoes. Sometimes, the shoes become difficult to bond because of high heels or heterogeneous shapes of soles. In general, the shoe designers do not pay attention to the bonding. 4. Adequate bonding operation. Several cases of poor adhesion in shoe bonding arise from a deficient operation. As a summary, the following aspects must be properly obeyed to assure an adequate upper-to-sole bonding: – Surface treatments of upper and sole. The way to produce the surface preparation and the instruments used are critical. Furthermore, proper surface preparation for each upper and sole must be selected. – Adhesive application. A thin film of adhesive (about 100 μm thick) is applied to each of the two substrates to be bonded, and the solvent is removed by natural or forced evaporation. A heavier coat of adhesive is more likely to result in a cohesive failure in the substrate. The application procedure (brush, doctor knife, spray gum, roller, coater) and the amount of adhesive must be carefully controlled. The choice of the adhesive application devices depends mainly on the type and size of the materials to be bonded as well as on the rheological properties of the adhesive. Furthermore, the viscosity of the adhesive must be controlled, and the operation times (evaporation rate of organic solvents or water, open time, shelf life) must be strictly obeyed. – Adhesive film drying. Controlled drying of the shoe bottom cements is preferable to natural drying. Removed excess water or organic solvent from an adhesive film is not entirely governed by the process of evaporation but also by the speed of absorption into the substrate. Porous substrates (such as leather) absorb water or organic solvents without detrimental effect on bond strength, whereas nonporous substrates need the complete solvent removal to produce adequate adhesion. Furthermore, a force drying may produce skin formation on the surface of the coating (especially if heavier adhesive coating is applied) leading to poor coalescence of the adhesive. A slight trace of organic solvent in the cement film on the upper at sole attaching is beneficial in giving complete coalescence. When necessary, although very fast drying can be achieved by radiant heat, the use of IR radiation and hot air is a more convenient method, and both provide more uniform heating. – Bond formation. The components with the dry adhesive film are placed for 10–30 s in a “flash heater” – hot air or IR radiation can also be used – where a radiant heat source raises suddenly the temperature of the adhesive film above the crystalline melting of the polyurethane (heat activation or reactivation process). Recommended reactivation temperature for polyurethane adhesives

51

Shoe Industry

1525

ranges between 45 and 85  C, depending on the substrates and the adhesive characteristics. An insufficient reactivation temperature causes non-coalescence of the adhesive, and an excessive reactivation temperature decreases creep because the substrate is too hot and the adhesive is still soft when sole and upper are attached (SATRA 1963). In some particular cases, the reactivation temperatures of the adhesives higher than 100  C are recommended to produce good performance. While still in their amorphous state, the adhesive films are brought together under pressure (typically 35–200 kPa) for 10–30 s (stuck-on process). The time interval after the heat activation process during which the films adhere is known as the “spotting tack.” Both the applied pressure and the pressing time of the upper to the sole must be adequately selected for each joint. – Crystallization or curing of the adhesive. After bond formation, the adhesive joint needs time to gain sufficient cohesion. Time can vary, although in general a minimum of 24 h is required for crystallization, being 72 h the optimum. For waterborne adhesives, the curing time may be longer. T-peel (see ▶ Chap. 20, “Fracture Tests”) and creep (see ▶ Chap. 34, “Creep Load Conditions”) tests are the most commonly used to establish the adhesion performance in shoe bonding. The peel test serves to determine the bonding properties of upper to sole in shoe industry. Peeling rate, material thickness, and size of test samples must be optimized. Standard T-peel tests require two rectangular test pieces of 150 mm long, 30 mm width, and 3 mm thick that are stuck together to cover it other to a length of at least 50 mm (Fig. 12). Standard peel rate is 100 mm/ min. The joints are stored for 72 h in standard atmosphere (23  C and 50% relative humidity) before carrying out the separation tests. Both the peel resistance and the way in which the separation occurs help to assess the bond. A minimum of five replicates, preferably ten, for each joint must be tested and averaged. Bond strength should be expressed as kN/m, although very often N/mm is used in the shoe industry. The loci of failure of the joints are generally expressed using the following capital letters: – A: Adhesion failure (detachment of the adhesive film from one of the materials) – C: Cohesive failure in the adhesive (separation within the adhesive film without detachment from the material) – N: Non-coalescence failure (failure of the two adhesive films without detachment from the material) – S: Surface cohesive failure of the material (breakdown of a substrate of low structural strength at its surface) – M: Cohesive failure of one of the substrates

Under the influence of force and when heated, the adhesive layers of footwear material bonds suffer plastic flow. The creep test at constant temperature serves to assess the behavior of shoe material bonds when heated under the influence of a

1526

J. M. Martín-Martínez

Fig. 12 T-peel test sample

constant peeling force over a fixed time. Weight between 0.5 and 2.5 kg can be fixed on the lower holders outside a heating oven where the test pieces can be heated at 50–70  C. The unbounded ends of five test pieces are bent apart carefully and inserted in the holders. Temperature is generally raised to 60  C after a warm-up period of 1 h. The test pieces are loaded with the chosen weight constantly for 10 min. Then, the heating oven is opened, and the separations of the bonds are marked. Moisture and temperature are the main agents able to degrade the shoe joints. Therefore, adequate aging tests have been developed to assure adequate performance. In general, addition of polyisocyanate hardener increases the durability and retards aging in most of the upper-to-sole adhesive joints. Several tests have been proposed to establish the aging resistance of upper-to-sole bonds. The most common aging tests involve the immersion in hot water, the exposition of the joints to 50  C and 95% relative humidity for 1 week, freeze-thaw cycles, and/or UV light. Generally, after aging T-peel tests are carried out to determine durability.

51.5

Conclusions

This review considered the current state of the art of the main issues involved in the bonding process of upper to sole in footwear manufacturing. The main innovations in the surface treatments of soles of different materials and the waterborne adhesives were addressed. Acknowledgment The chapter was not possible without the help of several people who work in my laboratory. The financial support from the Spanish Research Agency (CICYT, MCYT, MICINN), the Generalitat Valencia (Consellería de Educación, Cultura y Deporte, Consellería de

51

Shoe Industry

1527

Industria), and the University of Alicante is greatly appreciated. Last but not least, my deep recognition and gratitude to Antonia Armentia-Agüero (Toñi) for her love and for her help in several figure designs.

References Abbott SG (1992) Solvent-free adhesives for sole attaching. SATRA Bull 109 Abbott SG, Brewis DM, Manley NE, Mathieson I, Oliver NE (2003) Solvent-free bonding of shoesoling materials. Int J Adhes Adhes 23:225 Adams RD (ed) (2005) Adhesive bonding: science, technology and applications. Woodhead Publishing, Cambridge. Chapter 18 Alvarez-García S, Martín-Martínez JM (2015) Effect of the carbon black content on the thermal, rheological and mechanical properties of thermoplastic polyurethanes. J Adhes Sci Technol 29(11):1136 Amat-Amer JM (1999) Tecnología del Calzado, 3rd Diputación de Alicante, Elda, pp 133–146. Chapters 15 and 16 Arán-Aís F, Torró-Palau AM, Orgilés-Barceló AC, Martín-Martínez JM (2002) Characterization of thermoplastic polyurethane adhesives with different hard/soft segment ratio containing rosin resin as an internal tackifier. J Adhes Sci Technol 16:1431 Bahattab MA, Donate-Robles J, García-Pacios V, Martín-Martínez JM (2011) Characterization of polyurethane adhesives containing nanosilicas of different particle size. Int J Adhes Adhes 31:97 Bahattab MA, García-Pacios V, Donate-Robles J, Martín-Martínez JM (2012) Comparative properties of hydrophilic and hydrophobic fumed silica filled two-component polyurethane adhesives. J Adhes Sci Technol 26:303 Barni A, Levi M (2003) Aqueous polyurethane dispersions: a comparative study of polymerization processes. J Appl Polym Sci 88(3):716 Basaka GC, Bandyopadhyay A, Neogic S, Bhowmick AK (2011) Surface modification of argon/ oxygen plasma treated vulcanized ethylene propylene diene polymethylene surfaces for improved adhesion with natural rubber. Appl Surf Sci 257:2891 Brewis DM, Dahm RH (2003) Mechanistic studies of pretreatments for elastomers. In: Proceedings of Swiss Bonding 2003. Zurich, p 69 Brewis DM, Dahm RH, Mathieson I (1997) A new general method of pretreating polymers. J Mater Sci Lett 16(2):93 Cakić SM, Špírková M, Ristić IS, B-Simendić JK, M-Cincović M, Poręba R (2013) The waterborne polyurethane dispersions based on polycarbonate diol: effect of ionic content. Mater Chem Phys 138(1):277 Cantos-Delegido B, Martín-Martínez JM (2015) Treatment with Ar-O2 low-pressure plasma of vulcanized rubber sole containing noticeable amount of processing oils for improving adhesion to upper in shoe industry. J Adhes Sci Technol 29(13):1301 Carter AR (1969) Lacsol: problems and a new use. SATRA Bull 13(14):222 Carter AR (1971) Halogenation of thermoplastic rubbers. SATRA Bull 14(13):202 Cepeda-Jiménez CM, Pastor-Blas MM, Ferrándiz-Gómez TP, Martín-Martínez JM (2000) Surface characterization of vulcanized rubber treated with sulfuric acid and its adhesion to polyurethane adhesive. J Adhes 73:135 Cepeda-Jiménez CM, Pastor-Blas MM, Ferrándiz-Gómez TP, Martín-Martínez JM (2001) Influence of the styrene content of thermoplastic styrene-butadiene rubbers in the effectiveness of the treatment with sulfuric acid. Int J Adhes Adhes 21:161 Cepeda-Jiménez CM, Pastor-Blas MM, Martín-Martínez JM, Gottschalk P (2002) A new waterbased chemical treatment based on sodium dichloroisocyanurate (DCI) for rubber soles in footwear industry. J Adhes Sci Technol 16(3):257

1528

J. M. Martín-Martínez

Cepeda-Jiménez CM, Pastor-Blas MM, Martín-Martínez JM, Gottschalk P (2003a) Treatment of thermoplastic rubber with bleach as an alternative halogenation treatment in the footwear industry. J Adhes 79(3):207 Cepeda-Jiménez CM, Torregrosa-Maciá R, Martín-Martínez JM (2003b) Surface modifications of EVA copolymers by using RF oxidizing and non-oxidizing plasmas. Surf Coat Technol 174–175:94 Cuervo CR, Maldonado AJ (1984) Solution adhesives based on graft polymers of neoprene and methyl methacrylate. Du Pont Elastomers Bull Diputación de Alicante (1974) Estudio sobre control de calidad en la industria del calzado. Diputación de Alicante, Alicante, pp 53–115. Chapter III Dollhausen M (1985) Polyurethane adhesives. In: Oertel G (ed) Polyurethane handbook. Hanser, Munich, pp 548–562. Chapter 11 Dollhausen M (1988) Polyurethane adhesives based on Baycoll, Desmocoll and Desmodur, Technical report. Bayer AG, Leverkusen Donate-Robles J, Martín-Martínez JM (2011a) Comparative properties of thermoplastic polyurethane adhesive filled with natural or precipitated calcium carbonate. Macromol Symp 301:63 Donate-Robles J, Martín-Martínez JM (2011b) Addition of precipitated calcium carbonate filler to thermoplastic polyurethane adhesives. Int J Adhes Adhes 31:795 Donate-Robles J, Liauw CW, Martín-Martínez JM (2014) Flow micro-calorimetry and FTIR spectroscopy study of interfacial interactions in uncoated and coated calcium carbonate filled polyurethane adhesives. Macromol Symp 338:72 Extrand CW, Gent AN (1987) Contact angle and spectroscopic studies of chlorinated and unchlorinated natural rubber surfaces. Rubber Chem Technol 61:688 Fernández-García JC, Orgilés-Barceló AC, Martín-Martínez JM (1991) Halogenation of styrenebutadiene rubber to improve its adhesion to polyurethane. J Adhes Sci Technol 5:1065 Ferrándiz-Gómez T, Almela M, Martín-Martínez JM, Maldonado F, Orgilés-Barceló AC (1994) Effect of surface modification of leather on its joint strength with polyvinyl chloride. J Adhes Sci Technol 8:1043 Fisher W (1971) Einheitliche Prüfmethoden für Klebstoffe der Schuhindustrie. Adhäsion 15(1):372 Fisher W, Meuser H (1964) Mindestvoraussetzungen für die Prüfung von Schuklebstoffen. ShuhTech 55(9):1039 Frisch KC Jr (2002) Chemistry and technology of polyurethane adhesives. In: Chaudhury M, Pocius AV (eds) Adhesion science and engineering. surface chemistry and applications, vol 2. Elsevier, Amsterdam, pp 776–801. Chapter 16 García-Martín C, Andreu-Gómez V, Martín-Martínez JM (2010) Surface modification of vulcanized styrene-butadiene rubber with trichloroisocyanuric acid solutions of different active chlorine contents. Int J Adhes Adhes 30:550 García-Pacios V, Costa V, Colera M, Martín-Martínez JM (2010) Affect of polydispersity on the properties of waterborne polyurethane dispersions based on polycarbonate polyol. Int J Adhes Adhes 30:456 García-Pacios V, Colera M, Costa V, Martín-Martínez JM (2011a) Waterborne polyurethane dispersions obtained with polycarbonate of hexanediol intended for use as coatings. Prog Org Coat 71:136 García-Pacios V, Iwata Y, Colera M, Martín-Martínez JM (2011b) Influence of the solids content on the properties of waterborne polyurethane dispersions obtained with polycarbonate of hexanediol. Int J Adhes Adhes 31(8):787 García-Pacios V, Colera M, Iwata Y, Martín-Martínez JM (2013a) Incidence of the polyol nature in waterborne polyurethane dispersions on their performance as coatings on stainless steel. Prog Org Coat 76(12):1726 García-Pacios V, Jofre-Reche JA, Costa V, Colera M, Martín Martínez JM (2013b) Coatings prepared from waterborne polyurethane dispersions obtained with polycarbonates of 1,6-hexanediol of different molecular weights. Prog Org Coat 76(10):1484 Guggenberger SK (1990) Neoprene (polychloroprene)-based solvent and latex adhesives. In: Skeist I (ed) Handbook of adhesives, 3rd edn. Van Nostrand Reinhold, New York, pp 284–306. Chapter 15

51

Shoe Industry

1529

Guidetti G, Sacchetti G, Tribelhorn U (1992) In: Gum WF, Riese W, Ulrich H (eds) Footwear, in reaction polymers. Hanser, Munich, p 649. Chapter IV-H Hace D, Kovacevic V, Manoglovic D, Smit I (1990) The investigation of structural and morphological changes after the chlorination of rubber surfaces. Angew Makromol Chem 176:161 Harrington WF (1990) Elastomeric adhesives. In: Engineered materials handbook. Adhesives and sealants, 3. ASM International, Washington, DC, pp 143–150 Hu S, Luo X, Li Y (2015) Production of polyols and waterborne polyurethane dispersions from biodiesel-derived crude glycerol. J Appl Polym Sci 132(6):41425 Huang SM, Chen TK (2007) Effects of ion group content and polyol molecular weight on physical properties of HTPB-based waterborne poly(urethane-urea)s. J Appl Polym Sci 105(6):3794 Jang JY, Jhon YK, Cheong IW, Kim JH (2002) Effect of process variables on molecular weight and mechanical properties of water-based polyurethane dispersion. Colloids Surf A: Physicochem Eng Asp 196(2–3):135 Jaúregui-Beloqui B, Fernández-García JC, Orgilés-Barceló AC, Mahiques-Bujanda MM, MartínMartínez JM (1999) Rheological properties of thermoplastic polyurethane adhesive solutions containing fumed silicas of different surface areas. Int J Adhes Adhes 19:321 Jofre-Reche JA, Martín-Martínez JM (2013) Selective surface oxidation of ethylene-vinyl acetate and ethylene polymer blend by UV-ozone treatment. Int J Adhes Adhes 43:42 Keibal NA, Bondarenko SN, Kablov VF (2011) Modification of adhesive compositions based on polychloroprene with element-containing adhesion promoters. Polym Sci Ser D 4(4):267 Kelly DJ, McDonald JW (1963) Solution compatibility of neoprene with elastomers and resins. Du Pont Elastomers Bull Kim BK, Lee JC (1996) Waterborne polyurethanes and their properties. J Polym Sci A: Polym Chem 34(6):1095 Kim BK, Kim TK, Jeong HM (1994) Aqueous dispersion of polyurethane anionomers from H12MDI/IPDI, PCL, BD, and DMPA. J Appl Polym Sci 53(3):371 Kotrade Ph, Jofre-Reche JA, Martín-Martínez JM (2011) Surface modification of natural vulcanized rubbers containing excess of antiadherent moieties. In: Proceedings of adhesion 2011, London Kozakiewicz J (1991) Polyurethanes and isocyanates containing hydrophilic groups as potential components of water-borne adhesives. In: Allen KW (ed) Adhesion 15. Elsevier, London, pp 80–101. Chapter 6 Kueker P, Jeske W, Melchiors M (2016) Modern waterborne contact adhesives: 1k application property level. In: Proceedings of FEICA 2016 conference, Vienna. Paper 16-BOS-05-2 Landete-Ruiz MD, Martín-Martínez JM (2005) Surface modification of EVA copolymer by UV treatment. Int J Adhes Adhes 25:139 Landete-Ruiz MD, Martín-Martínez JM (2015) Improvement of adhesion and paint ability of EVA copolymers with different vinyl acetate contents by treatment with UV-ozone. Int J Adhes Adhes 58:34 Lawson DF, Kim KJ, Fritz TL (1996) Chemical modification of rubber surfaces: XPS survey of the reactions of trichloroisocyanuric acid at the surfaces of vulcanized elastomers. Rubber Chem Technol 69:245 Lee SK, Kim BK (2009) High solid and high stability waterborne polyurethane via ionic groups in soft segments and chain termini. J Colloid Interface Sci 336(1):208 Lijie H, Yongtao D, Zhiliang Z, Zhongsheng S, Zhihua S (2015) Synergistic effect of anionic and nonionic monomers on the synthesis of high solid content waterborne polyurethane. Colloids Surf A Physicochem Eng Asp 467(20):46 Liu X, Xu K, Liu H, Cai H, Su J, Fu Z, Guo Y, Chen M (2011) Preparation and properties of waterborne polyurethanes with natural dimer fatty acids based polyester polyol as soft segment. Prog Org Coat 72(4):612 Lyons D, Christell LA (1997) Waterborne polychloroprene adhesives. Adhes Sealants Ind 46 Maciá-Agulló TG, Fernández-García JC, Pastor-Sempere N, Orgilés-Barceló AC, Martín-Martínez JM (1992) Addition of silica to polyurethane adhesives. J Adhes 38:31

1530

J. M. Martín-Martínez

Madbouly SA, Otaigbe JU, Nanda AK, Wicks DA (2005) Rheological behavior of aqueous polyurethane dispersions: effects of solid content, degree of neutralization, chain extension, and temperature. Macromolecules 38(9):4014 Martin D (1971) Cork sandals. Weaknesses to guard against. SATRA Bull 14(13):195 Martínez-García A, Sánchez-Reche A, Martín-Martínez JM (2003a) Surface modifications on EVA treated with sulfuric acid. J Adhes 79(6):525 Martínez-García A, Sánchez-Reche A, Gisbert-Soler S, Cepeda-Jiménez CM, Torregrosa-Maciá R, Martín-Martínez JM (2003b) Treatment of EVA with corona discharge to improve its adhesion to polychloroprene adhesive. J Adhes Sci Technol 17(1):47 Martín-Martínez JM (2002) Rubber base adhesives. In: Chaudhury M, Pocius AV (eds) Adhesion science and engineering. surfaces chemistry and applications, vol 2. Elsevier, Amsterdam, pp 573–675. Chapter 13 Martín-Martínez JM (2008) Improving adhesion of rubber. In: Bhowmick AK (ed) Current topics in elastomers research. CRC Press, Boca Raton, pp 761–773 Moreno-Couranjou M, Choquet P, Guillot J, Migeon HN (2009) Surface modification of natural vulcanized rubbers by atmospheric dielectric barrier discharges plasma treatments. Plasma Process Polym 6:S397 Moyano MA, Martín-Martínez JM (2014) Surface treatment with UV-ozone to improve adhesion of vulcanized rubber formulated with an excess of processing oil. Int J Adhes Adhes 55:106 Mumtaz F, Zuber M, Zia KM, Jamil T, Hussain R (2013) Synthesis and properties of aqueous polyurethane dispersions: influence of molecular weight of polyethylene glycol. Korean J Chem Eng 30(12):2259 Oldfield D, Symes TEF (1983) Surface modification of elastomers for bonding. J Adhes 16:77 Orgilés-Calpena E, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C, Martín-Martínez JM (2009a) Addition of different amounts of a urethane-based thickener to waterborne polyurethane adhesive. Int J Adhes Adhes 29:309 Orgilés-Calpena E, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C, Martín-Martínez JM (2009b) Influence of the chemical structure of urethane-based thickeners on the properties of waterborne polyurethane adhesives. J Adhes 85:665 Orgilés-Calpena E, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C, Martín-Martínez JM (2009c) Effect of annealing on the properties of waterborne polyurethane adhesive containing urethanebased thickener. Int J Adhes Adhes 29:774 Orgilés-Calpena E, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C, Martín-Martínez JM (2009d) Addition of urethane-based thickener to waterborne polyurethane adhesives having different NCO/OH ratios and ionic groups contents. J Adhes Sci Technol 23:1953 Orgilés-Calpena E, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C (2016a) Influence of the chain extender nature on adhesives properties or polyurethane dispersions. J Dispers Sci Technol 33(1):147 Orgilés-Calpena E, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C (2016b) Novel polyurethane reactive hot melt adhesives based on polycarbonate polyols derived from CO2 for the footwear industry. Int J Adhes Adhes 70:218 Ortiz-Magán AB, Pastor-Blas MM, Ferrándiz-Gómez TP, Morant-Zacarés C, Martín-Martínez JM (2001) Surface modifications produced by N2 and O2 RF-plasma treatment on a synthetic vulcanised rubber. Plasmas Polym 6(1,2):81 Paiva RMM, Marques EAS, da Silva LFM, António CAC, Arán-Ais F (2016) Adhesives in the footwear industry. Proc IMechE Part L: J Mater: Des Appl 230(2):357 Pastor-Blas MM, Martín-Martínez JM, Dillard JG (1998) Surface characterization of synthetic vulcanized rubber treated with oxygen plasma. Surf Interf Anal 26:385 Pastor-Blas MM, Ferrándiz-Gómez TP, Martín-Martínez JM (2000) Chlorination of vulcanized styrene-butadiene rubber using solutions of trichloroisocyanuric acid in different solvents. J Adhes Sci Technol 14:561 Pastor-Sempere N, Fernández-García JC, Orgilés-Barceló AC, Torregrosa-Maciá R, Martín-Martínez JM (1995) Fumaric acid as a promoter of adhesion in vulcanized synthetic rubbers. J Adhes 50:25

51

Shoe Industry

1531

Penczek P, Nachtkamp K (1987) Resins used in adhesives. In: Frisch K, Reegen S (eds) Advances in urethane science and technology, vol 4. Technomic, Las Vegas, p 121 Pérez-Limiñana MA, Torró-Palau AM, Orgilés-Barceló AC, Martín-Martínez JM (2003) Modification of the rheological properties of polyurethanes by adding fumed silica: influence of the preparation procedure. Macromol Symp 194:161 Pérez-Limiñana MA, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C, Martín-Martínez JM (2006) Structure and properties of waterborne polyurethane adhesives obtained by different methods. J Adhes Sci Technol 20(6):519 Pérez-Limiñana MA, Arán-Aís F, Torró-Palau AM, Orgilés-Barceló C, Martín-Martínez JM (2007) Influence of the hard-to-soft segment ratio on the adhesion of water-borne polyurethane adhesive. J Adhes Sci Technol 21(8):755 Pettit D, Carter AR (1964) Adhesion of translucent rubber soling. SATRA Bull 11(2):17 Pettit D, Carter AR (1973) Behaviour of urethane adhesives on rubber surfaces. J Adhes 5:333 Radabutra S, Thanawan S, Amornsakchai T (2009) Chlorination and characterization of natural rubber and its adhesion to nitrile rubber. Eur Polym J 45:2017 Rahman MM (2013) Synthesis and properties of waterborne polyurethane adhesives: effect of chain extender of ethylene diamine, butanediol, and fluoro-butanediol. J Adhes Sci Technol 27(23):2592 Rahman MM, Kim H (2006) Synthesis and characterization of waterborne polyurethane adhesives containing different amount of ionic group (I). J Appl Polym Sci 102(6):5684 Rahman MM, Lee I, Chun H, Kim H, Park H (2014) Properties of waterborne polyurethanefluorinated marine coatings: the effect of different types of diisocyanates and tetrafluorobutanediol chain extender content. J Appl Polym Sci 131:39905 Romero-Sánchez MD, Martín-Martínez JM (2003) Treatment of vulcanised styrene-butadiene rubber (SBR) with mixtures of trichloroisocyanuric acid and fumaric acid. J Adhes 79:1111 Romero-Sánchez MD, Pastor-Blas MM, Martín-Martínez JM (2003a) Treatment of a styrenebutadiene-styrene rubber with corona discharge to improve the adhesion to polyurethane adhesive. Int J Adhes Adhes 23(1):49 Romero-Sánchez MD, Pastor-Blas MM, Martín-Martínez JM, Walzak MJ (2003b) UV treatment of synthetic styrene-butadiene-styrene rubber. J Adhes Sci Technol 17(1):25 Romero-Sánchez MD, Martín-Martínez JM (2008) UV-ozone surface treatment of SBS rubbers containing fillers: influence of the filler nature on the extent of surface modification and adhesion. J Adhes Sci Technol 22(2):147 SATRA-Preparation of leather uppers (1963) SATRA Bull 10(17):229 Schollenberger CS (1977) Polyurethane and isocyanate-based adhesives. In: Skeist I (ed) Handbook of adhesives, 3rd edn. Van Nostrand Reinhold, New York Souza EMM, da Costa W, Silva LGA, Wiebeck H (2016) Behavior of adhesion forces of the aqueous-based polychloroprene adhesive magnetically conditioned. J Adhes Sci Technol 30(15):1689 Sultan Nasar A, Srinivasan G, Mohan R, Radhakrishnan G (1998) Polyurethane solvent-based adhesives for footwear applications. J Adhes 68:21 Tanno T, Shibuya L (1967) Special behaviour of para tertiary phenol dialcohol in polychloroprene adhesives. Adhesives and Sealant Council Meeting Torregrosa-Coque R, Martín-Martínez JM (2011) Influence of the configuration of the plasma chamber on the surface modification of synthetic vulcanized rubber treated with low-pressure oxygen RF plasma. Plasma Process Polym 8:1080 Torregrosa-Coque R, Alvarez-García S, Martín-Martínez JM (2011a) Effect of temperature on the extent of migration of low molecular weight moieties to rubber surface. Int J Adhes Adhes 31(1):20 Torregrosa-Coque R, Alvarez-García S, Martín-Martínez JM (2011b) Migration of low molecular weight moiety at rubber-polyurethane interface: an ATR-IR study. Int J Adhes Adhes 31:389 Torregrosa-Coque R, Alvarez-García S, Martín-Martínez JM (2012) Migration of paraffin wax to sulphur vulcanized styrene-butadiene rubber (SBR) surface: effect of temperature. J Adhes Sci Technol 26(6):813

1532

J. M. Martín-Martínez

Tyczkowsky J, Krawczyk I, Wozniak B, Martín Martínez JM (2009) Low-pressure plasma chlorination of styrene-butadiene block copolymer for improved adhesion to polyurethane adhesives. Eur Polym J 45:1825 Vega-Baudrit J, Sibaja-Ballestero M, Nuñez S, Martín-Martínez JM (2009) Study of the relationship between nanoparticles of silica and thermoplastic polymer (TPU) in nanocomposites. J Nanotech Prog Int (JONPI) 1:24 Vélez-Pagés T (2003) Modificación de un serraje sin lijar por aplicación de un agente imprimante monocomponente para mejorar su adhesión a adhesivos de poliuretano. Master thesis, University of Alicante Vicent BJ, Natarajan B (2014) Waterborne polyurethane from polycaprolactone and tetramethylxylene diisocyanate: synthesis by varying NCO/OH ratio and its characterization as wood coatings. Open J Org Polym Mater 4(1):37 Vukov R (1984) Halogenation of butyl rubber – a model compound approach. Rubber Chem Technol 57(2):275 Wang H, Zhou Y, He M, Dai Z (2015) Effects of soft segments on the waterproof of anionic waterborne polyurethane. Colloid Polym Sci 293(3):875 Whitehouse RS (1986) Contact adhesives. In: Wake WC (ed) Synthetic adhesives and sealants. Wiley, Chichester, pp 1–29. Chapter 1 Yañez-Pacios A, Antoniac I, Martín-Martínez JM (2013) Surface modification and adhesion of vulcanized rubber containing an excess of paraffin wax treated with 2 wt% trichloroisocyanuric acid solution at different temperature. In: Proceeding of 36th adhesion society conference, Daytona Beach. Soft adhesives I -paper 1 Yang C, Yang H, Wen T, Wu M, Chang J (1999) Mixture design approaches to IPDI-H6XDI-XDI ternary diisocyanate-based waterborne polyurethanes. Polymer 40(4):871 Yang Z, Zhu Y, Peng F, Fu C (2014) Preparation and application of undecylenate based diol for bio-based waterborne polyurethane dispersion. Adv Mater Res 955–959:88 Yin L, Zhou H, Quan Y, Fang J, Chen Q (2012) Prompt modification of styrene-butadiene rubber surface with trichloroisocyanuric acid by increasing chlorination temperature. J Appl Polym Sci 124:661 Zhang K, Shen H, Zhang X, Lan R, Chen H (2009) Preparation and properties of a waterborne contact adhesive based on polychloroprene latex and styrene-acrylate emulsion blend. J Adhes Sci Technol 23(1):163 Zhang K, Huang C, Shen H, Chen H (2012) Modification of polychloroprene rubber latex by grafting polymerization and its application as a waterborne contact adhesive. J Adhes 88(2):119

52

Oil Industry Silvio de Barros, Luiz C. M. Meniconi, Valber A. Perrut, and Carlos E. Reuther de Siqueira

Contents 52.1 52.2

52.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Standards and Recommended Practices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.2.1 ISO 24817:2015: Petroleum, Petrochemical and Natural Gas Industries – Composite Repairs for Pipework – Qualification and Design, Installation, Testing and Inspection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.2.2 API 653: Tank Inspection, Repair, Alteration, and Reconstruction . . . . . . . . . . 52.2.3 ASME PCC-2–2015: Repair of Pressure Equipment and Piping . . . . . . . . . . . . Actual Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.1 FRP Pipe Connections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.2 Repair of Steel Pipes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.3 Ship Hull and Naval Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1534 1537

1538 1540 1541 1541 1542 1542 1543

S. de Barros (*) Department of Mechanical Engineering, Federal Center of Technological Education in Rio de Janeiro – CEFET/RJ, Rio de Janeiro, RJ, Brazil e-mail: [email protected] L. C. M. Meniconi Research and Development Center, CENPES, Petróleo Brasileiro S.A. – PETROBRAS, Rio de Janeiro, RJ, Brazil e-mail: [email protected] V. A. Perrut Research and Development Center, CENPES, Petróleo Brasileiro S.A. – PETROBRAS, Rio de Janeiro, RJ, Brazil Metallurgical and Materials Engineering Department, Federal University of Rio de Janeiro, Rio de Janeiro, Brazil e-mail: [email protected] C. E. Reuther de Siqueira Department of Offshore Engineering, Petróleo Brasileiro S.A. – PETROBRAS, Rio de Janeiro, RJ, Brazil e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_60

1533

1534

S. de Barros et al.

52.3.4 Metallic Storage Tanks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.5 Pipe Saddles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.3.6 Coamings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1548 1551 1552 1553 1554

Abstract

There has been an increase in using adhesives in the oil industry in the last years. The main advantage of using adhesive bonding is to avoid the fire risk concerning the traditional weld process. Adhesives also feature protection against corrosion which is a main concern in offshore installations. Different from other industrial applications, where adhesive bonding is used to manufacture new structures or some parts of them, the oil industry mainly uses adhesives to repair pipelines, pipework, and offshore structures. This chapter shows that the use of composites has become a real alternative to structural repairs in situ as these materials become more and more reliable. The chapter also brings a summary of the most applied standards and the best practices recommended to perform bonded repair in offshore units. Finally, some examples of actual applications of adhesives are presented.

52.1

Introduction

There has been an increase in the use of structural adhesives in the oil and gas industry recently. These materials have some inherent performance advantages over other types of joining, mainly welding. They avoid hot work process, which is one of the main advantages in some areas of the oil and gas industry. They also feature protection against corrosion, and the assembly process is faster compared to other types of joining. In oil installations, process plants, or ship platforms, some maintenance repairs are required during their operating life. Corroded pipes or sidewalls of ships and semisubmersible platforms are some examples. Usually the offshore production units must be sent to shipyards in order to perform repairs of metallic structures. Sometimes it is possible to stop the production for maintenance, but even in these cases an apparatus for inerting of tanks is needed for welding. It results in high direct and indirect costs. The use of composite repairs is an alternative to structural repairs in situ, avoiding the fire risk concerning the traditional weld process. Beyond the security issue, economical aspects should be considered. When welding is carried out in FPSO (floating, production, storage, and offloading) units, not only the tank to be repaired must be emptied but also the other surrounding tanks. It means that the storage capacity of the unit is compromised during the welding process. The reduction in storage capacity of the vessel leads to the reduction in the production of the unit causing financial losses. On the other hand, the use of composite patch repairs kept the reduction in storage capacity of the vessel to a minimum while the repair took place (Turton et al. 2005).

52

Oil Industry

1535

In what concerns corroded pipelines, adhesively bonded composite systems are also seen to be more economical than other repair alternatives. Duell et al. (2008) state that in this case, the use of composite repairs can be one fourth less expensive than welded steel sleeve repairs and three fourths less expensive than the complete replacement of the damaged segment. Adhesively bonded composite repairs can be considered an effective repair for mechanical damage, such as dent on the longitudinal weld, by increasing the fatigue life of the defective pipe. In most applications, the use a primer adhesive is recommended. But sometimes the lamination resin is at the same time responsible for the adhesion between the repaired metallic substrate and the composite. Depending on the surface conditions, it can be necessary to use a resin as a filler to provide a smooth surface. Thickened resins are often used to fill corroded areas before applying the first layer of the composite material (Rohem et al. 2016). Many commercial carbon fiber repair systems use epoxy resins (Alnaser and Keller 2015). An epoxy-based primer can also be applied to prevent galvanic corrosion and increase adhesion. The carbon fiber fabric is normally wetted with the epoxy resin, and the wetted fabric is wrapped by the hand around the pipe (Goertzen and Kessler 2007). When using glass fiber fabric, some authors reported that a vinyl ester resin was found to provide an attractive combination of on-site processability, metal–composite bond toughness, and overall resistance to corrosion (Mableson et al. 2000). Originally developed to be applied to onshore pipelines, glass fiber composite repairs were gradually proposed to be used in offshore pipelines such as risers. The complex combined loads acting on risers are one of main concerns, since it is completely different from onshore pipelines where the main load is internal pressure (Alexander and Ochoa 2010). Some examples that show the viability of the use of bonded composite repairs to the two most widely encountered damage scenarios in floating offshore units, fatigue cracking and thickness loss due to corrosion, can be found in the literature. McGeorge et al. (2009) have demonstrated that these repairs can be applied for full-scale trials even under severe sea conditions. In situ, submerged repairs of damaged piping using composites face several challenges. According to Dhanalakshmi et al. (1997), the primary challenge is the curing and material property development of composite systems that are based on liquid resins that wet out reinforcement and then are applied while submerged. The choice of the appropriate materials to be used is not the only issue. The entire repair operation must be carefully studied, and logistics issues, including the use of divers and the application of repairs to moving structures, must be considered (Mally et al. 2013). Da Costa Mattos et al. (2014) report the use of a polyurethane pre-impregnated bidirectional glass fiber composite applied to repair corroded pipelines (Fig. 1). According to them, the water-activated polyurethane resin reduces composite preparation time by 50% and is fully cured after 2 h at 24
C. Different from other industrial applications, where adhesive bonding is used to manufacture new structures or some parts of them, the oil industry mainly uses adhesives to repair pipelines, pipework, and offshore structures. In most of those

1536

S. de Barros et al.

Fig. 1 Polyurethane pre-impregnated repair

Fig. 2 Pipeline repair (courtesy of TEAM/ FURMANITE) https://twitter. com/furmanite/status/ 723008415182798848

situations, field conditions are very difficult needing much skill to accomplish the repair operation (Fig. 2). The use of sophisticated surface preparation methods before applying the adhesive is so laborious that the repair becomes too expensive. Mechanical surface preparation methods are more indicated rather than chemical modification of the surface. Abrasive methods are commonly applied to obtain a specific surface roughness (De Barros et al. 2015). Traditional abrasive wear and erosive wear, such as sandpapering and sandblasting, are still used, but they have been replaced by new methods. The bristle blasting system MBX (Monti Tools Inc., Houston, USA) has been widely used because it can be easily taken to field situations (De Barros et al. 2016). This system uses high-carbon steel wire bristles as shown in Fig. 3. A new abrasive method is the use of different kinds of Sponge Media™ (SpongeJet Inc., Newington, USA). These composite abrasives incorporate micro-abrasive particles within high-performance synthetic sponge (Fig. 4). The main advantage, compared to sandblasting, for instance, is that the synthetic sponge can capture most of the contaminant particles resulting from the blasting operation.

52

Oil Industry

1537

Fig. 3 MBX system

Fig. 4 Sponge-Jet system

This chapter is structured in three sections. After this introduction showing not only the advantages but also the challengers involving the use of adhesives in oil and gas industry, the next section brings a summary of the most applied standards and the best practices recommended to perform bonded repair in offshore units. The chapter ends with six examples to illustrate actual applications that have been proposed for adhesives in field situations.

52.2

Standards and Recommended Practices

In offshore structures, an important document, approved by classification societies, is “Design, Fabrication, Operation and Qualification of Bonded Repair of Steel Structures” – DNVGL-RP-C301 from DNV GL (Det Norske Veritas – Germanischer Lloyd). This Recommended Practice (RP) is based upon a project guideline developed within the Joint Industry Project (JIP) “Qualification of Adhesive Bonding in Structural Repair of FPSO’s” sponsored by ConocoPhillips, Norsk

1538

S. de Barros et al.

Hydro, Petrobras, Petronas, Shell, and Statoil. The main focus of this document was placed on FPSOs and similar ship-like floating structures, but it can be also equally applicable to the repair of other types of offshore units. The main objectives of this document are to: – Provide an international RP of repairs practices for floating units using composite adhesive patches – Serve as a technical reference document in contractual matters – Reflect the state-of-the-art and consensus on accepted industry practice and server as a RP for composite adhesive patch repair design, analysis, application, and maintenance Section 7 of this RP describes materials and bonding agents and points out special considerations for adhesive composite patches on steel substrates. In what concerns composite repairs, a complementary document is the DNVGLOS-C501 – Composite Components. It contains more information on the use of adhesives in composite laminated joints. A list of important standards is summarized following.

52.2.1 ISO 24817:2015: Petroleum, Petrochemical and Natural Gas Industries – Composite Repairs for Pipework – Qualification and Design, Installation, Testing and Inspection The objective of this standard is to ensure that pipework, pipelines, tanks, and vessels repaired using composite systems that are qualified, designed, installed, and inspected using this standard will meet the specified performance requirements. Repair systems are designed for use within the petroleum, petrochemical, and natural gas industries and also within utility service applications. The main users of this standard will be plant and equipment owners of the pipework and vessels, design contractors, suppliers contracted to provide the repair system, certifying authorities, installation, maintenance, and inspection contractors. The qualification, design, installation, testing, and inspection procedures for composite repair systems in this standard cover situations involving the repair of damage commonly encountered in oil, gas, utility pipework systems, and vessels. The procedures are also applicable to the repair of pipelines, caissons, and storage tanks with appropriate consideration. Procedures described in this standard cover the repair of metallic and GRP pipework, pipework components, and pipelines originally designed in accordance with a variety of standards, including ISO 15649, ISO 13623, ISO 14692, ASME B31.1, ASME B31.3, ASME B31.4, ASME B31.8, and BS 8010. This standard is not a defect assessment standard. Within this standard, no statements are made regarding whether a specific defect is acceptable or unacceptable for repair. The standard assumes that a defect assessment has already been performed to, for example, ASME B31G or API RP 579. The starting point

52

Oil Industry

1539

for this standard is that a decision has been taken to repair a given defect with a composite repair system and the output from the defect assessment, e.g., MAWP (maximum allowed work pressure) or minimum remaining wall thickness is used as input for the repair design. The main concern of this standard is with the subsequent activities of repair qualification, design, installation, and inspection. Repair systems are applied to restore structural integrity. The following repair situations are addressed: – External corrosion, where the defect is or is not through wall. In this case, the application of a repair system will usually arrest further deterioration. – External damage such as dents, gouges, and fretting (at supports). – Internal corrosion, erosion, where the defect is or is not through wall. In this case, corrosion and/or erosion can continue after application of a repair system, and therefore the design of the repair system shall take this into account, i.e., the size of the defect at the end of the required design life of the repair should be taken as the size of the defect when designing the repair. – Crack-like defects, where the defect is or is not through wall. It is a requirement that the length of the crack is known and will not increase during the lifetime of the repair. For through-wall cracks, the crack should be modeled as a type B defect (through-wall defect), either a circumferential or axial slot (depending on the crack orientation). For non-through-wall cracks, the crack should be modeled as type A defect (non-through-wall defect). – Strengthening and/or stiffening in local areas. Services that are covered within the scope of this standard include those normally found in an oil and gas production or processing installation. These include the following: – Utility fluid, diesel, seawater, and air – Chemicals (liquids) – Production fluids, including liquid hydrocarbons, gaseous hydrocarbons, and gas condensates The upper temperature limit is defined in this standard for different class of repairs in a detailed table. The lower temperature limit is dependent on the type of repair laminate being used. This limit is determined by the design requirements presented in detailed in a table of this standard. The upper pressure limit is a function of defect type (internal, external, or through wall), defect dimensions (depth and extent), pipe diameter, design temperature, and repair design lifetime. Therefore, a unique number cannot be quoted, but rather the limit is derived for a given set of conditions by calculations in accordance with this standard using the qualification test data. The lower pressure limit, e.g., vacuum conditions, is determined by the design requirements based on external loads. All equations for thickness design are described on this standard.

1540

S. de Barros et al.

The composite materials constituting the repair laminate considered within this standard are typically those with aramid (AFRP), carbon (CFRP), glass (GRP), or polyester (or similar material) fiber reinforcement in a polyester, vinyl ester, epoxy, or polyurethane polymer matrix. Other fiber and matrix types are also permissible once qualified. The pipework and vessel substrates considered within the standard include carbon steel, 6 moly steel, stainless steel, duplex steel, super duplex steel, GRP, cunifer, aluminum, galvanized steel, and titanium. Careful consideration is required before repair of GRP lines because the damage in the pipe may be more extensive than is visible on the surface and may affect a longer length of the pipe than is immediately obvious; advice of the GRP pipe manufacture and repair system supplier shall be sought before a repair is installed.

52.2.2 API 653: Tank Inspection, Repair, Alteration, and Reconstruction This standard covers steel storage tanks built to API 650 and its predecessor API 12C. It provides minimum requirements for maintaining the integrity of such tanks after they have been placed in service and addresses inspection, repair, alteration, relocation, and reconstruction. The scope is limited to the tank foundation, bottom, shell, structure, roof, attached appurtenances, and nozzles to the face of the first flange, first threaded joint, or first welding-end connection. Many of the design, welding, examination, and material requirements of API 650 can be applied in the maintenance inspection, rating, repair, and alteration of in-service tanks. In the case of apparent conflicts between the requirements of this standard and API 650 or its predecessor API 12C, this standard shall govern for tanks that have been placed in service. This standard employs the principles of API 650; however, storage tank owner/ operators, based on consideration of specific construction and operating details, may apply this standard to any steel tank constructed in accordance with a tank specification. This standard is intended for use by organizations that maintain or have access to engineering and inspection personnel technically trained and experienced in tank design, fabrication, repair, construction, and inspection. This standard does not contain rules or guidelines to cover all the varied conditions which may occur in an existing tank. When design and construction details are not given, and are not available in the as-built standard, details that will provide a level of integrity equal to the level provided by the current edition of API 650 must be used. This standard recognizes Fitness-for-Service assessment concepts for evaluating in-service degradation of pressure-containing components. API 579-1/ASME FFS-1, Fitness-for-Service, provides detailed assessment procedures or acceptance criteria for specific types of degradation referenced in this standard. When this standard does not provide specific evaluation procedures or acceptance criteria for

52

Oil Industry

1541

a specific type of degradation or when this standard explicitly allows the use of Fitness-for-Service criteria, API 579-1/ASME FFS-1 may be used to evaluate the various types of degradation or test requirements addressed in this standard.

52.2.3 ASME PCC-2–2015: Repair of Pressure Equipment and Piping This standard provides methods for repair of equipment and piping within the scope of ASME Pressure Technology Codes and Standards after they have been placed in service. These repair methods include relevant design, fabrication, examination, and testing practices and may be temporary or permanent, depending on the circumstances. The methods provided in this standard address the repair of components when repair is deemed necessary based on appropriate inspection and flaw assessment. These inspection and flaw evaluation methods are not covered in this standard, but are covered in other post-construction codes and standards. This document is divided into five parts: Part 1 covers the scope, organization, and intent and is applicable to all articles in this standard and also provides guidance for the applicability of repair methods listed in this standard. Part 2 covers repair methods and techniques that include the use of welding, brazing, soldering, or other methods involving metal deposit. Part 3 covers mechanical repairs, with or without sealant, such as bolted clamps or fixtures, and includes all repair methods not covered in Part 2 or Part 4. Part 4 covers repairs using nonmetallic means, such as nonmetallic liners and wraps, and bonding (e.g., joining by epoxy), including bonding of metallic components. Part 5 covers examination and testing methods and techniques. This standard provides technical information, procedures, and recommendations for repair methods that were determined by consensus to be recognized and generally accepted good engineering practice. Where equipment repair is subject to jurisdictional regulation, jurisdictional approvals may be required.

52.3

Actual Applications

Steel and fiber-reinforced plastic (FRP) are commonly used materials in the oil industry. Sometimes there is a need to join them using adhesive bonding. There are three kinds of situations where adhesives have been used, and their application depends on the substrates to be joined – FRP/FRP joint, FRP/steel joint, or steel/steel joint: FRP/FRP joints – FRP pipe connections – Nonmetallic pipe saddle supports

1542

S. de Barros et al.

FRP/steel joint – Repair of steel pipes – Repair of naval structures – Repair of storage tanks Steel/steel joint – Metallic pipe saddle support In what follows, some examples of actual applications of adhesives are presented considering these three types of substrate material combination.

52.3.1 FRP Pipe Connections Two joining systems are normally adopted for the assembly of offshore production plant pipework, namely, socket/spigot and laminated joints, as displayed in Fig. 5. The first one is easier to apply in the field and has a self-aligning design but depends on good workmanship to provide a safe and sound connection. If the adhesive joint surfaces are badly cleaned or the adhesive itself is not properly prepared, the joint will have a high probability of failure in service. The laminated type, on the other hand, is considered to be more reliable as it can be made as thick and long as required, to provide the necessary reinforcement. As several layers of reinforcing cloth and resin are applied from inside out, its final resistance is less variable. In the end, the choice between one type and another is much dependent on the contractor design philosophy.

52.3.2 Repair of Steel Pipes Composite sleeve repair systems, now of widespread use in the market, were originally developed for the reinforcement of corroded pipelines and continue to gain wider acceptance in the pipeline industry for repairing a range of pipeline anomalies (Meniconi et al. 2002). Figure 6 presents the visual aspect of steel pipes before and after the application of the composite repair. ISO/TS 24817 (2015) and ASME PCC-2 are the codes that cover all aspects of the technology, encompassing material qualification, repair design, installation, testing, and inspection. The documents allow reinforcement of a defected metallic pipe due to presence of general wall thinning, local wall thinning, pitting, gouges, blisters, laminations, cracks, and through-wall penetration. The defects can be located at the external or internal side of the pipe wall. Concerning the last case, the effectiveness of the repair is time-dependent, according to the evolution of the defect. The kinds of services covered are those normally found on hydrocarbon processing plants, like utility fluids, seawater, air, chemicals, production fluids, and

52

Oil Industry

1543

a pipe with socket end

adhesive fillet

pipe with spigot end

Socket and spigot tapered ends

laminate overlap length

pipe wall

b

beveled joint

Laminated Fig. 5 Two basic types of FRP pipe connections. (a) Socket and spigot. (b) Laminated

liquid and gaseous hydrocarbons. The composite materials most commonly used for the repair laminates are glass or carbon fibers for reinforcement and vinyl ester or epoxy resins for matrix. The upper limits for pressures and temperatures are obtained from qualification tests. The ISO document establishes three classes of repairs. The higher the class, the more demanding is the application and the information required, concerning material data, design capability, surface preparation, and degradation with time. In relation to lifetime, repairs are classified as short term (2 years), medium term (10 years), or long term (20 years).

52.3.3 Ship Hull and Naval Structures Offshore production structures like floating, production, storage, and offloading (FPSO) vessels are designed to remain in station for 25 years or more. This is a major deviation from the traditional ship maintenance scheme, which involves dry docking every 5 years or so, for overhaul maintenance. Due to this scenario, in place

1544

S. de Barros et al.

Fig. 6 Steel pipes before (a) and after (b) the application of the composite repair

repair techniques are an excellent option to restore structural integrity without the need of interrupting production. Composite patch repairs are one of those techniques, because no hot work is involved, turning the operation intrinsically safe. Many successful application cases of this technique are reported (Grabovac and Whittaker 2009). Figure 7 shows the first composite repair applied to the metallic hull of a ship platform operating in offshore Brazil. The repair procedure based on reliability design method, as proposed by DNVGL-RP-C301, provides repair solutions that assure a minimum safety level in relation to a given limit state. For a specific repair, there is a need to identify all the possible failure modes that can cause bad performance in relation to its functional requirements. Examples of composite repair failure modes are debonding, matrix cracking, delamination, fiber rupture, fluid permeation, etc. Examples of functional requirements are load-bearing capacity, level of stiffness, fluid tightness, chemical inertness, erosion resistance, fire survival, etc. The concept of risk adopted is the

52

Oil Industry

1545

Fig. 7 Composite repair applied to a ship hull

usual combination of probability and consequence of failure. Statistical methods are applied to guarantee a minimum reliability, corresponding to a probability of failure of 103 once the repair is completed. Composite bonded repairs are separated in four reliability repair classes with different levels of effort and optimization: – Class 0 – Emergency repairs, done empirically without integrity and efficiency qualification by any code or standard. – Class I – Noncritical repairs qualified by reliable estimates based on short-term and small-scale tests. Repair efficiency is evaluated by standard statistical methods. – Class II – Noncritical repairs qualified by accelerated time degradation tests together with long-term behavior simulation models. The reliability level is higher than that of class I repairs. Classes I and II are the ones considered by the design document. – Class III – Repairs qualified by long-term degradation tests, designed to assure the desired reliability level along the whole life of the structure. Given the little experience accumulated so far with composite bonded repairs, class III is not covered by DNVGL-RP-C301, yet. The design equations criteria are formulated in terms of load and resistance characteristic values, affected both by partial safety factors, which take into account the random nature of design input data. Basically, the design process aims to satisfy the following inequality: γ F :γ Sd :Sk 

RK γ M :γ Rd

(1)

1546

S. de Barros et al.

where Sk is the characteristic effect of the load applied to the repaired structure that can lead to the failure mode under analysis. It is multiplied by partial load factors γF, associated to the variability of loading, and γSd related to the uncertainties of the adopted model of structural behavior. Similarly, Rk is the characteristic structural resistance, divided by the partial resistance factors γM, referred to time and sampling variations of material strength and γRd, concerning geometric aspects and fabrication processes. When dealing with the static capacity of the adhesive layer between the metallic substrate and the composite patch, the equation criterion above is written as: J

RII γM

(2)

The load J, expressed as the energy per unit area of resulting adhesive debonding introduced by mechanical factored load combination, needs to be smaller than or at most equal to the mode II toughness of the interface, RII, divided by the partial material factor γM, obtained from the Recommended Practice. For the first applications of DNVGL-RP-C301, only class I repairs were considered. This class of repair is qualified by small-scale, short-term tests which results are statistically treated to make sure that the necessary repair capacity is achieved within design lifetime. This generally implies that larger partial safety factors are adopted, to account for aging effects. A minimum reliability level of 103 is achieved when all the requirements are fulfilled. The first scenario chosen for the application of composite repairs was the reposition of metallic thickness loss due to corrosion, given its simpler nature. Nevertheless, this kind of repair is classified as structural, because it attracts mechanical loading once deployed. The repair is also two-dimensional in nature, making it easier to design and qualify. It can be applied whenever there is sufficient lateral area to transfer the load from the metallic substrate to the composite patch repair, via shear stresses in the adhesive layer. Furthermore, it is a type of remedial action with great demand from an aging production platform fleet. The two-dimensional repair qualified is rated “simple” according to the design document because it can be calculated by means of a basic theoretical model, since it fulfilled four requisites: – The lamination directions coincide with those of the main loading. – The overlap length L is long enough to provide regions in the bondline at the borders of the repair where the shear stresses at the adhesive were equal to zero, according to Fig. 8.

Fig. 8 Modified double-strap specimen where notches are used to obtain the desired overlap length

52

Oil Industry

1547

This is achieved when the following expression holds: L>2

F or L > 2:em w:τP

(3)

where F is the repair failure load, estimated from simple models, and w is the composite patch width. The parameter τp is the adhesive critical plastic shear stress and em is the maximum adhesive layer overlap effective length, obtained from double-lap shear tests (Meniconi et al. 2010). Shorter overlap lengths can be accepted if: – It is demonstrated that the minimum shear stress in the adhesive is smaller than 10% of the interface critical plastic shear stress. – Bending stresses are limited, resulting in reduced adhesive peel stresses. Adhesives normally display limited strength against this kind of loading. – Adhesive critical plastic shear stress is smaller than the out-of-plane shear strength of the repair laminate, estimated in a conservative way. This implies a failure mode by debonding of the adhesive layer rather than by delamination of the composite repair laminate. For the bondline, there are two distinct failure modes that need to be considered. The most fundamental one is debonding propagating from the center at the discontinuity in steel. This corresponds to the case where a cracked steel plate is repaired. The other one is debonding propagating from the edges of a patch repair, but in general this only occurs after failure in steel by yielding had already developed thus showing that bondline fracture toughness is not the governing factor for the capacity of the repaired plate. Common practice is to provide repair border edge thickness tapering widths of 10–20 times the total laminate thickness, which are generally enough to efficiently reduce the out-of-plane stresses that would trigger debonding from there. The thickness reduction at the boundaries of composite patch repairs also helps to protect them from lateral impacts and other contact damage. Table 1 lists the functional requirements proposed for a patch repair designed for metallic plating thickness loss reposition. Some demonstration repairs already performed in the field showed that the design methodology proposed by DNVGL-RP-C301 proved feasible, at least for the Table 1 Functional requirements to be achieved by a thickness reposition patch repair Functional requirements Strength Stiffness Corrosion resistance Temperature resistance Electric insulation

Comments Restore steel plating strength Restore steel plating stiffness Not to be degraded by exposition to ambient Material properties still acceptable at 60

C To prevent the formation of a galvanic pair

1548

S. de Barros et al.

simpler case of restoration of steel thickness loss greater than the maximum allowable by the classification society of the offshore platform. Even though ambient conditions are generally far from ideal, it was possible to provide enough cleanness, illumination, temperature, and humidity to make manual lamination work feasible. The importance of adequate surface preparation should never be underestimated. It is strongly advised to pay special attention in planning surface preparation and execution of the first sealing layer of the repair. The previous cleaning and drying of the repair area are also important. Grit blasting is essential to get the necessary surface profile. Polymer-coated grit helped in performing a clean operation, but conventional grit will also suffice, if one considers that normally the areas to be repaired are not too extensive. The volume of dust to be aspirated afterward will not be excessive. It must be assured that the initial tasks – surface preparation, cleaning, and sealing – are performed sequentially without any delay, given the need of a dry, rust-free surface to obtain good adhesion between the patch repair and the metallic substrate. This is the optimal situation, but it demands good logistics to move and replace quickly the equipment and personnel involved in those tasks. It should be remembered that the repair location can be situated in a remote and difficult-toaccess part of the platform. Inspections executed afterward confirmed that the repairs already performed remain intact. More inspections will occur along the years, to make sure this situation continues. As this is a new development, only repair performance during the entire platform operational lifetime in real operational conditions will assure that the proposed solution works (Meniconi et al. 2014).

52.3.4 Metallic Storage Tanks Metallic storage tanks are cylindrical structures which hold oil and their derivatives and are very important to the supply chain in the oil industry. These tanks are made of thin steel plates welded together, and the large-capacity ones often are of the floating roof type. During operation, corrosion wall thickness loss frequently occurs at the inner metallic surface of the tank shell in contact with the roof seal, where rainwater tends to accumulate. The thickness loss can happen in wide areas, as the roof position changes during operation. Figure 9 shows defects found inside steel tank plating, characterized by general wall loss due to corrosion. This corrosion mechanism may lead to a value less than acceptable minimum corroded wall thickness according to API 579 and can affect the structural integrity of the tank. In this case, welding repair is the most commonly used to restore integrity. This repair technique is quite complex and expensive, and hot working is necessary. Special care is needed to prevent severe deformation of the slender tank wall as it is heated during the welding process. In this context, composite repair seems to be an excellent alternative as this technology is very simple, fast, and costeffective. The repair is accomplished by applying layers of carbon or glass fiber, at the outer surface of the tank wall, opposite to the corroded area. This is to avoid

52

Oil Industry

1549

Fig. 9 Defects found in steel tank: inside view (left) and corroded surface detail (right)

Fig. 10 Repairing a storage tank: repair overview (left) and repair application (right)

interference with the correct operation of the floating roof. An example of a repair in a storage tank can be seen in Fig. 10. Differently from pipelines, there is no specific code to cover the repair of storage tanks by composite materials. Due to this factor, a design methodology for composite repairs of corroded storage tanks was proposed that considers the requirements of API 653, ISO/TS 24817 and DNVGL-RP-C301. Two main factors turn composite repairs very suitable for this application. Firstly, the wide tank surface provides an extensive area to transfer loading from the steel plates to the composite laminates by means of shear stress in the adhesive layer between the two materials. Secondly, the steel tank wall plating is in a state of plane stress, mainly unidirectional, adequately resisted by composite laminates, which shows high stiffness and strength in lamination plane and fiber direction. Moreover, the materials, equipment, and workmanship involved are quite similar to those needed for painting operations, very common in this industrial environment. The properties of the adhesive interface already mentioned, the critical plastic shear stress τp and the maximum effective overlap length em, define a limit in steel thickness loss that can be reinstated by a patch repair. Figure 11 illustrates this

1550

S. de Barros et al. Tlam em

σ

σ

Tremain Tlost

Tnom

Fig. 11 Schematic description of a patch repair, with relevant parameters

principle, by means of a segment of tank shell with unitary height. There is a corrosion thickness loss at the inside surface of the shell, remediated by a composite laminate reinforcement at the corresponding outside surface. The local circumferential stress at the non-corroded plating away from the defected area is σ. The composite laminate thickness Tlam is readily obtained as the main design drive is to reinstate the stiffness loss caused by corrosion: Elam :T lam ¼ Esteel :T lost

)

T lam ¼

Esteel :T lost Elam

(4)

The laminate elastic modulus Elam (as well as all other composite properties) are referred to the circumferential direction of the tank wall. Once the circumferential stress is transferred from the metallic wall to the composite laminate through shear stresses at the adhesive interface, the highest unitary load fp that can be transmitted is given by: f p ¼ τp :em ¼ σ lam :T lam

(5)

The equilibrium of resultant forces implies that: σ:T nom ¼ σ remain :T remain þ σ lam :T lam σ:T nom ¼ σ remain :ðT nom  T lost Þ þ τp :em so σ remain :T lost ¼ T nom :ðσ remain  σ Þ þ τp :em

(6)

The formulation above considers linear elastic behavior. Since the elastic modulus of the laminate is usually smaller that of steel, the laminate thickness turns considerably greater than the steel thickness loss. Knowing composite laminates typical elastic moduli, high-strength carbon fiber is applied instead of fiberglass, to keep laminate thickness feasible to be applied in practice. As high-strength carbon is the chosen repair material, once the stiffness replacement criterion is fulfilled, the composite laminate circumferential stress σlam results low when compared with the corresponding strength. For this reason, it is possible to allow the stress acting on the

52

Oil Industry

1551

remaining steel wall, σremain, to reach steel yield strength Sy. The design code for atmospheric storage tanks limits the average circumferential stress at the shell, σ, to a typical value of 0.75 Sy, during hydrostatic test. Considering this limit, the expression for the maximum steel thickness loss prone to be replaced by a patch repair then reads: T lost ¼ 0:25 T nom þ τp :em =Sy

(7)

Typical figures are nominal steel plating thicknesses of 19 mm (at the level corrosion normally occurs), critical adhesive plastic shear stresses (τp) of 10 MPa, maximum effective overlap lengths em of 90 mm, and steel yield stresses of 250 MPa. In conclusion, up to 8.3 mm of steel thickness loss can be replaced by a composite patch repair laminate. This limit covers most corrosion cases that happen in practice. The major elastic modulus of the laminate is of the order of 100 GPa, compared with a modulus of 210 GPa for steel, so its thickness would be around 17 mm, roughly double the steel thickness lost. The average longitudinal stress at the composite laminate turns to be about 50 MPa, one sixth of its ultimate tensile limit of 300 MPa. The low stresses (and strains) at the laminate are important contributions for a long operational life of the repair and, moreover, of the adhesive interface.

52.3.5 Pipe Saddles A pipe saddle is a structure used to support the pipe by transmitting the load or forces to the adjacent structure. It avoids direct contact between the pipe and the support, reducing stress concentration. Some of them are designed to allow longitudinal movement due to thermal expansion and contraction of pipe and are also designed to prevent damage to pipe thermal insulation. Figure 12 shows examples of adhesively bonding saddles with different substrates. Carbon steel is the main material used for manufacturing the saddles. Stainless steel is another material that can be found. Saddles are installed just below the pipe mainly by welding, and this process can lead to some problems related to corrosion and metallurgical defects. These areas are susceptible to higher corrosion rates because of crevice and galvanic effects, and historically inspecting them has involved costly and time-consuming operations to lift the pipe or to cut the supports. Corrosion under support is a well-known issue within all oil and gas assets around the world. The external corrosion forms itself between the pipe support and flow line and tends to be nonuniform and unpredictable. The corrosion is often caused by movement of the pipe on the support which damages the protective coating on the pipe. This leaves the exposed metal free to corrosion attack from the environment in an area where it cannot be detected and in most cases not even protected. The general problem with corrosion under pipe supports is that it is very hard to inspect the remaining wall thickness. In many cases the pipe would have to be lifted to understand the real impact of the corrosion attack. This is of course very labor intensive, risky, and expensive as the plant operations would have to be shut down.

1552

S. de Barros et al.

Fig. 12 Pipe saddles: (a) steel pipe and steel saddle (b) steel pipe and FRP saddle (c) FRP pipe and FRP saddle

Alternatively, pipe saddle can be glued onto the pipe surface instead of welding. In addition, it is possible to use pipe saddles made of nonmetallic materials as FRP. They provide many advantages to the system: to prevent from further corrosion on support area, to avoid galvanic effect, to not require hot work to install, and to be faster and easier to install compared to welding. It is important to mention that the adhesive must be thoroughly applied to avoid crevice corrosion caused by trapped water (Fig. 13).

52.3.6 Coamings Coaming is a raised metal border around a ship’s deck or an equipment, designed to avoid oil leaking or chemical fluids into the sea. Welding is the main process used to place coamings over the metallic surface (Fig. 14). The installation of these structures demands lots of time, workmanship, and logistic to perform welding. Furthermore, hot work can damage surround and internal paintings, which makes necessary to repaint these affected areas. This activity can be very complex, especially for internal surfaces where it might be required to install scaffolding inside the oil tank, for instance. When installed at the beginning of the assembly process, these coamings create a physical barrier making it difficult for technicians, equipment, and assembly devices to move over the platform, which is still in the test and integration phase.

52

Oil Industry

1553

Fig. 13 Adhesive application: over pipe surface (left) and the saddle (right)

Fig. 14 Coaming: around an equipment (left) and detail of welded coaming (right)

In this context, the possibility of using alternative technologies, such as adhesive bonding, to the traditional method (welding), arises, allowing to increase productivity and to prevent rework in the painting process.

52.4

Conclusion

In this chapter, it was shown that the application of adhesives in many different situations in the oil industry has increased in the last years. The use of composites has become a real alternative to structural repairs as these materials become more and more reliable. In this sense, the development of adhesive bonding methods has become inexorable. On the other hand, it was shown that field conditions do not allow the use of sophisticated process to perform the adhesive bonding. It makes necessary to adjust the surface preparations methods to these unfavorable situations, for instance. The chapter also brings a summary of the most applied standards and the best practices recommended to perform bonded repair in offshore units. Besides, some examples of actual applications of adhesives are presented. In what concerns the materials to be bonded, the three main situations where the adhesives have been used

1554

S. de Barros et al.

in oil industry are presented, namely, FRP/FRP joint, FRP/steel joint, and steel/steel joint. Despite of all the challengers involving the use of adhesives in field situations, the advantages of applying this technique are pushing engineers and researchers to propose the necessary solutions. There is no doubt that adhesive bonding is becoming essential as a safe joint technique in oil industry.

References Alexander C, Ochoa OO (2010) Extending onshore pipeline repair to offshore steel risers with carbon–fiber reinforced composites. Compos Struct 92(2):499–507. ISSN 0263-8223, https:// doi.org/10.1016/j.compstruct.2009.08.034 Alnaser IA, Keller MW (2015) Comparison of coupon and full-scale determination of energy release rate for bonded composite repairs of pressure equipment. Eng Fract Mech 146:31–40. ISSN 0013-7944, https://doi.org/10.1016/j.engfracmech.2015.07.023. (http://www.sciencedirect. com/science/article/pii/S0013794415003902) API 579-1/ASME FFS-1 Fitness-For-Service (n.d.) API 653 (n.d.) Tank Inspection, repair, alteration, and reconstruction ASMEPCC-2–2015 (n.d.) Repair of pressure equipment and piping da Costa Mattos HS, Reis JML, Paim LM, da Silva ML, Amorim FC, Perrut VA (2014) Analysis of a glass fibre reinforced polyurethane composite repair system for corroded pipelines at elevated temperatures. Compos Struct 114:117–123. ISSN 0263-8223, https://doi.org/ 10.1016/j.compstruct.2014.04.015. (http://www.sciencedirect.com/science/article/pii/S026 3822314001822) De Barros S, Kenedi PP, Ferreira SM, Budhe S, Bernardino AJ, Souza LFG (2015) Influence of mechanical surface treatment on fatigue life of bonded joints. J Adhes. https://doi.org/10.1080/ 00218464.2015.1122531. https://doi.org/10.1080/00218464.2015.1122531 De Barros S, Banea MD, Budhe S, de Siqueira CER, Lobão BSP, Souza LFG (2016) Experimental analysis of metal-composite repair of floating offshore units (FPSO). J Adhes. https://doi.org/ 10.1080/00218464.2016.1177514. https://doi.org/10.1080/00218464.2016.1177514 Design, fabrication, operation and qualification of bonded repair of steel structures – DNVGL-RPC301 from DNV GL (Det Norske Veritas – Germanischer Lloyd) (n.d.) Dhanalakshmi M, Maruthan K, Jayakrishnan P, Rengaswamy NS (1997) Coatings for underwater and wet surface application. Anti-Corros Methods Mater 44(6):393–399 DNVGL-OS-C501 – Composite Components (n.d.) Duell JM, Wilson JM, Kessler MR (2008) Analysis of a carbon composite overwrap pipeline repair system. Int J Press Vessel Pip 85(11):782–788. ISSN 0308-0161, https://doi.org/10.1016/j. ijpvp.2008.08.001. (http://www.sciencedirect.com/science/article/pii/S0308016108001026) Goertzen WK, Kessler MR (2007) Dynamic mechanical analysis of carbon/epoxy composites for structural pipeline repair. Compos Part B 38(1, 1):–9. ISSN 1359-8368, https://doi. org/10.1016/j.compositesb.2006.06.002. (http://www.sciencedirect.com/science/article/ pii/S1359836806000825) Grabovac I, Whittaker D (2009) Application of bonded composites in the repair of ships structures – a 15-year service experience. Compos A: Appl Sci Manuf 40(9):1381–1398. ISSN 1359-835X, https://doi.org/10.1016/j.compositesa.2008.11.006. (http://www.sciencedirect.com/science/ article/pii/S1359835X08002984) ISO 24817: 2015 – Petroleum, petrochemical and natural gas industries – composite repairs for pipework – qualification and design, installation, testing and inspection (2015) Mableson AR, Dunn KR, Dodds N, Gibson AG (2000) Refurbishment of steel tubular pipes using composite materials. Plast Rubber Compos 29:10,558–10,565. https://doi.org/10.1179/ 146580 100101540770

52

Oil Industry

1555

Mally TS, Johnston AL, Chann M, Walker RH, Keller MW (2013) Performance of a carbon-fiber/ epoxy composite for the underwater repair of pressure equipment. Compos Struct 100:542–547. ISSN 0263-8223, https://doi.org/10.1016/j.compstruct.2012.12.015 McGeorge D, Echtermeyer AT, Leong KH, Melve B, Robinson M, Fischer KP (2009) Repair of floating offshore units using bonded fibre composite materials. Compos A: Appl Sci Manuf 40 (9):1364–1380. ISSN 1359-835X, https://doi.org/10.1016/j.compositesa.2009.01.015. (http:// www.sciencedirect.com/science/article/pii/S1359835X09000153) Meniconi LM, Freire JF, Vieira RD, Diniz JC (2002) Stress analysis of pipelines with composite repairs. ASME. International pipeline conference, 4th international pipeline conference, pp 2031–2037. https://doi.org/10.1115/IPC2002-27372 Meniconi LCM, Porciuncula IN, McGeorge D, Pedersen A (2010) Structural repair at a production platform by means of a composite material patch. Offshore Technol Conf. https://doi.org/10.4043/ 20657-MS Meniconi et al (2014) Experimental fatigue and aging evaluation of the composite patch repair of a metallic ship hull. Appl Adhes Sci 2:27. 10.1186/s40563-014-0027-8 Rohem NRF, Pacheco LJ, Budhe S, Banea MD, Sampaio EM, de Barros S (2016) Development and qualification of a new polymeric matrix laminated composite for pipe repair. Compos Struct 152 (15):737–745. ISSN 0263-8223, https://doi.org/10.1016/j.compstruct.2016.05.091. (http:// www.sciencedirect.com/science/article/pii/S0263822316307140) Turton TJ, Dalzel-Job J, Livingstone F (2005) Oil platforms, destroyers and frigates – case studies of QinetiQ’s marine composite patch repairs. Compos A: Appl Sci Manuf 36(8):1066–1072. ISSN 1359-835X, https://doi.org/10.1016/j.compositesa.2004.11.006. (http://www.sciencedirect.com/sci ence/article/pii/S1359835X04002702)

Part X Emerging Areas

Molecular Dynamics Simulation and Molecular Orbital Method

53

Ya-Pu Zhao, Feng-Chao Wang, and Mei Chi

Contents 53.1 53.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Molecular Dynamics Simulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.1 Basic Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.2 Force Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.3 Boundary Conditions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.4 Energy Minimization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.5 Integrators . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.6 Analysis of the MD Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.7 Water Models Used in MD Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.2.8 Examples of MD Implements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.3 Molecular Orbital Method . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.3.2 Density Functional Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.3.3 Examples of DFT Calculations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.4 Hybrid Quantum Mechanics/Molecular Mechanics (QM/MM) Methods . . . . . . . . . . . . . 53.4.1 Basic Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.4.2 QM/MM Boundary Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.4.3 Examples of QM/MM Simulations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.5 Ab Initio Molecular Dynamics (AIMD) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53.6 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1560 1561 1562 1562 1564 1566 1567 1568 1570 1571 1576 1576 1578 1582 1585 1586 1588 1588 1590 1592 1593

Abstract

Computer simulations have provided a powerful technique in understanding the fundamental physics and mechanics of adhesion. In this chapter, various Y.-P. Zhao (*) · F.-C. Wang · M. Chi State Key Laboratory of Nonlinear Mechanics (LNM), Institute of Mechanics, Chinese Academy of Sciences, Beijing, China e-mail: [email protected]; [email protected]; [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_52

1559

1560

Y.-P. Zhao et al.

simulation methods pertaining to adhesion technology are introduced, such as the molecular dynamics simulations, the quantum mechanics calculations, the molecular orbital method, the density functional theory, and the molecular mechanics simulations. Besides, some combined methods such as the hybrid quantum mechanics/molecular mechanics simulations, ab initio molecular dynamics, and the density-functional-based tight-binding method are reviewed. General features and routines of these methods as well as the basic theory are described. The advantages and disadvantages of these methods are compared and discussed. Each method has the distinctive advantage and is suitable for specific condition. Some examples are proposed to give the direct perception when investigating adhesion issues using various simulation methods. All these instances are expected to be helpful to readers when performing the corresponding simulations and analyzing of the results.

53.1

Introduction

As hierarchical, multiscale, and complex properties in nature, as shown in Fig. 1, adhesion is an interdisciplinary subject, which undergoes vast experimental, numerical, and theoretical investigations from microscopic to macroscopic levels. As an example, adhesion in micro- and nano-electromechanical systems (MEMS/NEMS) is one of the outstanding issues in this field including the micromechanical process Adhesion force Solid 1

Apparent contact

Solid 2 Capillary

Solid 1

Micro roughness Solid 2

Nano roughness True contact

Fig. 1 Hierarchical and multiscale nature of adhesion

Van der waals force

53

Molecular Dynamics Simulation and Molecular Orbital Method

1561

of making and breaking of adhesion contact; the coupling of physical interactions; the trans-scale (nano–micro–macro) mechanisms of adhesion contact, adhesion hysteresis; and the new effective ways of adhesion control (Zhao et al. 2003). Computer simulations including quantum mechanics (QM) calculations and molecular dynamics (MD) simulations have provided a powerful technique in understanding the fundamental physics of micro- and nanoscale adhesion (Shi et al. 2005; Yin et al. 2005; Chi and Zhao 2009; Yuan et al. 2009). QM method has been proven accurate and provided fundamental understanding for a wide variety of phenomena, including the energy levels of atoms, the band structure of lattice, and the chemical reaction of the materials. The key rule of the QM is to establish the Hamiltonians of the system and to solve the many-body Schrödinger equation (Leach 2001). QM method has the advantage over MD simulations that it includes the effects of charge transfer, polarization, and bond breaking and forming from the beginning. However, the QM computation is always expensive. So QM simulation is employed for the investigation of the chemical reaction in which a few number of atoms are needed, while MD simulation is well suited for the investigation of the dynamics process and time-dependent properties for relatively large systems (Allen and Tildesley 1989). Besides, the molecular mechanics (MM) simulation method is used to study the conformational transformation of the system. Due to some drawbacks of classical MD, ab initio MD (AIMD) is proposed to describe the molecular system dynamic behavior directly from the electronic structure. The development of the hybrid quantum mechanics/molecular mechanics (QM/MM) algorithm (Field et al. 1990) is guided by the general idea that an electronically important region of the system is treated with QM and the remainder admits a classical description, which combines both merits of QM and MM. The QM/MM method is a powerful and effective tool to investigate the adhesion issues such as the polymer interaction with metals. The density-functional-based tightbinding (DFTB) method is also one of the combined simulation methods which integrates the density functional theory (DFT) and the tight binding (TB) method. It should be noted that any one of the aforementioned methods has both advantages and disadvantages. One should pay extra attention when choosing a method before performing simulations.

53.2

Molecular Dynamics Simulation

Molecular dynamics (MD) is a computer simulation technique with which one can obtain the time evolution of a system of interacting particles (atoms, molecules, coarse-grained particles, granular materials, etc.). It was first introduced to study the interactions of hard spheres by Alder and Wainwright in the late 1950s (Alder and Wainwright 1959). With the rapid developments of the computer technology, MD simulations have become a powerful technique and are widely used in the study of proteins and biomolecules, as well as in materials science. This method is now routinely used to investigate the structure, dynamics, and thermodynamics of

1562

Y.-P. Zhao et al.

biological molecules and their complexes addressing a variety of issues including tribology and adhesion (Bharat 2004).

53.2.1 Basic Theory The basic idea of the MD simulations is to generate the atomic trajectories of a system of N particles by numerical integration, of a set of Newton’s equations of motion for all particles in the system !

mi

! @ 2 ri ¼ Fi 2 @t

(1)

where Fi is the force exerted on particle i, mi is the mass of particle i, and ri is its current position. The force can be derived from the gradient of the potential energy !

Fi ¼ 


 ! ! ! @U r 1 , r 2 , . . . , rN !

@ rt

(2)


 ! ! ! in which U r 1 , r 2 , . . . , r N is the empirical interaction potential, depending on the position of the N particles. Given certain boundary conditions, as well as an initial set of positions and velocities, which are referenced as the initial condition, the equations of motions for a specific interatomic potential can be solved and then the information including atomic positions and velocities generated. One may have an interest in the conversion of this microscopic information to macroscopic observables such as pressure, energy, temperature, etc., which requires statistical mechanics. In contrast with the Monte Carlo method, MD is a deterministic technique: The subsequent time evolution is in principle completely determined. Two basic assumptions of the MD simulations are: (i) The motions of all the particles obey the classical Newton’s laws of motion, and (ii) the interactions between any pair of the particles satisfy the superposition theorem. To summarize, the highly simplified description of the MD simulation algorithm is illustrated in Fig. 2.

53.2.2 Force Fields In a MD simulation, a force field which includes the functional form and parameter sets needs to be specified in order to describe the interaction between atoms and molecules, which is also called the potential energy of a system of particles. Force field functions and parameter sets are derived from both experimental work and high-level quantum mechanical calculations. Most force fields are empirical and consist of a summation of bonded forces associated with chemical bonds, bond angles, bond dihedrals, out-of-plane bending, and nonbonded forces associated with

53

Molecular Dynamics Simulation and Molecular Orbital Method

1563

Fig. 2 Simplified flowchart of a standard MD simulation

van der Waals forces and the Coulomb forces. The most commonly used functional forms are U ¼ U nobond þ U bond þ U angle þ U dihedral þ U improper þ U elec

(3)

Van der Waals interaction can be expressed as the best well-known Lennard–Jones (LJ) 12–6 function (Jones 1924), U nobond ¼ 4e


  σ 12
σ 6  r r

(4)

where e is the depth of the potential well, σ is the distance at which the interatom potential is zero, and r is the distance between atoms. In a MD simulation, the truncation schemes are always adopted. Here, we introduce the simplest one in which the potential is discontinuous at r = rc  U nobond ðrÞ ¼

U nobond ðrÞ  U nobond ðrc Þ, 0

r1000 kDa), in sum an elastic storage protein with exceptional adhesive characteristics. It is the main component beyond water in seitan (meat substitute). Gluten is chiefly used in the baking industry but also in the paper industry, and furthermore as an additive in concrete and mortar, and in adhesives. A protein that obtained significant interest is casein, maintained from milk by lowering the pH; it was already used as paint in caves to bind pigments. It is also known as thermoset, namely, galalith, when cured with formaldehyde. The water resistance of casein paints can be achieved by adding chalk milk to it. Because of the food competition and its high price, casein was replaced widely. Still, it is used even today in the food sector for printing casein inks on sausage casings or for labeling bottles. Indoor paintings and paints in kindergartens are also often casein based (Türk 2014). Casein fibers have most comfortable, excellent water transportation, air permeability, and long-lasting antibacterial effects and have received valid international certifications for ecological textiles. These characteristics are attractive for the fashion sector, but in medical applications they were considered cytotoxic. As a plethora of interesting, proteinaceous structures exists and corresponding unique properties have been identified, the areas of applications as adhesives have rapidly become extremely complex, as the following review focused on just the three major classics beyond casein shows. Treatment of the marine adhesive proteins is postponed and covered in the section polyphenol-based and bioinspired adhesives.

54

Bioadhesives

1607

Fibrous Proteins (Collagen, Animal Glues, and Fish Glue) Collagen, keratin (wool), silk, and elastin are common fibrous proteins. As structuring molecules some of them are constructed hierarchically just like cellulose. Without pretreatment they can be used as reinforcing agents. Collagen is a by-product from leather manufacturing and is used in foods, pharmaceutical products, and medical devices (wound dressing and tissue engineering). In particular medical applications promise a substantial added value. Animal glues are obtained by boiling bones (bone glue), skin (hide glue) or swim bladders, and skin and bones from fish (isinglass or fish glue). Cartilage glue, on the other hand, is not based on the same compound as other animal glues, but is made up of chondrin, a protein-carbohydrate complex, with less adhesive power (chondroitin sulfate glue). Boiling of the gelatine leads to a water-soluble substance that consists mainly of partially unfolded collagen, whose composition is similar to that of gelatine. All animal glues are high-polymer proteins and derivatives of collagen that form colloids (NPCS Board of Consultants and Engineers 2017). Degradation, e.g., by hydrolysis and heating of collagen (mainly type I), results in gelatine, which was already used by the ancient Egyptians as an adhesive. Gelatine is able to swell and to absorb water due to a high content of polar residues (about 65%). It is amphiphilic, which makes it suitable as an additive to stabilize foams (Türk 2014). The sum of excellent properties inaugurates versatile niche applications as, for example, in stucco work as a gypsum additive (delaying hardening and increasing strength), in gas masks as an antifogging additive, in compostable candles on graves, and in gum blends as a smoothener and as barrier adhesive. Currently the price for gelatine is relatively low compared to that of synthetic or plant proteins. Proteins in Tissue Engineering Substrates involved in surgeries are challenging substrates that require advanced adhesives and need to obey certain natural concepts: fibrinogen and fibrin are proteins present in blood plasma (see chapters in von Byern and Grunwald 2010) and so is albumin, which is inter alia located in the blood. Both and furthermore collagen possess remarkable properties and excellent biocompatibility. Therefore they are used as medical adhesives with FDA approval (albumin limited to surgical repair of acute thoracic aortic dissections). Albumins are globular proteins with amphiphilic character and a high content of the sulfur containing amino acid cysteine. They have polar domains exposed to the outer site, while the lipophilic ones are directed to the inner sphere. This structural feature enables them to function as transporters of water-insoluble compounds. Albumin-based adhesives are usually cross-linked with glutaraldehyde or similar organic compounds and reach an initial tack in about 30 s. Their bonding strength becomes maximal within a few minutes and ranges between 10 and 40 kPa (Bochynska et al. 2016). Fibrin and fibrinogen belong to the blood clotting system. In combination with thrombin, calcium ions, and factor XIII, they are the most commonly used bio-based tissue adhesives, although their bonding strength (approx. 0.01 MPa) is one order of magnitude lower than gelatineresorcinol-formalin adhesives (approx. 0.1 MPa) (see chapters in von Byern and Grunwald 2010). Collagen and gelatine, a water-soluble derived form of collagen,

1608

K. Richter et al.

became integral parts of the surgeon’s toolbox as well: its applications are established in tissue reconstruction, e.g., after drastic damage from combustion, owing to the high potential as a surgical sealant. The substrate can be enzymatically cross-linked by transglutaminase and gelates in less than 5 min with bonding strengths about 15–45 kPa. Gelatine is considered biodegradable, non-immunogenic, and biocompatible. It usually needs a cross-linking agent as well as a polymeric additive (e.g., alginate based), with additional functional groups available for cross-links to achieve sufficient mechanical strengths. But it can also be designed as a two-component adhesive such as photochemically induced crosslinking of oxidized urethane dextran and gelatine (Mathias et al. 2016).

Vegetable Proteins Vegetable proteins like those of soybeans, maize, peas, or wheat gluten, as described above, are in principle all suitable for adhesion applications. As raw materials of by-products, they bear potential for sustainable substitutes, e.g., in wood adhesives. In contrast to casein, they benefit from mild hydrolysis (enzymatical or chemical) or physical treatment to alter their molecular structure by unfolding into a suitable conformation for adhesional processing. Due to its comparatively low costs and good adhesion characteristics, soy protein has established itself as a binder for fillers and pigments in (paper) coatings. Thus, it gives paper a glossy white surface. It was already used in the USA until the beginning of the last century as a low-budget (0.02 dollar/pound; Weakley and Mehltretter 1965) adhesive composed of soy protein, a nonvolatile starch dialdehyde, and lyophilized blood. Just cross-linked soy proteins are not suitable as pure material; they need to be blended with plasticizers because of their brittleness. Also the water content has a considerable impact on the material properties. Up to now just 0.5% of soy protein is used in industrial applications (Türk 2014). Nevertheless, strength and water resistance of vegetable proteins are still far removed from commercial adhesives. Their potential, however, can be increased by blending them with poly(vinyl acetate) or poly(vinyl alcohol) but also with other proteins as described for the American low-budget formulation with blood and casein (Mathias et al. 2016).

54.2.6 Polyphenol-Based Adhesives Polyphenols are natural compounds found in animals and plants. These bulky molecules comprise an aromatic p-system and hydroxyl groups in phenolic groups like catechols. It is assumed that the catechol functionality is responsible for adhesive and cross-linking characteristics. In plywood industry the most common adhesives are phenol based and cross-linked by formaldehyde (phenolformaldehyde). It stands out with excellent weather and water resistance. Based on the structural similarity, several attempts have been made to replace phenol by lignin derivatives. Lignin is composed out of p-coumaryl alcohols, coniferyl, and sinapyl. These three phenylpropanoid monomers give an amorphous polyphenol linked via a multitude of interunit bonds. Lignin is linked covalently to hemicellulose. It serves

54

Bioadhesives

1609

as a waterproof and antimicrobial binder for cellulose fibers in wood and is a highly valuable by-product of the lignocellulosic bio-refinery and pulping industry. The latter is less suitable as the lignin coming from the bio-refinery for the use as adhesion. Lignin coming from the bio-refinery has fewer methyl groups and that enhances reactivity. Therefore the lignocellulosic residues have a high content of reactive lignin with access to condensation reactions with phenol and formaldehyde in alkaline conditions. Lignin can replace phenol up to 50% in the formulation (plywood based of such a formulation reached the outdoor grade pursuant to the Chinese National Standard) (Mathias et al. 2016). Homogeneous and constant batches of lignin can also be obtained directly starting from different biomass sources by environmentally friendly enzymatic processes. These enzymes involving laccases, tyrosinases, and lipoxygenases transform lignin into a reactive biopolymer that can be used, e.g., as prepolymers, in adhesives or coatings. A life cycle assessment explored the environmental sustainability of the production. Laccasemodified and reduced lignin-soy protein adhesives yielded more than half of the strength of a common polyurethane adhesive (Mathias et al. 2016). Adhesives from marine sessile organisms are well designed for hydrated surroundings. Several mechanisms are responsible for this performance, depending on the purpose of shorter- or longer-lasting adhesions. In many adhesives key compounds in the composition could be identified such as the amino acid lysine, in case of the sperm cells from goby fish (Mathias et al. 2016), hydroxyproline, or 3,4-dihydroxyphenylalanine (DOPA). The marine bacterium Alteromonas colwelliana uses L-DOPA, tyrosine, and related quinones beyond exopolysaccharides for moisture-resistant adhesion just like some higher organisms like mussels do. DOPA is very versatile in its reactions that are depicted in Fig. 1. It can act as cross-linker or as initial adherent as well. Due to these abilities to adhere in wet conditions and to build up stable networks, these glues gave inspiration for biomedical applications and other technical uses (see Sect. 4.2). DOPA is just one representative phenolic compound. Tannins, lignin, urushi, and cashew nut shell liquid (CNSL) are further examples. The shell of cashew nuts contains about 20% of CNSL, which is composed of an allergenic phenolic lipid, mainly anacardic acid (71–82%) and also cardol (4–20%), cardanol (1–9%), and 2-methylcardol (1–4%). Anacardic acid is a potent skin irritant, similar to the allergenic oil urushiol (toxin of ivy). The extraction process causes decarboxylation of the anacardic acid leading to cardanol (Maiorana et al. 2015). The structural increments are related to the properties of the resins, made up of phenolic compounds: (i) hydroxyl groups are responsible for good adhesion and proper reactivity at moderate temperatures; (ii) a long aliphatic side chain gives resistance to water, good anticorrosive properties, low viscosity, long open/application time, and flexibility; and (iii) the aromatic ring ensures excellent chemical resistance. The oldest adhesive used by mankind, birch tar, is also considered to be a polyphenolic representative. Another well-known tar is bitumen, whose use was documented at the site of Umm el Tlell and Hummal (Syria). It dates back to about 70,000 years to the Middle Paleolithic (Boëda et al. 2008). Even though it cannot be considered as actually renewable, it is a natural petroleum tar, consisting of

Fig. 1 Catechole can act as chelating ligand and complexate metall ions (a), the hydroxy group is capable to interact by hydrogenbridging (b), the phenol can use its π-system for interactions with p-Orbitals und π-Orbitals (c). Quinone can form covalent bonds by reactions with thiol (d); form Schiff Base and undergo a

1610 K. Richter et al.

54

Bioadhesives

1611

macromolecular chains that become solid by entanglement without any crosslinking, a natural hot melt. Resins (see Sects. 2.1 and 2.3.1) and tars have widespread uses as adhesives. Bitumen is available in liquid form, as water emulsion or in solvents, and also in solid form like hot melts. Additives like proteins or thixotroping agents are used in bitumen formulations to adjust their viscosity, tack, and mechanical strength. The tack can be improved by an addition of ionomeric elastomers and flexibility benefits from rubber latex in the formulation. Cigarette butts encapsulated with bitumen are used in construction and combine recycling with improving the mechanical properties of asphalt concrete by reducing thermal conductivity. This reduces the feared Urban Heat Island effect and also hot spots on roads caused by sunlight (Mohajerani et al. 2017).

54.3

Biological Adhesives

Bioadhesives from bacteria, plants, and animals have proven their efficacy for 500 million years and become adapted to suit the needs and requirements of the organisms producing them. However, still very little is known about the composition, production, secretion, and mechanical properties of the vast majority of these systems. Generally speaking, biological adhesion tends to be based on two principles: (i) Attachment via mechanical systems and interfacial forces, i.e., by van der Waals forces or capillary interactions as seen in geckos, flies, beetles, tree frogs, and ivy or by reduced-pressure systems as given in cephalopod suckers (see ▶ Chap. 55, “Biological Fibrillar Adhesives: Functional Principles and Biomimetic Applications”). Among these systems, hairy and smooth toe pads have been studied most extensively, and the structures present on the feet of gecko and tree frog species have become model systems with successful prototypes and applications. (ii) Usage of chemical bonds based on, e.g., proteins or other macromolecules. These secretions are released through specialized glands and serve not only for solitary attachment and locomotion as for mechanical system but also for other purposes such as construction or defense. Around 100 marine and terrestrial organisms are known to secrete adhesives (see Graham 2005; Hennebert et al. 2015; von Byern et al. 2018 and contributions in Smith and Callow 2006; von Byern and Grunwald 2010; Smith 2016), but of these only a few organisms have already been characterized in detail or implemented into functional prototypes (see Tables 1, 2, 3, 4, 5). However, as even closely related ä Fig. 1 (continued) kind of Michael reaction (e) or by radical addition forming a aryl-aryl-coupling (f). The DOPA-quinon can also tautomerise to a melanin precursor (g) Image from Rischka et al. (2010) and republished with permision

Arthropoda

Mollusca

Resistant to UV irradiation, temperature (18–30 ), high humidity range (20–90%), tensile strength 5–32 μN/mm

Graham (2005) and Sahni et al. (2014)

Klöppel et al. (2013)

Worm snail (Vermetus, Dendropoma) Orb-weaver spider (Leucauge, Araneus)

Onychophora

n.d.

von Byern et al. (2017b) and von Byern et al. (2017a)

n.d.

84–90% water, 55% protein (8–1300 kDa, major protein Er_p1 = 250, 350 kDa), major amino acids Gly (27 mol%) and Pro (13 mol%), 1.3% sugar (mannose, galactose), lipids, no toxin Mucopolysaccharides and glycosylated proteins, toxic components Proteins (>65 kDa), major amino acids Gly (22 mol%) and Pro (16 mol%), sugars (N-acetylgalactosamine), lipids, choline, GABamide, isethionate, elements (KNO3, H2PO4-, Na, Cl, Ca)

Velvet worm (Principapillatus, Euperipatoides)

Ericales, Nepenthales, Lamiales Ctenophora

Predation

Sticky mucus net, produced by modified pedal gland system Glue droplets on viscid silk capture thread, produced by aggregate glands at abdominal spinneret

Major references See contribution in von Byern and Grunwald (2010) von Byern et al. (2017b)

Bonding properties Viscoelastic (102 N.s/m2), hygroscopic, bonding lost 120 kDa) (unpubl.), other molecules and elements not characterized yet 60% protein (17–220 kDa), major amino acids Gly (12–30 mol%) and Glx (9–16 mol%) depending on the species, 40% neutral sugar, toxin holothurin Two major proteins (12 and 130 kDa); some species additionally contain hydrogen cyanide (HCN); species-specific pH range Tenuirostritermes resin composed of 62% α-pinene, 27% myrcene, and 11% limonene; some species also contain toxins and pheromones

Table 2 Compilation of organisms using biological adhesives mainly for defense (no claim is made to completeness)

1614 K. Richter et al.

Chordata

Milky secretion from dorsal and ventral skin glands (two-component system)

One dorsal skin gland (granular gland)

Lateral slime glands (two-component system), secreting proteinaceous threads and mucin vesicles

Salamander (Ambystoma, Plethodon)

Burrowing ground frog (Notaden)

Hagfish (Eptatretus, Myxine)

70% water, 78% protein (15–120 kDa), 0.41% sugar (mannose, α-Lfucose, N-acetyl-Dglucosamine), lipids, chemical elements (Na, Cl, K, S); some species secrete distasteful substances with its glue 85–90% water, 55–60% protein (13–500 kDa, dominant Nb-1R), major amino acids Gly (16 mol %) and Pro (9 mol%), 0.75 % sugar 99.9% water, 0.002% thread (proteins) and 0.0015% mucin (acidic sulfated glycoproteins), chemical elements (Na, Cl, K)

von Byern et al. (2017b) and von Byern et al. (2017a)

See contributions in von Byern and Grunwald (2010), Smith (2016), and Graham (2005)

von Byern et al. (2017b)

Hardening within seconds, tensile strength >1.7 MPa (unpubl.), hydrophobic in cured stage

Tensile strength >70 kPa (within 24 h), shear stress >2800 kPA (cured 1 week on wood)

Elastic and coherent soft hydrogel, high extensibility. No adhesiveness

54 Bioadhesives 1615

Bioadhesive main function Construction

Sticky silk threads secreted from the abdominal spinneret to line and stabilize the burrow in the desert sand Salivary gland secretions (silk and glue coat) as binder for organic/inorganic material constructions Silk as double fiber with 5 μm diameter, core (0.5 μm thick

Huntsman spider (Cebrennus, Leucorchestris)

Caddisfly larvae (Hesperophylax)

Wax glands (three cell types) located ventrally on the 4–7 abdominal segment

Glue production Salivary secretion used as binder for organic material to build papery nest. The number of glands varies between species

Honeybee (Apis)

Phylum (zool.)/ order (bot.) Genus Arthropoda Eusocial wasp (Polistes, Vespa)

Silk core: H- and L-fibroin (>350 and 25 kDa); major amino acids Pro, Glu, Cal, Lys; chemical element as Ca Glue coat: neutral and acid glycoproteins, heme peroxidase

n.d.

Chemical composition Species dependent: presumably mostly proteinaceous (>73%, silklike) and insoluble polysaccharides, major amino acids (Ala, Ser, Gly), chemical elements as K, S, Zn, Fe Mixtures of wax esters, hydrocarbons, free fatty acids, and alcohols

Stress 4.5 MPa (pH dependent), elongation up to 120%, recovery bisphasic and Ca2+ dependent

Mechanical manipulation at 35  C, melting point at 62–64  C, fragile at low temperature, not adhesive n.d.

Bonding properties Water resistant, light weight (thickness 43 μm), tensile strength (0.25–1.02 MN/m2)

Table 3 Compilation of organisms using biological adhesives mainly for construction (no claim is made to completeness)

See contribution in Smith (2016)

Foelix et al. (2016) and Henschel (2017)

Tulloch (1970) and Hepburn et al. (1991)

Major references McGovern et al. (1988), Singer et al. (1992), and Cole et al. (2001)

1616 K. Richter et al.

Swiftlet (Collocalia)

Chordata

Stickleback fish (Gasterosteus)

African lungfish (Protopterus)

Sandcastle worm (Phragmatopoma) Tube worm (Sabella)

Annelida

Salivary secretions used as binder or solitary for bird nest building Three types of goblet cells, which differ morphologically and histochemically Skin secretions form a protective outer cocoon against dry periods in the soil Nest construction with organic material. Glue produced in the kidney, stored in the urinary bladder

Four distinct secretory cell types in the building organ release the cement to stick sand grains toward a tube

Protein (Spiggin, 203 kDa), major amino acid Cys (8 mol%)

Positive staining for sugars (PAS)

Two glycoproteins, sialic acid-rich O-glycosylproteins

Proteins rich in DOPA, major amino acids Ser (29 mol%), Gly (26 mol%), Ala (10 mol%)

Highly elastic thread, silklike appearance

n.d.

Tensile strength >2,4 MPa

Van Iersel (1953) and Jakobsson et al. (1999)

Kitzan and Sweeny (1968) and Greenwood (1986)

See contributions in Graham (2005), Smith (2016), Graham (2005), and von Byern et al. (2017a) Oda et al. (1998)

54 Bioadhesives 1617

Bioadhesive main function Attachment

Yeast (Candida)

Diatom (Toxarium)

Ochrophyta

Genus Pseudomonas, Caulobacter

Fungi

Phylum (zool)/ order (bot)/ kingdom Bacteria

Glue secreted through slit in the silica cell wall (termed raphe)

Glue production Extracellular matrix, synthesized by three different mechanisms: ABC transporter, Wzx/Wzy dependent, synthase dependent. Attachment via fimbria/pilus Depending on function (substrate or spore bonding), i.e., appressorium, hyphodium, conidia, epithelial cells Species-specific macromolecules (Eap1, 110 kDa protein, 90 kDa mannoprotein, etc.), mucilaginous matrix (containing mycosporinealanine, cutinase, esterase, etc.), sugars (i.e., mannose, galactose) Protein (single >220 kDa), major amino acids (Gly 22 mol%, Asx 14 mol%, His 11 mol%), sulfate, chemical elements as Ca, Mg, sugars (mannose, xylose, etc.)

Chemical composition Adhesin (i.e., FimH) and exopolysaccharides as mannose, N-acetylgalactosamine, rhamnose, etc.

0.8 nN, selfhealing properties

See contribution in Smith (2016)

Appressorium See contribution turgor pressure up in Smith (2016) to 8 MPa and force of 17 μN

Bonding properties Major references Binding partly See contribution effected by acidic in Smith (2016) pH (4–5), adhesion μN range

Table 4 Compilation of organisms using biological adhesives mainly for attachment (no claim is made to completeness)

1618 K. Richter et al.

Orchid (Catasetum)

Arrowhead vine (Syngonium) English ivy (Hedera)

Sea lettuce (Ulva)

Giant kelp (Macrocystis, Durvillaea)

Asparagales

Alismatales

Ulvales

Laminariales

Apiales

Flowering plant (Tilia) Mistletoe (Viscum)

Malvales Santalales

Vesicles (physodes) released from fertilized zygote for permanent attachment

Vesicles with adhesive content in free-swimming zoospore

Pollen: 1. Pollenkitt produced by anther tapetum 2. Viscid thread is part of the ektexine Viscid disc, attachment of pollinarium to pollinator. Temporary attachment Aerial root hair, permanent attachment to all types of substrata

Phlorotannin with phloroglucinol monomer. >150 compounds with MW from 126 Da to 650 kDa, cross-linked Ca alginates

Protein (110 kDa), glycan moieties

Arabinogalactan proteins, pectic polysaccharides, chemical element (Ca)

Proteins and polysaccharides

Pollenkit: lipids, carotenoids, protein Viscid thread: polymer bases on sporopollenin Viscin: polysaccharide only Glycoprotein including sucrose, glucose, fructose

Combination of physical contact, chemical secretion, and shape modification. Tensile strength, 3.4 MPa (single root hair) Vesicle discharge and curing within 1 min, Ca2+ involvement Tissue adaptation, thallus breakage, shear stress >0.3 MPa

n.d.

n.d.

Viscid threads are flexible but not elastic threads

(continued)

See contributions in Stevens et al. (2002), Smith and Callow (2006), and Graham (2005)

See contribution in Smith and Callow (2006)

Melzer et al. (2010) and Huang et al. (2016)

Yang and Deng (2017)

Schlee and Ebel (1983)

See contribution in von Byern and Grunwald (2010)

54 Bioadhesives 1619

Bioadhesive main function

Mussel (Mytilus)

Mollusca

Violet sea snail (Janthina)

Freshwater polyp (Hydra) Beadlet anemone (Actinia)

Genus Demosponge (Lubomirskia)

Cnidaria

Phylum (zool)/ order (bot)/ kingdom Porifera

Table 4 (continued)

Byssus threads with distal adhesive plaque, originated from glands in the foot organ. Permanent bonding but mobility given by thread breakage and re-anchorage Anterior part of the foot (propodium), mucus used as float, >12 cm long, >2 cm width

Glue production Baso-pinacocyte secreting spongin for permanent attachment. Chitin glands not determined yet Gland cells at the basal disc part of the peduncle for temporary bonding

n.d.

Hydra: proteins (presumably glycosylated), protein characterization in progress Actinia: 96% water, 24% protein (12-200 kDa), 8% carbohydrate, 1% lipid, 67% chemical elements (Cl, Na, Mg), equinatoxins 6 foot proteins (115, 42–47, 5–7, 79, 9.5, 11.6 kDa) with different amounts of DOPA, collagenous core

Chemical composition (a) Spongin (collagen-like protein) (b) Chitin (close to α-chitin)

Mussel plaque tensile strength >0.85 MPa (with seasonal variation), plaque area varies to surface energy Quick curing. Enclosure of gas to a foamlike buoy. Clear glue color

Muscle-mediated detachment

Laursen (1953)

See contributions in von Byern and Grunwald (2010), Smith (2016), and Graham (2005)

Stabili et al. (2015) and Rodrigues et al. (2016)

Bonding properties Major references n.d. Evans (1977) and Ehrlich et al. (2013)

1620 K. Richter et al.

Salivary gland. Used as temporary attachment in the skin as a dowel

Unicellular gland type in the body secrete cement for permanent bonding

Ticks (Dermacentor, Amblyomma)

Acorn barnacle (Balanus, Megabalanus)

Arthropoda

Ectoparasite (Entobdella)

Adhesive organ at the tail. Duo-gland system (one gland for attachment, other for release). Temporary attachment Anterior adhesive pad with two gland types, with rodor spheroid-like content

Adhesive glands on digital tentacle (Nautilus), dorsal mantle (Idiosepius), other body regions. Temporary attachment

Platyhelminthes Flatworm (Macrostomum)

Cephalopoda (Nautilus, Idiosepius)

Protein, lack of sugars, lipids, and L-DOPA, similarities in amino acid composition to sea star, limpets, and barnacles Two proteins (RIM36, 64P), major amino acids Gly, Pro, Tyr. Proteins similarity to keratin, collagen, glutenins. Protein 64 P homology to Phragmatopoma (see Sect. 3.3) 84–90% protein (majors from 16- 110 kDa), 1% sugar, major amino acids Ser (9–11 mol%), Leu 8–9 mol%), Pro (8 mol%)

Proteins (characterization in progress) and carbohydrates (PNA lectin); transcriptome

Histochemical confirmation of acidic (1.0–2.5) and basic proteins (pH  8.0) and sugars

Tensile strength >2 MPa (cyprid larva and adult)

Cement-like glue, resistant to shear force from water current, temporary bonding Glue with quickhardening core and slow-hardening cortex, high bonding strength

Weak bonding, fast release, for some species a mechanical detachment is assumed n.d.

(continued)

See contributions in Smith (2016), Graham (2005), and von Byern et al. (2017a)

Kemp et al. (1982), Graham (2005), and Hennebert et al. (2015) Johannes

Kearn and EvansGowing (1998) and Hamwood et al. (2002)

Lengerer et al. (2016) and Lengerer et al. (2017)

See contribution in von Byern and Grunwald (2010)

54 Bioadhesives 1621

Bioadhesive main function

Chaetognatha

Phylum (zool)/ order (bot)/ kingdom

Table 4 (continued)

Egg anchorage on substratum or prey: Dinocras: follicle cells with ovaries Drosophila: abdominal gland Liris: tubiform Dufour gland on the abdomen Some species bear adhesive structures (appear as long rigid fingerlike processes) ventrally in the tail area. Temporary attachment

Stoneflies (Dinocras) Fruit fly (Drosophila) Sphecoid wasps (Liris)

Arrow worm (Spadella)

Glue production Unicellular gland type in the stalk. Cement used as attachment and by Dosima also as buoy

Genus Goose barnacle (Lepas, Dosima)

n.d.

Chemical composition Dosima: 92% water, 84% protein (60–85 kDa), major amino acids Ser (6–9 mol%), Gly and Ala (10 mol%), Leu (8–10 mol %), 1.5% sugars Dinocras and Drosophila: protein, polysaccharide Liris: proteins (14–200 kDa, straight-chain hydrocarbons as pentadecane, (Z)-8heptadecane

n.d.

n.d.

Bonding properties Gas volume 19%, hardness 2.5 kPa, tensile strength 0.2 MPa

Michel (1984)

See contribution in von Byern and Grunwald (2010)

Major references Zheden et al. (2015) and von Byern et al. (2017a)

1622 K. Richter et al.

Bioadhesive main function Locomotion

Demosponge (Tethya)

Gastropods (Cornu, Patella)

Mollusca

Genus Diatoms (Crasopedostauros, Pinnularia)

Porifera

Phylum (zool)/ order (bot) Ochrophyta Glue production Glue secreted through slit in the silica cell wall (termed raphe), gliding effected through actin filament Filamentous podia with three distinct gland types, formation of adhesive discs and central fiber Mucus released by numerous glands in the pedal sole (ventral body side) Species-specific: protein (25–50%, mostly pH 1–4, 20–220 kDa); major amino acid Gly (13 mol %); sugar (i.e., mannose, fucose) composition; elements (Cl, K, Ca, S); saturated fatty acids as myristic, palmitic, and stearic acids; water content >93% (in limpets)

Chemical composition Extracellular polymeric substances (EPS) containing mostly sugars (mannose, glucose, galactose). Major amino acids (Gly 18 mol%, Ser 22 mol%, Thr 12 mol%) Mucopolysaccharides (not specified)

Tensile strength >518 kPa

Movement up to 5–8 cm/week. Bonding to hard substrata

Bonding properties Gliding speed >25 μms-1, thickness >150 nm

Table 5 Compilation of organisms using biological adhesives mainly for locomotion (no claim is made to completeness)

Bioadhesives (continued)

See contributions in Smith (2016), Graham (2005), and Graham (2005)

Fishelson (1981)

Major references Poulsen et al. (2014) and see contribution in Smith (2016)

54 1623

Bioadhesive main function

Phylum (zool)/ order (bot) Echinodermata

Table 5 (continued)

Sea urchin (Paracentrotus)

Genus Sea star (Asterias)

Glue production Two types of secretory cells (one adhesive, one de-adhesive) in the tube foot Detachment enzymes ensure the glue breakage for foot release

Chemical composition 21% protein (one major protein Sfp1), major amino acid Gly (10 mol%), 8% sugar 6.4% protein (inter alia nectin and cohesive proteins), 2.5% lipid, 1.2% sugar, large inorganic fraction (45.5%)

Reversible adhesion through release of proteases and glycosylases, glue remains as footprint, tensile strength 340 kPa

Bonding properties Tensile strength >198 kPa

Major references See contributions in Smith (2016) and Graham (2005) Lebesgue et al. (2016) and contributions in Smith (2016) and Graham (2005)

1624 K. Richter et al.

54

Bioadhesives

1625

Fig. 2 Biological adhesives used for different purposes such as for protection (1 barnacles), locomotion (2 starfish), defense (3 sea cucumber), and attachment (4 mussels)

species can differ with regard to gland morphology and/or adhesive composition and the secretions have become optimized for different purposes or environments, there are still a huge number of species whose adhesive secretions have not been investigated yet. Biological adhesives are not exclusively used for settlement, but often also fulfill other purposes in connection with defense, predation, locomotion (Fig. 2), or nest construction and are superbly adapted morphologically, chemically, and physically to the needs and requirements of the organisms that produce them. Different from man-made synthetic systems, biological adhesion works over a wide range of temperatures, in different environments (aquatic, terrestrial, subterranean, arid), and under changing physicochemical conditions. Bonding can occur irrespective of texture or biological interference within milliseconds to all sorts of surfaces, be they natural, synthetic, biological, hard, or soft and with irregular or of complex substratum chemistry. Some organisms form permanent bonds for predation, attachment, or constructions; others use temporary adhesives to enable a holdfast on demand or for locomotion. This diversity of organisms and variety of adhesive systems as well as the low amounts of secretion available make bioadhesion research challenging, but with access to advanced technological approaches, considerable progress in its characterization has already been made.

1626

K. Richter et al.

In particular most animal-based biological adhesives have been found to be typically polymers, formed by proteins and polysaccharides, which provide robust cohesion strength between the molecules and strong interaction with the contact area. In most animal glues, a high protein content is present (up to 90%), and technical progress has been made in some species, reaching full-length sequences of key proteins involved (Hennebert et al. 2015). For many other biological adhesives listed in the tables below, a rough estimation of protein number, molecular mass (i.e., 4–650 kDa), and major amino acid residues (in particular serine, glycine, proline, and/or leucine) are available. Difficulties still appear in view of the carbohydrate characterization, as many adhesives contain a relatively low sugar fraction (3 cm), while spitting spiders and onychophorans (see paragraph in Sect. 3.1) use an oscillating behavior toward the predator or prey. Helicid snails as Cornu not only retract in their shell to protect themselves from predators but also secrete a foamlike slimy secretion through the shell opening. The

1628

K. Richter et al.

defensive glue of the American slug Arion subfuscus is formed by the matrilin-like proteins, heparan sulfate-like polysaccharides, and metal ions and currently used as template for medical sealant (Li et al. 2017). Although the secretion of the hagfishes as Eptatretus is not adhesive (von Byern et al. 2018), the mucus they release from lateral skin glands strongly swells and clogs the gills and mouths of the predators. In general, the most defensive secretions are released during contact with the pesterer and dispersed over its body and in particular mouth only a few fires directly toward the predator. Exposure to air causes in Henia, Tenuirostritermes, and Plethodon an immediate hardening of the glue (von Byern et al. 2018), while the adhesives of Arion, Cornu, and Eptatretus remain viscoelastic for a certain time (von Byern et al. 2018; Smith 2016). The Cuvierian tubules in Holothuria in contrast remain sticky and form a large web, entangling the predator. In many defensive secretions (sea cucumber, salamander, chilopoda, termites), also toxic, distasteful, or noisomely components could be determined, serving to distract the predator additionally (see respective contribution in von Byern and Grunwald 2010; Smith 2016 von Byern et al. 2018 as well as Nutting et al. 1974), while other bioadhesives (e.g., Notaden, Plethodon) confirm its biocompatibility in cell culture and medical tests (see Sect. 4.2).

54.3.3 Construction Biological adhesion and mucus are mainly used as binders for constructions incorporating inorganic/organic materials; a few species also use the secretions for other functions. Annelids as the sandcastle worm Phragmatopoma and the tube worm Sabella build permanent tubes, in which they reside (see respective contributions in Smith 2016). As given for mussels (see below), also these annelids use the amino acid L-DOPA as binder to stick the sand grains together and by this form strong and stable tubes (up to 40 cm) (Smith 2016). As the animals lift themselves from the soil with increasing tube length and thereby lose the possibility to uptake new construction material from the ground, suspended sand grains are collected, selected, and used for tube repair and partial increase. In relation to sabellids, other animals form with their secretions temporary architectures. Stickleback fish as Gasterosteus build in the spring a tunnellike nest of plant origin for breeding. The threadlike glue is secreted from the kidney and dispersed by the fins on the nest material (Van Iersel 1953). Caddisfly larvae and hunting spiders combine silk and adhesives to form stabile cases/burrows; however, differences are given in the production site. The hunting spider Cebrennus releases the sticky silk threads from its abdominal spinneret to line and stabilize the burrow in the desert sand (Foelix et al. 2016). The caddisfly larva instead secretes its silk and glue from the salivary glands and uses all kinds of inorganic and organic material to build portable or substrate-bonded cases; some species even design multilevel cases. Most caddisflies are not as selective as the sabellids (see above) in view of its construction material and case design; some species use all types of material and design their bizarre-looking cases.

54

Bioadhesives

1629

Salivary liquid and organic material is also used by social wasps as Polybia and birds (i.e., Delichon) as nest-cementing substance (Oda et al. 1998). Although the nest constructions may not appear as stable as the caddisfly case, the papery constructions of wasps are quite long-lasting (even after the inhabitant leave) and have very good stability and even water-resistant properties (McGovern et al. 1988). Besides its binder function, adhesives are also solitary used for constructions. Swiftlets build with their concentrated salivary breeding nests (also known as Edible bird’s nests), while the African lungfish Protopterus (Greenwood 1986) form with its mucus protective cocoons to survive dry periods. The waxlike secretion of honeybees provides not only the basis for the honeycombs in nature but is also commonly used in cosmetics, food, and pharmaceutical industry. Besides bee wax, also other secretions as fish slime (Antony 1954, see Sect. 2) and the Phragmatopoma adhesives are known to have a high potential for medical applications, e.g., in the field of fetal membrane defect sites (Mann et al. 2012; Papanna et al. 2015) and spina bifida repair (Papanna et al. 2016) (see Sect. 4.2).

54.3.4 Attachment The usage of adhesives as a support for surface bonding is surely the most common purpose, used by bacteria, plants, and animals. Especially aquatic organisms use such secretions to withstand currents and avoid drifting. Some seal themselves or their eggs permanently to the substratum or biological surfaces (i.e., Ulva, Macrocystis, Lubomirskia, Mytilus, Balanus, Lepas) and developed cement-like binders with high adhesive strength (up to 2 MPa), adapted to the strong hydrodynamic forces typical of the intertidal regions. Others like Hydra, Nautilus, Idiosepius, Macrostomum, Entobdella, and Spadella adhere only temporarily, “on demand.” Detachment is presumably achieved mechanically through muscle contractions or body movements except for the case of Macrostomum, where a second gland type (so-called duo-gland system) secretes a detachment enzyme. Janthina and Dosima release a foamlike glue, enabling them to float on the water and in the water flea Simocephalus (Meyer-Rochow 1979); suction in combination with an adhesive may be employed for temporary attachments. Terrestrial and in particular epiphytic plants (Syngonium, Hedera) use adhesives for attachment; furthermore a large number of flowers use pollenkitt or viscid threads for pollen binding and transport (Catasetum, Tilia, Viscum). Many insects as Dinocras, Drosophila, and Liris in particular use glues for egg anchorage; animals itself in particular use mechanical tools (see ▶ Chap. 55, “Biological Fibrillar Adhesives: Functional Principles and Biomimetic Applications”) for attachment and bonding. Ectoparasites as Entobdella or ticks (Dermacentor or Amblyomma) secure themselves in the host tissue with an adhesive called cement. The cement portion is secreted after insertion of the mouthparts and solidifies almost immediately to tightly anchorage the animal in the tissue during the feeding phase (Kemp et al. 1982). The secretion of the second portion takes up to several days. After

1630

K. Richter et al.

feeding, ticks detach from the cement cones, which remain attached in or on the host’s skin (Kemp et al. 1982; Suppan et al. 2017).

54.3.5 Locomotion Also the here listed animals use glue or mucus to bond temporary to the substratum as given for temporary bonding species (see Sect. 3.4). However, the glue is synthesized and released mostly through structures, specialized for locomotion, i.e., the podia of Tethya, the gastropod pedal sole, or the tube foot of sea stars and sea urchins. Best known are surely the gastropods, which produce a highly effective, hydrogel-like mucus, enabling the animals to cross any smooth, sharp (razor blades), and even extreme superhydrophobic, anti-adhesive surfaces as polytetrafluoroethylene (PTFE, commonly known as Teflon) (Shirtcliffe et al. 2012). The mucus acts as glue, allowing the animals to adhere and push themselves forward mechanically by means of wavelike movements of the sole (Miller 1974). In contrast, Paracentrotus and Asterias likewise use a duo-gland system as Macrostomum (see Sect. 3.4), in which one gland produces the glue and another secretes enzymes to break the bonding and enable movement. In particular the viscous gastropod mucus and its bonding ability on any surface make this secretion interesting not only for new bioinspired wound sealants (Li et al. 2017) but also to design anti-adhesive surfaces for medical implants and superhydrophobic paints (see Sect. 4.1).

54.4

Bioinspired Applications

Despite the diversity and superiority of biological adhesives (biocompatible, biodegradable, lack of heavy metals, or absence of volatile organic compounds), synthetic adhesives in particular (e.g., phenol/resorcinol/urea-formaldehyde) dominate today’s adhesive market and are used as ingredients in many sealants, furniture products, and cosmetics. Although many synthetic adhesives offer necessary features for specific applications such as for wood manufacturing or in microelectronics, as well in dentistry, their performance is far from optimal and they can have unacceptable properties with respect to their suitability as an essential biomedical tool (harmful, toxic, not biodegradable, low adhesive strength, microbial contamination). Technological advances in the area of superior bioinspired materials have been made with scientific progress in biological adhesion, e.g., the ability to produce key elements like L-DOPA recombinants. Such key components enabled the development of tissue adhesive like Cell-TakTM (USA), the first example (year 1986, TM-No. 73604754) of a marine-derived sealant, based on mussel adhesive proteins only; other biomimetic analogues are currently in the research focus. As a result of our increased understanding of the interrelationships between biological adhesives and underlying mechanisms or similar boundary conditions, it is increasingly possible to transform this into

54

Bioadhesives

1631

biomimetic solutions. In contrast to the desired copies in the field of bonding according to biological models, these perfectly adapted organisms cause extensive economic loss due to their adhesive powers. For this reason, the progress made in the course of clarifying adhesive phenomena also serves not least to prevent fouling.

54.4.1 Antifouling Bacterial biofilms are prevalent and possess benefits and damage at once. They cause problems, e.g., by inducing biocorrosion and biofouling in drinking water supplies, and they are liable for impaired safety in surgery and wound healing. But they are also used for waste water treatments (van der Kooij and Van der Wielen 2014). In contrast to nonadherent bacteria, these microorganisms use biofilms as matrices for cooperation and to establish synergistic interactions like preventing the washaway of enzymes and nutrients or in order to survive in hostile environments. Biofilm formation allows them to withstand periods of starvation by quorum sensing. Consequently the vast majority of microorganisms exist in nature in the form of biofilms. They adhere mostly on wet solid surfaces. Within the process of biofilm formation, several steps are distinguished, whereby the first two are interesting in terms of bioadhesion: (1) adsorption to the substrate by cell adhesion (this initial step can last from a few minutes to one hour of exposure and is based on proteins) and (2) formation of extracellular polymeric substances (EPS) composed of polysaccharides to achieve irreversible attachment by polymer bridging (Vandevivere and Kirchman 1993). A well-known example for bacteria attached to teeth and aggregated in a self-made hydrated polymeric matrix is dental plaque. Micro- and macrofouling (Fig. 3) are causes of considerable economic damage. A 5% increase in biofouling has been shown to increase ship fuel consumption by 17%, with a 14% increase in greenhouse gas CO2, NOx, and SO2 emissions. Biofouling of marine energy turbines is regarded as the primary and most persistent source of failure among these devices. Control of biofouling for aquaculture and fishery operations can account for up to 15% of total annual operating costs. Beyond that is biofouling of sensors a significant barrier to technology advancement due to the impact on data quality (Chapman et al. 2014). Consequently, preventive measures are a challenge for material science. To avoid biofouling is inter alia the purpose for the findings on the principles of action of biological adhesives. Different approaches ranging from the use of “confusing surfaces” – to prevent (micro) organisms from adaption – like dendrimer-based coatings to antimicrobial concepts or surface modifications with barettin have proven their potential. Also surface topology mimicking the skin of a shark has been shown to be effective.

54.4.2 Biomimetic Adhesives Today plenty of commercial adhesive agents are biomimetic. The most prominent representative is the gecko tape. The production of such kinds of adhesives is

1632

K. Richter et al.

Fig. 3 Macrofouling caused by tough adherent organism, e.g., mussels and barnacles on different surfaces under harsh conditions (salinity, marine current)

estimated to account for 25 billion m2 (Schwotzer et al. 2012). But for medical issues the market is attractive as well, as the modern surgery is affected by two major factors: cost containment and an aging population (Spotnitz and Burks 2008). Stopping bleedings or sealing and reconnecting tissues after surgical procedures are challenging subjects. The bonding technology emerged as a convenient alternative for wound closure instead of sutures and other tissue applications (Annabi et al. 2015). Recently the slime of slugs (genus Arion) was successfully used as an adhesive and sealant for a porcine heart (Li et al. 2017). Many biomimetic tissue adhesives are based on secretions from marine organism, e.g., mussels and algae, but also from terrestrial organism like snails and frogs. Their performance in aquatic conditions is acceptable but requires polymeric additives to adjust the threedimensional network. These glues received much attention during the last few years and ended up in mimetics of natural structures like DOPA-containing co-polypeptides and DOPAfunctionalized poloxamers or catechol-modified alginate hydrogels (Lee et al. 2013; Bochynska et al. 2016). Red and brown algae also use a mechanism based upon phenolic compounds. Under hydrated conditions their secretions are able to adhere to hydrophilic as well as hydrophobic surfaces. An enzyme (vanadate-peroxidase) induces the cross-link between the polyphenol and the extracellular carbohydrate fibers and causes adhesion to the surface. Hence carbohydrate-based adhesives are worth imitating. Thus, an alginate-polyphenol hybrid adhesive, mimicking Fucus serratus, showed higher adhesion to Teflon™ than to glass (Kumar Patel et al. 2013).

54

Bioadhesives

1633

As promising candidate the Australian frog Notaden was also identified as it exhibited five times stronger adhesive strength than fibrin glue on sheep meniscus tissue (Graham et al. 2005). Various types of medical adhesives are in use as hemostats to form a barrier and stop bleeding (protein and polysaccharide based) or as sealants in order to establish an impervious barrier to gas and many body liquids (fibrin sealant, polyethylene glycol, albumin, and glutaraldehyde) or as adhesives to connect tissue or medical devices (cyanoacrylates, albumin, glutaraldehyde, and fibrin sealant) (Spotnitz and Burks 2008; Ferguson et al. 2010). This classification refers to their function in surgical context. With respect to the former sections, a division could also be used to summarize the medical adhesives according to the compounds’ origins: (i) Natural or biological adhesives (protein based, sprayable foams or dry fibrin sealants, collagen, gelatine; polysaccharide based, chitosan, alginates, dextran, chondroitin sulfate glue) (ii) Synthetic and semisynthetic adhesives (albumin and glutaraldehyde, cyanoacrylates, polyurethanes, polyvinyl alcohol, polyethylene glycols, dendrimers) (iii) Biomimetic adhesives (mussel, sandcastle worm, frog, slug) The development of high-performance adhesives for use with living tissue as shown in Fig. 4 keeps the surgical toolbox expanding (Spotnitz and Burks 2008). More insight is given in ▶ Chap. 58, “Adhesion in Medicine.”

Fig. 4 Investigation and characterization of biological adhesives by MALDI-TOF MS, application of a biomimetic adhesive to soft tissue and testing the joints (Images from Ingo Grunwald, Fraunhofer Institute for Manufacturing Technology and Advanced Materials (IFAM), Bremen, Germany, and reproduced with his permission)

1634

54.5

K. Richter et al.

Challenges and Opportunities

Each of the reviewed bioadhesives has advantages and drawbacks. In terms of bacterial adhesion, approaches for a forensic usage are in discussion. Considering that everyone’s microbiome is highly individual like DNA is, the microbial fingerprint is a valuable hint in criminal investigations. But depending on the kind of test and recognition of the microbiome, this approach can be risky: in case of sequencing the genetic material, the latter can be manipulated by malicious software encoded into the physical nucleic acid strand. The resulting data can corrupt gene-sequencing software and take control of the underlying computer-aided analysis (https://www. wired.com/story/malware-dna-hack/). Adhesion at this minor scale in crime context as well as with regard to antifouling efforts reveals its high potential. So does the use of some algae and their adhesion to surfaces triggered by certain wavelengths of the light. Based on such knowledge, adhesion could be prevented by manipulations involving intelligent materials, for instance, modified materials in bioreactors. Referring to carbohydrates chitosan displays impressive potential as fully biobased adhesive (5 MPa shear strength on pine wood) (Kumar Patel et al. 2013). Microbial polysaccharides blended with surfactants showed trendsetting results as well. However, as with sugars, there is an increasing tendency to abandon starch solutions in choosing binders for clay and pigments. Preferable would be combinations of synthetic latexes with casein or vegetable proteins instead of starch dispersions. Coming to proteins, many investigations in terms of constructional applications have indicated potential: gluten as a substituent (25%) in phenol-formaldehyde resin masses for use in particleboard production (medium-density fiberboard, MDF) did not affect the mechanical performance. As gluten is a by-product from the glucose syrup production, the costs (about 300 euros/t referred to protein desiccant) are more than one sixth less than urea-formaldehyde resins (363 euros/t) (Türk 2014). Assuming that the cost of such binders contributes about 25% to the production costs, a substitution might be more economic, even if more process time is required. Native proteins are mostly insoluble due to their tertiary or quaternary structures. Processing like degradation or reduction can impart the solubility for its use in adhesive formulations. In general proteinaceous adhesives combine biodegradability more favorably and are much more water resistant compared with carbohydrate-based adhesives. Except for chitosan “bio-based” mostly have a positive impact to any life cycle assessment as a substitute to petrochemical compounds, and they show just a few or no emissions. Complete substitution of synthetic components to compete with commercial adhesives still needs to be realized. Intermediate bio-based adhesives are therefore just a compromise. The construction and building sector remains as one of the most significant markets for adhesives. Their uses of adhesives range from flooring to structural purposes or building envelopes. The family of medical adhesives represents another relevant market. Although “classical” medical adhesives and those specific for medical as well as dental devices

54

Bioadhesives

1635

are rather heterogeneous, they can be grouped into one family. The global market for medical adhesives and sealants is estimated to reach over 15 billion US dollars in the year 2022 (http://www.strategyr.com/MarketResearch/Medical_Adhesives_and_ Sealants_Market_Trends.asp). Driven by an aging population, increasing surgical operational interventions, the rise in healthcare needs, and new treatment methods like the preferences for minimal invasive methods or the use of new tissue scaffold applications, the CAGR (compound annual growth rate) in the area of medical adhesives will reach top values of about 10% until 2021, with even higher values expected for the Asia-Pacific area (http://www.strategyr.com/MarketResearch/Medi cal_Adhesives_and_Sealants_Market_Trends.asp). The consciousness of the benefits involved by using bioadhesives in multiple areas (e.g., health, environmental, ethical) is evolving. Bioadhesives are popular; six billion US dollars have been estimated for the global market in 2019. It is possible to increase the sustainability of industrial processes while maintaining or increasing the competitiveness of enterprises. As long as the legislatures are aware of the need to go ahead with the development of bioadhesives and the technologies of secondgeneration biofuels, the development of bioadhesives will lead to products that can cope with commercial adhesives. Last but not least, the requirements for reducing the emissions of volatile organic compounds (VOCs) and the requirement for recyclable materials push the development of bioadhesives (Mathias et al. 2016). The classical medical adhesives are normally grouped into (depending on the classification system) hemostats which stop bleedings, sealants which seal leakages (gas and nonclotting liquids), or adhesives that “glue” two tissues together. This segment of classical medical adhesives alone will have a market size of about three billion US dollars in 2021 (markets) – indicating a highly attractive field of application for medtech companies today and in the future. The market is dominated by the USA and Europe, but emerging markets like those of South America and Asia offer quite significant growth opportunities. The growth rate is observed and checked, for instance, by the process of regulatory affairs (the USA with the FDA, Europe with CE marking) for the certification of new medical adhesives. At least in the EU, the process of the approval of medical devices like adhesives will be tightened due to new regulations. Regardless of this situation, the medical adhesive market and the research associated with it are expected to have a bright future.

54.6

Conclusion

Biological and microbiological adhesives have an enormous potential for industrial applications. The most prominent ones are constructional, packaging, and medical applications. Sustainable packaging will prevail. There is, however, also a need for manipulating and adapting biological adhesion for healthcare or food safety issues. On account of their impressively efficient and precise way to adhere to industrially challenging substrates, these materials need to be studied and the conditions under which initial adhesion occurs want to be understood. Thus, any activity furthering

1636

K. Richter et al.

insights into the principles of these compounds is significant, reflecting, for instance, the current worth of the stem cell industry 3.5 million pounds in the UK alone (Waugh et al. 2016). Developing biocompatible adhesives with strong adhesive properties for medical applications is highly desirable, but the delicate balance between adhesive performance and biocompatibility is crucial and part of the legal and technical requirements. Although “tissue adhesives are far from ideal” (Li et al. 2017), there are many promising approaches under investigation (Balcioglu et al. 2016; Li et al. 2017). Nearly all of the reviewed adhesive classes are important subjects of research to develop alternatives for petroleum-based materials, and the quest exists to develop bioinspired solutions and to achieve tailored adhesives for specific tasks including the prevention of fouling. As carbohydrates and vegetable proteins can be obtained from waste streams of renewable resources and may be produced in great abundance, they do not have to compete with food production and can be considered as a sustainable option. Nature is a prime candidate for inspiration and provides raw materials for applications in adhesives. Bioadhesives can be converted into a variety of useful adhesive products with a promising future, even if natural variations in quality are weaknesses. The interdisciplinary approach with inputs from chemistry, physics, bionics, biology, and engineering offers a wide range of possibilities. Thus, outstanding adhesive performance in nature can be elucidated on a molecular level and translated into technical concepts. Mechanical discrepancies can be bridged with hybrid materials from biological concepts in combination with synthetic components. Looking just at the largest market for adhesives (construction), a fast-growing market (packaging), and the exclusive market (medical adhesives), the future prospects for bioadhesives are positive. This trend is clearly evidenced by the increase in relevant publications covering the field over the last few years (Babu et al. 2013).

References Annabi N et al (2015) Elastic sealants for surgical applications. Eur J Pharm Biopharm 95 (Pt A):27–39 Antony AC (1954) Use of fish slime in structural engineering. J Bombay Nat Hist Soc 50 (3&4):682–683 Babu RP, O’Connor K, Seeram R (2013) Current progress on bio-based polymers and their future trends. Prog Biomater 2:8 Balcioglu S et al (2016) Design of xylose-based semisynthetic polyurethane tissue adhesives with enhanced bioactivity properties. ACS Appl Mater Interfaces 8(7):4456–4466 Baroncini EA et al (2016) Recent advances in bio-based epoxy resins and bio-based epoxy curing agents. J Appl Polym Sci 133(45):44103 Bochynska AI et al (2016) Tissue adhesives for meniscus tear repair: an overview of current advances and prospects for future clinical solutions. J Mater Sci Mater Med 27(5):85_1–85_18 Boëda E et al (2008) New evidence for significant use of bitumen in Middle Palaeolithic technical systems at Umm el Tlel (Syria) around 70,000 BP. Paléorient 34(2):67–83 Buch K et al (2009) Investigation of various shellac grades: additional analysis for identity. Drug Dev Ind Pharm 35(6):694–703 Chapman J et al (2014) Bioinspired synthetic macroalgae: examples from nature for antifouling applications. Int Biodeterior Biodegrad 86A:6–13

54

Bioadhesives

1637

Cohen ER et al (2008) Quantities, units and symbols in physical chemistry. IUPAC/RSC Publishing, Cambridge Cole MR, Hansell MH, Seath CJ (2001) A quantitative study of the physical properties of nest paper in three species of Vespine wasps (Hymenoptera, Vespidae). Insect Soc 48:33–39 Eberhard WG (1980) The natural history and behavior of the bolas spider Mastophora dizzydeani sp. n. (Araneidae). Psyche J Entomol 87(3–4):143–169 Ehrlich H et al (2013) First report on chitinous holdfast in sponges (Porifera). Proc R Soc B Biol Sci 280(1762):20130339_1–20130339_9 Erren TC et al (2013) Oceans of plastics: possible risks of cancer in marine wildlife and humans. In: Allodi S, Nazari EM (eds) Exploring themes on aquatic toxicology. Research Signpost, Trivandrum Evans CW (1977) The ultrastructure of larvae from the marine sponge Halichondria moorei Bergquist (Porifera, Demospongiae). Cah Biol Mar 13:427–433 Ferguson J, Nürnberger S, Redl H (2010) Fibrin: the very first biomimetic glue – still a great tool. In: von Byern J, Grunwald I (eds) Biological adhesive systems: from nature to technical and medical application, 1st edn. Springer, Wien/New York Fishelson L (1981) Observation of the moving colonies of the genus Tethya (Desmospongia, Porifera) I. Behaviour and cytology. Zoomorphology 98:89–99 Foelix RF et al (2016) Über den Bau der Wohnröhre bei wüstenlebenden Spinnen. Arachne 21 (1):4–17 Geyer R, Jambeck JR, Law KL (2017) Production, use, and fate of all plastics ever made. Sci Adv 3 (7):e1700782_1–e1700782_5 Gohad NV et al (2014) Synergistic roles for lipids and proteins in the permanent adhesive of barnacle larvae. Nat Commun 5:4414_1–4414_9 Gosline JM et al (1999) The mechanical design of spider silks: from fibroin sequence to mechanical function. J Exp Biol 202(Pt 23):3295–3303 Graham LD (2005) Biological adhesives from nature. In: Bowlin GL, Wnek G (eds) Encyclopedia of biomaterials and biomedical engineering. Taylor & Francis, Oxon Graham LD et al (2005) Characterization of a protein-based adhesive elastomer secreted by the Australian frog Notaden bennetti. Biomacromolecules 6(6):3300–3312 Greenwood PH (1986) The natural history of African lungfishes. J Morphol 190(Suppl. 1):163–179 Gross RA, Kalra B (2002) Biodegradable polymers for the environment. Science 297 (5582):803–807 Hamwood TE et al (2002) Preliminary characterization and extraction of anterior adhesive secretion in monogenean (platyhelminth) parasites. Folia Parasitol 49:39–49 Hennebert E et al (2015) Experimental strategies for the identification and characterization of adhesive proteins in animals: a review. Interface Focus 5(1):20140064 Henschel JR (2017) The biology of Leucorchestris arenicola (Araneae: Heteropodidae), a burrowing spider of the Namib dunes. In: Seely MK (ed) Namib ecology: 25 years of Namib research, Transvaal museum monograph, vol 7. Transvaal Museum, Pretoria Hepburn HR et al (1991) Synthesis and secretion of beeswax in henexbees. Apidologie 22:21–36 Hon DNS (1989) Cellulosic adhesives. In: Hemingway RW, Conner AH (eds) Adhesives from renewable resources, ACS symposium series, vol 385. American Chemical Society, Washington, DC Huang Y et al (2016) Nanospherical arabinogalactan proteins are a key component of the highstrength adhesive secreted by English ivy. Proc Natl Acad Sci U S A 113(23):E3193–E3202 Jakobsson S et al (1999) An 11-ketotestosterone induced kidney-secreted protein: the nest building glue from male three-spined stickleback, Gasterosteus aculeatus. Fish Physiol Biochem 20:79–85 Karak N (2016) Biopolymers for paints and surface coatings. In: Pacheco-Torgal F, Ivanov V, Karak N, Jonkers H (eds) Biopolymers and biotech admixtures for eco-efficient construction materials, Woodhead publishing series in civil and structural engineering, vol 63. Elsevier, Amsterdam Kearn GC, Evans-Gowing R (1998) Attachment and detachment of the anterior adhesive pads of the monogenean (platyhelminth) parasite Entobdella soleae from the skin of the common sole (Solea solea). Int J Parasitol 28(10):1583–1593

1638

K. Richter et al.

Kemp DH, Stone BF, Binnington KC (1982) Tick attachment and feeding: role of the mouthparts, feeding apparatus, salivary gland secretions and the host response. In: Obenchain F, Galun R (eds) Physiology of ticks. Pergamon Press, Oxford Kitzan SM, Sweeny PR (1968) A light and electron microscope study of the structure of Protopterus annectens epidermis. I: mucus production. Can J Zool 46:767–772 Klöppel A et al (2013) Detection of bioactive compounds in the mucus nets of Dendropoma maxima, Sowerby 1825 (Prosobranch Gastropod Vermetidae, Mollusca). J Mar Biol 2013:1–9 Köckritz A, Martin A (2008) Oxidation of unsaturated fatty acid and vegetable oils. Eur J Lipid Sci Technol 110(9):812–824 Koerner L, Gorb SN, Betz O (2012) Functional morphology and adhesive performance of the stickcapture apparatus of the rove beetles Stenus spp. (Coleoptera, Staphylinidae). Zoology 115(2): 117–127 Kumar Patel A, Mathias JD, Michaud P (2013) Polysaccharides as adhesives: a critical review. Rev Adhes Adhes 1(3):312–345 Laursen D (1953) The genus Ianthina. Dana-Report 38:1–40 Lebesgue N et al (2016) Deciphering the molecular mechanisms underlying sea urchin reversible adhesion: a quantitative proteomics approach. J Proteome 138:61–71 Lee C et al (2013) Bioinspired, calcium-free alginate hydrogels with tunable physical and mechanical properties and improved biocompatibility. Biomacromolecules 14(6):2004–2013 Lengerer B et al (2016) Adhesive organ regeneration in Macrostomum lignano. BMC Dev Biol 16(20):1–16 Lengerer B et al (2017) Organ specific gene expression in the regenerating tail of Macrostomum lignano. Dev Biol. https://doi.org/10.1016/j.ydbio.2017.07.021 Li A, Li K (2014) Pressure-sensitive adhesives based on epoxidized soybean oil and dicarboxylic acids. ACS Sustain Chem Eng 2(8):2090–2096 Li J et al (2017) Tough adhesives for diverse wet surfaces. Science 357(6349):378–381 Maiorana A et al (2015) Bio-based epoxy resin toughening with cashew nut shell liquid-derived resin. Green Mater 3(3):80–92 Mann LK et al (2012) Fetal membrane patch and biomimetic adhesive coacervates as a sealant for fetoscopic defects. Acta Biomater 8(6):2160–2165 Mathias JD, Grédiac M, Michaud P (2016) Bio-based adhesives. In: Pacheco-Torgal F, Ivanov V, Karak N, Jonkers H (eds) Biopolymers and biotech admixtures for eco-efficient construction materials, Woodhead publishing series in civil and structural engineering, vol 63. Elsevier, Amsterdam McGovern JN, Jeanne RL, Effland MJ (1988) The nature of wasp nest paper. TAPPI J 71(12): 133–139 Melzer B et al (2010) The attachment strategy of English ivy: a complex mechanism acting on several hierarchical levels. J R Soc Interface 7:1383–1389 Meyer-Rochow VB (1979) The attachment mechanism of the waterflea Simocephalus. Microscopy 33: 551–553 & 558 Meyer-Rochow VB (2007) Glowworms: a review of Arachnocampa spp. and kin. Luminescence 22:251–265 Meyer-Rochow VB et al (2015) Commentary: Plastic ocean and the cancer connection: 7 questions and answers. Environ Res 142:575–578 Michel HB (1984) Chaetognatha of the Caribbean Sea and adjacent areas, NOAA technical report NMFS, vol 15. NOAA/National Marine Fisheries Service, Seattle Miller SL (1974) Adaptive design of locomotion and foot form in prosobranch gastropods. J Exp Mar Biol Ecol 14(2):99–156 Miserez A et al (2008) The transition from stiff to compliant materials in squid beaks. Science 319:1816–1819 Mohajerani A et al (2017) Physico-mechanical properties of asphalt concrete incorporated with encapsulated cigarette butts. Constr Build Mater 153:69–80

54

Bioadhesives

1639

NPCS Board of Consultants & Engineers (2017) The complete book on adhesives, glues & resins technology (with process & formulations). Asia Pacific Business Press Inc, Delhi Nutting WL, Blum MS, Fales HM (1974) Behaviour of the North American termite Tenuirostritermes tenuirostris, with special reference to the soldier frontal gland secretion, its chemical composition and use in defense. Psyche 84(1):167–177 Oda M et al (1998) Study on food components: the structure of N-linked asialo carbohydrate from the edible bird’s nest built by Collocalia fuciphaga. J Agric Food Chem 46:3047–3053 Papanna R et al (2015) Cryopreserved human amniotic membrane and a bioinspired underwater adhesive to seal and promote healing of iatrogenic fetal membrane defect sites. Placenta 36 (8):888–894 Papanna R et al (2016) Cryopreserved human umbilical cord patch for in-utero spina bifida repair. Ultrasound Obstet Gynecol 47(2):168–176 Patachia S, Croitoru C (2016) Biopolymers for wood preservation. In: Pacheco-Torgal F, Ivanov V, Karak N, Jonkers H (eds) Biopolymers and biotech admixtures for eco-efficient construction materials, Woodhead publishing series in civil and structural engineering, vol 63. Elsevier, Amsterdam Penning M (1996) Aqueous shellac solutions for controlled release coatings. In: Karsa DR, Stephenson RA (eds) Chemical aspects of drug delivery systems. The Royal Society of Chemistry, Cambridge Poulsen N et al (2014) Isolation and biochemical characterization of underwater adhesives from diatoms. Biofouling 30(4):513–23 Rischka K et al (2010) Bio-inspired polyphenolic adhesives for medical and technical applications. In: von Byern J, Grunwald I (eds) Biological adhesive systems: from nature to technical and medical application, 1st edn. Springer, Wien/NewYork Rodrigues M et al (2016) Profiling of adhesive-related genes in the freshwater cnidarian Hydra magnipapillata by transcriptomics and proteomics. Biofouling 32(9):1115–1129 Sahni V et al (2014) Prey capture adhesives produced by orb-weaving spiders. In: Asakura T, Miller T (eds) Biotechnology of silk. Springer, Dordrecht Sankaranarayanan Y (1989) Adhesives & Shellac. Shellac Export Promotion Council, Calcutta Schlee D, Ebel F (1983) Note on the chemical nature of the adhesive of viscid discs in Catasetum fimbriatum Lindl. (Orchidaceae). Am J Bot 70(6):872–876 Schwotzer W et al (2012) Biologische und biomimetische Klebstoffe (Teil 2) – Praxistauglichkeit und Trends. Adhäsion 10:38–43 Shibata M (2013) Bio-based epoxy resin/clay nanocomposites. In: Mittal V (ed) Thermoset nanocomposites. Wiley-VCH Verlag, Weinheim Shirtcliffe NJ, McHale G, Newton MI (2012) Wet adhesion and adhesive locomotion of snails on anti-adhesive non-wetting surfaces. PLoS One 7(5), e36983. https://doi.org/10.1371/journal. pone.0036983 Singer TL, Espelie KE, Himmelsbach DS (1992) Ultrastructural and chemical examination of paper and pedicel from laboratory and field nests of the social wasp Polistes metricus say. J Chem Ecol 18(1):77–86 Smith AM (2016) Biological adhesives. Springer, Cham Smith AM, Callow JA (2006) Biological adhesives. Springer, Heidelberg Specht F et al (1998) The application of shellac acidic polymer for enteric coating. Pharm Technol 10:20–28 Spotnitz WD, Burks S (2008) Hemostats, sealants, and adhesives: components of the surgical toolbox. Transfusion 48(7):1502–1516 Stabili L et al (2015) The mucus of Actinia equina (Anthozoa, Cnidaria): an unexplored resource for potential applicative purposes. Mar Drugs 13(8):5276–5296 Stevens CL, Hurd CL, Smith MJ (2002) Field measurement of the dynamics of the bull kelp Durvillaea antarctica (Chamisso) Heriot. J Exp Mar Biol Ecol 269:147–171 Sung HW et al (1998) Feasibility study of a natural crosslinking reagent for biological tissue fixation. J Biomed Mater Res 42(4):560–567

1640

K. Richter et al.

Suppan J et al (2017) Tick attachment cement – reviewing the mysteries of a biological skin plug system. Biol Rev. https://doi.org/10.1111/brv.12384 Tulloch AP (1970) The composition of beeswax and other waxes secreted by insects. Lipids 5(2): 247–258 Türk O (2014) Stoffliche Nutzung nachwachsender Rohstoffe. Grundlagen – Werkstoffe – Anwendungen. Springer, Wiesbaden Türünc O et al (2015) From plant oils to plant foils: straightforward functionalization and crosslinking of natural plant oils with triazolinediones. Eur Polym J 65:286–297 van der Kooij D, Van der Wielen PWJJ (2014) Microbial growth in drinking-water supplies: problems, causes, control and research needs. IWA Publishing, London Van Iersel JJA (1953) An analysis of the parental behaviour of the male three-spined stickleback (Gasterosteus aculeatus L.). E. J. Brill, Leiden Vandevivere P, Kirchman DL (1993) Attachment stimulates exopolysaccharide synthesis by a bacterium. Appl Environ Microbiol 59(10):3280–3286 von Byern J, Grunwald I (2010) Biological adhesive systems: from nature to technical and medical application. Springer, Wien/NewYork von Byern J et al (2017) Chemical characterization of the adhesive secretions of the salamander Plethodon shermani (Caudata, Plethodontidae). Sci Rep 7(1):6647_1–6647_13 von Byern J et al (2018) Examples of bioadhesives for defence and predation. In: Gorb SN, Gorb EV (eds) Functional surfaces in biology III: diversity of the physical phenomena. Biologicallyinspired systems, vol 10. Springer International Publishing AG, p XV, 268 Waugh DG et al (2016) Surface treatments to modulate bioadhesion: a critical review. Rev Adhes Adhes 4(1):69–103 Weakley FB, Mehltretter CL (1965) Low cost protein glue for southern pine plywood. For Prod J 15(1):8–12 Wilks AM et al (2015) Double-network gels and the toughness of terrestrial slug glue. J Exp Biol 218(Pt 19):3128–3137 Wu Y, Li A, Li K (2015) Pressure sensitive adhesives based on oleic acid. J Am Oil Chem Soc 92(1):111–120 Yang X, Deng W (2017) Morphological and structural characterization of the attachment system in aerial roots of Syngonium podophyllum. Planta 245(3):507–521 Yeargan KV (1994) Biology of bolas spiders. Annu Rev Entomol 39:81–99 Zheden V et al (2015) Mechanical properties of the cement of the stalked barnacle Dosima fascicularis (Cirripedia, Crustacea). Interface Focus 5(1):20140049

Biological Fibrillar Adhesives: Functional Principles and Biomimetic Applications

55

Stanislav N. Gorb and Lars Heepe

Contents 55.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.2 Dynamic Adhesion for Locomotion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.3 Pad Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.4 Nature of Attractive Forces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.5 Multiple Contacts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.6 Contact Shape . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.7 Slope and Hierarchy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.8 Adhesion Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.9 Environmental Conditions Affecting Animal Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.10 Biomimetic Implications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55.11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1642 1644 1645 1646 1649 1651 1655 1657 1662 1664 1669 1671

Abstract

Specific mechanisms of adhesion found in nature are discussed in the previous chapter (▶ Chap. 54, “Bioadhesives”). One of the most discussed biological systems in the last decade are the so-called fibrillar adhesives of insects, spiders, and geckos. These systems are adapted for dynamic adhesion of animals during locomotion and, therefore, have some extraordinary properties, such as (1) directionality, (2) preload by shear, (3) quick detachment by peeling, (4) low dependence on the substrate chemistry, (5) reduced ability to contamination and selfcleaning, and (6) the absence or strong reduction of self-adhesion. In the present chapter, we review functional principles of such biological systems in various animal groups with an emphasis on insects and discuss their biomimetic potential.

S. N. Gorb (*) · L. Heepe Department of Functional Morphology and Biomechanics, Zoological Institute at the University of Kiel, Kiel, Germany e-mail: [email protected]; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_54

1641

1642

S. N. Gorb and L. Heepe

The data on ultrastructure and mechanics of materials of adhesive pads, movements during contact formation and breakage, the role of the fluid in the contact between the pad and substrate are presented here. The main goal is to demonstrate how a comparative experimental approach in studies of biological systems aids in the development of novel adhesive materials and systems. The microstructured adhesive systems, inspired by studies of biological systems of insects, spiders, and geckos, are also shortly reviewed.

55.1

Introduction

Representatives of several animal groups, such as insects, arachnids, tree frogs, and lizards, are able to attach to and walk on smooth vertical surfaces and even on the ceiling (reviews by Gorb 2001; Scherge and Gorb 2001; Autumn 2006; Wolff and Gorb 2016). This ability is due to highly specialized attachment devices located at different parts of their legs. Biomimetic transfer of such systems into industrial applications is a very challenging task, and several working groups of materials scientists worldwide are currently attempting to reach this goal by applying various approaches of microfabrication and surface science in general. However, biological systems are far too complex for copying them exactly. That is why intensive structural and experimental comparative studies are required to find out which structural and mechanical features of real biological systems are essential for biomimetics. In order to understand functional principles of animal attachment devices, it is not enough to have information about external morphology of the structures. An integrative approach is necessary to combine data on ultrastructure and mechanical properties of materials, arrangement of muscles, joint design, movements during contact formation and breakage (motor control), sensorics, forces and their directionality, contact mechanics at micro- and nanoscale, potential substrates (surface profile and surface energy), and the presence and properties of the fluid in the contact between the pad and the substrate (chemical composition, viscosity, transporting system). Two important questions have to be answered in order to select the right biological prototype for particular technological requirements. The first question is which product has to be developed, and the second one is which biological system possesses the ability we are going to mimic. Normally, the first and most important requirement from the industry of adhesives is that the new material has to be as sticky as possible. If this is the only requirement, one has to select a prototype among animals and plants using glue-like systems for permanent or long-term adhesion to the substrate (for review, see Gorb 2009 and this book ▶ Chap. 54, “Bioadhesives”). However, if the requirement for the new material is its ability to allow multiple adhesive cycles, for example, in robotic applications for robot locomotion on smooth walls and the ceiling, with the ability (1) to build and break the contact as fast as possible, (2) to attach to unpredictable surfaces, (3) to have lifetime of millions of attachment-detachment cycles, and (4) to not attach to itself, then we definitely have to search for prototypes among locomotory attachment devices of animals.

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1643

To understand principles, at which biological systems operate, detailed studies on ultrastructure, material properties, force range, and motion pattern during locomotion are necessary. Such studies became possible during the last years, because of new developments (1) in microscopical visualization techniques (atomic force microscopy, freezing and environmental scanning electron microscopy), (2) in the characterization of mechanical properties of biological materials and structures in situ and in vivo (measurements of stiffness, hardness, adhesion, friction) at local and global scales, and (3) in computer simulations. In general, there are two different working strategies that aid in extracting structural and mechanical features of biological objects for further development of bioinspired systems. The first strategy deals with a careful detailed characterization of one particular biological system defined as a model object (one of the most prominent examples for fibrillar adhesives is the gecko toe). The second strategy is focused on a comparative study of a large variety of biological objects, which possess similar functional systems that appeared independently in the course of biological evolution. In our opinion, the second approach is more promising for biomimetics, because structural similarities, which evolved independently in different lineages of organisms, may indicate some kind of “optimal” solution for this type of system. Moreover, a comparative approach aids in extracting essential features and abandoning less important ones for designing artificial prototypes. Although different, these two approaches are complementary and help in the course of the biomimetic process. Figure 1 provides a summary and categorization of essential features that contribute to the so-called gecko effect. In this chapter, along with this categorization,

Fig. 1 Summary and categorization of essential features that contribute to the so-called gecko effect: physical mechanisms, attachment pad structure, material and surface properties, and locomotion control. Various substrate properties and environmental factors may influence the adhesive performance of climbing animals (Adapted from Heepe and Gorb 2014. Photo copyright by Alex Mortillaro)

1644

S. N. Gorb and L. Heepe

we report on our 15-year-long experience in studies of locomotory attachment devices in various animal groups with an emphasis on insects. The main goal is to demonstrate how a comparative experimental biological approach can aid in understanding biological attachment systems and derive design rules for the development of novel adhesive materials and systems.

55.2

Dynamic Adhesion for Locomotion

The adhesive contact between the leg and the substrate is mostly important for a rather short period of time of the stance phase and must be released during the swing phase (Fig. 2). These phases repeat infinitely in step cycles during locomotion. Therefore, adhesive structures of animal legs must fulfill several tasks, sometimes opposite to each other and reversibly executed in repeated step cycles (Gladun et al. 2009). (1) The sticky surface of the leg must be ready to immediately build adhesive contact with the substrate. The contact formation must be fast and established with minimal load on the ceiling. (2) Adhesion of few legs must be strong enough to hold the animal on the inclined and even on the overhanging substrate (reliability). (3) At the same time, adhesion must be low enough to allow smooth and quick detachment of the swinging leg. In other words, adhesion must be reversible. (4) The adhesive surface of the pad must be sticky enough to generate sufficient adhesive force to the substrate, but it should be secured from adhering to itself, to prevent disability of the adhesive surface for the next step cycle.

Fig. 2 Ceiling situation. If the mass m of the animal is in the center of gravity, then its weight is W = m g, the angle of the ground reaction is α, the normal pull-off force component is Fn, and the shear force component is Fs. Thus, the weight of the animal walking on the ceiling should be counterbalanced by two forces: friction (Ff), which prevents leg sliding along the substrate, and adhesion (Fa), which prevents leg separation from the substrate. In order to walk on the ceiling, it is not sufficient just to generate strong adhesive bond between the leg and substrate. Two additional problems have to be solved: (1) contact formation must be fast, reliable, and performed with the minimal load on the ceiling and (2) contact must be released in the fast manner and with the minimum of the applied load (From Gladun et al. 2009)

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

55.3

1645

Pad Structure

To date, both the morphological and ultrastructural bases of the ability of animals to dynamically adhere to vertical surfaces have been studied in detail only in representatives of selected taxa, including insects (reviews by Gorb 2001, 2005; Federle 2006), arachnids (Homann 1957; Kesel et al. 2003; Niederegger and Gorb 2006; Wolff and Gorb 2016), tree frogs (Barnes 2006; Smith et al. 2006), and lizards (Hiller 1968; Autumn 2000, 2002; Huber, 1995a, b). In their evolution, animals have developed two distinctly different mechanisms to attach themselves to a variety of substrates: smooth pads and fibrillar (hairy, setose) pads. Due to the flexibility of the material of the attachment pads or fine surface outgrowths, both mechanisms can maximize the possible contact area with a wide range of substrate profiles (Gorb and Beutel 2001; Gorb 2005). Interestingly, in different lineages, both types of attachment pads can be found. For example, within insects, such groups as orthopterans, hymenopterans, and thysanopterans bear smooth pads, whereas dipterans, coleopterans, and dermapterans possess mainly fibrillar pads. Within arachnids, representatives of Solifugae have smooth pads, whereas Araneae are characterized by fibrillar ones. Among reptiles, Gekkonidae and Anolinae have fibrillar pads, and some skinks have smooth ones. Additionally, these highly specialized structures are not restricted to one particular area of the leg. For example, in insects, they may be located on claws, derivatives of the pretarsus, tarsal apex, tarsomeres, or tibia. Recent phylogenetic analyses of hexapods based on the pad characters, processed together with characters of other organ systems, aided in resolving the question of the attachment pad evolution in hexapods and showed that these structures have evolved several times, independently (Gorb and Beutel 2001; Beutel and Gorb 2001, 2006). Fibrillar systems always contain cuticle protuberances on their surfaces (Fig. 3). Protuberances on the fibrillar pads of Coleoptera, Dermaptera, and Diptera belong to different types. Representatives of the first two lineages have socketed setae originating each from two different cells. Setae range in their length from a few micrometers to several millimeters. Dipteran outgrowths are acanthae: single sclerotized protuberances originating from a single cell (Richards and Richards 1979). Structural features of adhesive outgrowths have been previously reported, for example, for flies, the acanthae are hollow inside (Bauchhenss 1979), and some of them contain pores under the terminal plate (Gorb 1998). These pores, presumably, deliver an adhesive secretion directly in the contact area. Pore canals at the base of the shaft may additionally transport secretions to the surface. In beetles, the setal bases are embedded in rubberlike material, which provides flexibility to the supporting material and helps the seta to adapt to various surface profiles. Gecko setae are multibranched, hierarchical structures of 80–100 μm in length (Ruibal and Ernst 1965; Stork 1983; Schleich and Kastle 1986; Roll 1995; Huber et al. 2005a; Rizzo et al. 2006). In anole lizards, setae are smaller and unbranched (ca. 20 μm long) (Ruibal and Ernst 1965; Stork 1983; Rizzo et al. 2006). The gecko setae are not hollow inside. They are formed by proteinaceous fibrils held together by a matrix and surrounded by a proteinaceous sheath. Ordered protein constituent in these structures exhibits a diffraction pattern characteristic of β-keratin. Raman microscopy

1646

S. N. Gorb and L. Heepe

Fig. 3 Wet fibrillar adhesive system of the fly (a–c) and dry fibrillar adhesive system of gecko (d–g). (b, c, e, g) Scanning electron microscopy images of adhesive setae. Both systems are similar according to the contact mechanics but differ in relative contribution of different physical forces to the overall adhesion. Insect mainly employ wetting phenomena, whereas gecko system is dominated by intermolecular van der Waals forces

of individual setae, however, shows the presence of additional protein constituents, some of which may be identified as α-keratins (Rizzo et al. 2006). In anoles, setae emerge from the oberhautchen layer of the epidermal shedding complex into the cells of the clear layer, which act as templates for their growth (Alibardi 1997). The clear layer cells template the formation of spatular terminal elements. This means that setae are fully formed prior to shedding of the external epidermis during the shedding cycle.

55.4

Nature of Attractive Forces

There are two different adhesive mechanisms in fibrillar systems: a wet one, primarily based on capillarity, and a dry one, primarily based on intermolecular van der Waals interactions. Wet adhesive systems (insects) generate a thin fluid layer in the contact between single hairs and substrate, whereas dry ones (spiders, geckos)

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1647

do not. Pad fluids have been previously reported from fibrillar pads of reduviid bugs (Edwards and Tarkanian 1970), flies (Bauchhenss 1979; Walker et al. 1985), coccinellid (Ishii 1987; Kosaki and Yamaoka 1996), and chrysomelid (Eisner and Aneshansley 2000) beetles. The pad secretion contains nonvolatile, lipid-like substances that can be observed in footprints stained with Sudan Black. Microscopy data combined with gas chromatography and mass spectrometry showed that secretion of smooth pads of ants and locusts is an emulsion containing water-soluble and lipid-soluble fractions (Federle et al. 2002; Vötsch et al. 2002). For fibrillar pads, two chemically different secretion compositions were reported. Tarsal secretions of beetles were mainly lipid based (Geiselhardt et al. 2009, 2010; Betz 2010) with a small water fraction, whereas in flies it is assumed to be a mainly water-based microemulsion (Gorb 2001) with low lipid content. Footprints of secretion in flies, visualized by cryo-SEM, indicate as well an emulsion-like composition (Fig. 4). Detailed fluid evaporation measurements in beetles Coccinella septempunctata and flies Calliphora vicina by atomic force microscopy supported these observations (Langer et al. 2004; Peisker and Gorb 2012). In flies, a larger volatile fraction was observed, if compared to beetles (Peisker and Gorb 2012). This result is also reflected in the corresponding viscosities of tarsal secretions in flies and beetles. For flies (C. vicina) a viscosity of approximately 11 mPa
s was measured by microrheological technique, whereas for beetles (C. septempunctata) it was two times higher, and a viscosity of approximately 22 mPa
s was found (Fig. 5) (Peisker et al. 2014). Both secretions showed a Newtonian behavior. In wet fibrillar systems, numerous experiments showed that the main contribution of attractive forces to the overall adhesion is due to capillarity. When hairy pads of the bug Rhodnius prolixus were treated with organic solvents, attachment was impaired (Edwards and Tarkanian 1970). Experiments with freely walking beetles strongly suggested that cohesive forces, surface tension, and molecular adhesion, mediated by the pad secretion, may be involved in the mechanism of attachment (Stork 1980a). Multiple local force-volume measurements, carried out on individual terminal plates of the setae of the fly C. vicina by application of atomic force microscopy (Langer et al. 2004), showed that adhesion strongly decreases as the volume of the secretion decreases, indicating that a layer of pad secretion covering the terminal plates is crucial for the generation of strong attractive force. These data provide direct evidence that, beside van der Waals forces, also attractive capillary forces, mediated by the pad secretion, contribute to the fly’s attachment mechanism (Langer et al. 2004). Similarly, for beetles, traction force experiments on nanoporous plant (Gorb and Gorb 2002) and nanoporous artificial surfaces (Gorb et al. 2010), capable of adsorbing tarsal secretions, showed strong minimization of attachment ability. In the hairy attachment system of gekkonid lizards, van der Waals interactions are responsible for the generation of attractive forces (Hiller 1968; Autumn et al. 2000, 2002). Experiments, in which the force-displacement curves were determined for individual spatulae by atomic force microscopy, showed that these smallest elements of the gecko’s attachment devices generate forces of about 10 nN. An estimate of adhesion energy (γ) from Kendall’s classical peeling theory (Kendall 1975), which describes the force required to peel a tape of certain width (200 nm for

1648

S. N. Gorb and L. Heepe

Fig. 4 Fluid in wet fibrillar systems. (a, b) Tenent setae of the syrphid fly Episyrphus balteatus bearing lumen (LU) filled with the fluid and pores (black arrows) under platelike terminal contact elements (PL). (c) Pattern of microdroplets left by the fly Calliphora vicina on the glass surface, as viewed in phase contrast light microscope. (d, e) Same droplets frozen and further replicated with carbon-platinum coating technique and visualized in transmission electron microscope (black arrow indicates direction of coating). Please note pattern of nanodrops on the surface of major droplets. (f, g) Menisci formed around single terminal contact elements of the setae of the fly C. vicina. Fly leg was frozen in contact with smooth glass, carefully removed, and the fluid residues are viewed in cryo-SEM ((a, b) From Gorb 1998. c–e From Gorb 2001. (f, g) From Gorb 2007))

gecko spatula) from a rigid substrate in perpendicular direction, gives γ = 50 mJ
m 2 (Huber et al. 2005a). This value corresponds well to the range expected for adhesion by intermolecular forces (10–100 mJ
m 2; Israelachvili 1992). Using this value of adhesion energy, we may estimate the order of the theoretical force a gecko can produce, assuming all spatulae readily make intimate contact with the substrate and peel off simultaneously. Assuming an effective total peeling line, which is the sum of widths of all spatulae, of 1.8 km for the tokay gecko (Varenberg et al. 2010) and a peeling angle of 30 , shown to be the critical angle of detachment for gecko setae (Autumn et al. 2000), we obtain a maximum force of order 700 N or correspondingly a maximum supported weight of order 70 kg. This large value indicates that only small portions of spatulae are needed in contact with the substrate at each single step. While these estimates show that van der Waals forces can account for the measured adhesion, we cannot unambiguously exclude contributions from capillary forces (Huber et al. 2005b).

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1649

Fig. 5 Plots of the mean square displacement (MSD) over the lag time for different fluids (Adapted from Peisker et al. 2014). (a) Dependence of the MSD on lag time (see black line with slope 1) for water (blue dots), the fly fluid (black dots), and the beetle fluid (red dots). The linear relationship in the log-log scale indicates purely viscous (Newtonian) properties of all fluids studied. For the fly and the beetle, all measurements were averaged. Representative MSDs for (b) water, (c) beetle fluid, and (d) fly fluid. Black solid lines are the MSD data points; the gray-shaded area is the standard deviation in each MSD data point; the white-dashed lines are linear fits to the data

55.5

Multiple Contacts

Based on the studies of different animal groups, an interesting correlation between the setal density and animal weight was found: the heavier the animal, the smaller and more densely packed are the terminal contact elements (Scherge and Gorb 2001) (Fig. 6). Using contact mechanics theories, this scaling effect was explained by applying the principle of contact splitting to the Johnson–Kendall–Roberts (JKR) contact theory (Johnson et al. 1971), according to which splitting up the contact into finer subcontacts increases adhesion on any substrate profile (Arzt et al. 2003). This

1650

S. N. Gorb and L. Heepe

Fig. 6 Dependence of the contact density of terminal contacts on the body mass in fibrillar pad systems in representatives from diverse animal groups: 1, 2, 4, 5, flies; 3, beetle; 6, bug; 7, spider; 8, gekkonid lizards (Adapted from Scherge and Gorb 2001). The systems, located above solid horizontal line, preferably rely on van der Waals forces (dry adhesion), whereas other ones mostly rely on capillary and viscose forces (wet adhesion))

relationship holds because animals cannot increase the area of the attachment devices proportionally to the body weight due to the different scaling rules for mass and surface area. Therefore, the increase of the attachment strength in hairy systems is realized by increasing the number of single contact points, i.e., by increasing the hair density. It has been shown later that this trend is different within each single lineage of organisms (Peattie and Full 2007). However, a closer look at the terminal contact elements in biological hairy adhesive systems involved in locomotion revealed that they are predominately spatula shaped (Gorb and Varenberg 2007) and cardinally differ from flat-punch- or hemisphere-ended structures (Fig. 7). Therefore, the application of the contact splitting principle to Kendall’s peeling model (Kendall 1975) also provides a proper explanation of multiple contact geometry with spatula-like terminal elements. It demonstrates that an animal’s attachment ability grows with an overall length of the peeling line, which is the sum of widths of all thin-film elements participating in contact (Fig. 8). This robust principle is found to manifest itself across eight orders of magnitude in an overall peeling line ranging from 64 μm for a little red spider mite to 1.8 km for a tokay gecko, generalizing the critical role of spatulate terminal elements in biological fibrillar adhesion (Varenberg et al. 2010). Furthermore, other advantages of such hairy adhesive systems have been proposed. For example, in a smooth, continuous adhesive contact, once a crack is initiated at the adhesive interface, the crack may easily propagate over the entire contact area until complete detachment occurs. By contrast, in a fibrillar system, the crack leads initially to the detachment of one seta, and the elastically stored energy in

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1651

Fig. 7 Cryo-SEM images of spatula-shaped thin-film terminal elements while in contact with smooth glass, in hairy attachment pads found in animals of evolutionary remote lineages. (a) Beetle (Gastrophysa viridula). (b) Fly (Calliphora vicina). (c) Spider (Cupiennius salei). (d) Tokay gecko (Gekko gecko). Arrows point in distal direction (From Varenberg et al. 2010)

the seta is released and can no longer contribute to the propagation of the crack (Jagota and Bennison 2002; Hui et al. 2004; Chung and Chaudhury 2005; Tang and Hui 2005). This principle is known as crack trapping. Thus, the crack has to be reinitiated at each individual subcontact. In addition, fibrillar systems adapt better to uneven and rough surfaces (Persson 2003; Persson and Gorb 2003; Kim and Bhushan 2007; Filippov et al. 2011). Because each seta is virtually independent of other neighboring setae, a seta can contact the asperities and valleys of a rough surface without being mechanically influenced by neighboring setae (Fig. 9a). In contrast, in a smooth system, the region around the contact with a surface asperity can be prevented from forming contact (Fig. 9b). Moreover, the effective stiffness of a fibrillar adhesive system is strongly reduced compared with the stiffness of the bulk material of setae. Therefore, the fibrillar system is more compliant to rough surfaces (Persson 2003). Since these effects are based on fundamental physical principles and mostly related to the geometry of the structure, they must also hold for artificial surfaces with similar geometry.

55.6

Contact Shape

Real biological adhesive contacts display a variety of shapes and only rarely resemble a hemisphere (Fig. 10). The influence of various contact shapes on the pull-off force for single contacts, as well as their scaling potential in contact arrays, was previously theoretically estimated (Spolenak et al. 2005) and experimentally

1652

S. N. Gorb and L. Heepe

Fs ~ 3b

b

a

Fs

F~b q

Total peeling line, L (m)

c

b/2

q

Insecta Arachnida Reptilia 104 3 Geckos 10 Anoles 102 101 Spiders Beetles 100 10–1 10–2 Flies 10–3 Mites –4 10 10–5 10–5 10–4 10–3 10–2 10–1 100 101 102

Body mass, m (g)

Fig. 8 Graphical representation of the effect of adhesion enhancement by increasing the peeling line length. (a) A single piece of elastic film peeling from a rigid substrate. Peeling force F is proportional to the film width b (at constant peel angle θ for given pair of materials). (b) A series of elastic films covering the same contact area as in (a). Peeling force Fs is threefold the value F in (a) due to an overall growth of the peeling line length calculated as a sum of individual film widths. (c) Total peeling line (the sum of widths of all terminal elements) in fibrillar attachment systems of different animals as a function of their body mass (From Varenberg et al. 2010)

Fig. 9 (a) Fibrillar and (b) smooth adhesive systems making contact with a rigid substrate with an asperity. Whereas in a fibrillar system, individual fibers can deform and follow surface irregularities without affecting neighboring setae (a), in smooth elastic pads, regions around the asperity can be prevented from forming contact (arrows; b) (From Heepe and Gorb 2014)

demonstrated for artificial fibrillar adhesive systems (del Campo et al. 2007). It was concluded that non-spherical shapes, such as toroid, should lead to better attachment. In insects, spiders, and geckos, as shown in Fig. 7, the topmost hierarchical level of a seta that is responsible for the formation of intimate contact with the substrate

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1653

Fig. 10 Shapes of attachment devices in nature and their hypothetical evolutionary pathways (shown as arrows). (a) Bug Pyrrhocoris apterus, smooth pulvillus. (b) Grasshopper Tettigonia viridissima, surface of the attachment pad. (c) Fly Myathropa florea, unspecialized hairs on the leg. (d) Fly Calliphora vicina, seta of the pulvilli. (e) Ladybird beetle Harmonia axyridis, seta of the second tarsal segment. (f) Beetle Chrysolina fastuosa, seta of the second tarsal segment. (g) Male diving beetle Dytiscus marginalis, suction cups on the vertical side of the foreleg tarsi (Adapted from Spolenak et al. 2005). Band-like (tape-like, spatula-like) terminal contact elements (highlighted in gray) are most often observed in biological fibrillar adhesive systems capable of ceiling locomotion

resembles thin film-like or spatulae (Stork 1980b; Gorb 1998, 2000; Persson and Gorb 2003; Spolenak et al. 2005). Additionally, spatulae bear a gradient of thickness from the base to the tip of the spatula (fly, Gorb 1998; gecko, Persson and Gorb 2003; beetle, Eimüller et al. 2008) and a gradient in width as they become wider toward its free end. This geometrical property of setal tips can be observed in many animal species (Figs. 3 and 7). In contact, spatulae are aligned and orientated to the distal direction of the pad. Several hypotheses have been previously proposed to explain the functional importance of such a contact geometry: (1) enhancement of adaptability to the rough substrate (Persson and Gorb 2003), (2) contact formation by shear force rather than by normal load (Autumn et al. 2000), (3) increase of total peeling line due to using an array of multiple spatulae (Varenberg et al. 2010), and (4) contact breakage by peeling off (Gao et al. 2005).

1654

S. N. Gorb and L. Heepe

Fig. 11 The role of the spatula-like terminal contact elements at the tips of adhesive setae in the adaptation to the surface profile (a–c) and in the increasing of the total peeling line. (a) Contact of the spatula of the beetle Gastrophysa viridula with the microrough surface. (b–c) Thick elastic material requires additional load to form adhesive contact (b), whereas adhesion interaction pulls the elastic thin film (even made of stiff material) into complete contact with the rough substrate surface (Adapted from Persson and Gorb 2003)

It appeared that the role of highly flexible terminal spatula elements as compliant contacting surfaces is critical (Persson and Gorb 2003). Since the effective elastic modulus of thin plates is very small, even for relatively stiff materials such as keratin or arthropod cuticle, this geometry is of fundamental importance for adhesion on rough substrates (Persson and Gorb 2003), due to the low deformation energy stored in the material during intimate contact formation (Fig. 11). For adhesion activation, it is well known that the application of a normal load can enhance adhesion (Popov 2010). However, for hairy attachment systems, the adhesion force always remains smaller than the initially applied normal force. This would potentially be not enough to enable walking on the ceiling. Another possibility to activate or enhance adhesion forces is the application of a shear force. We have previously shown that an applied shear force to a particular direction of the spider’s spatulated setae results in an increase of real contact area (Niederegger and Gorb 2006; Filippov et al. 2011). Also, flies employ shear movements during contact formation by their attachment devices (Niederegger and Gorb 2003). Previous authors revealed strong shear dependence of measured pull-off force in the gecko attachment system and even called this effect by the somewhat misleading term frictional adhesion (Autumn et al. 2000, 2006), which can be simply explained by Kendall’s peeling model (Kendall 1975), when the tape is loaded/peeled off at very shallow angles to the substrate. In contrast, by peeling the tape off at very high angles, a very strong reduction in the adhesive force is obtained, which allows for a simple, quick, and energy-efficient detachment mechanism.

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

55.7

1655

Slope and Hierarchy

The observed pull-off force of an adhesive structure may be defined by the difference between attractive forces arising due to, e.g., van der Waals or capillary interaction and the contact reaction forces, arising due to elastic deformation of the adhesive structure during contact deformation (Varenberg et al. 2011). In order to increase the pull-off force, either the attractive interactions have to increase, or the contact reaction forces have to decrease. The attractive forces, e.g., due to van der Waals interactions, reach a natural upper limit as soon as the adhesive structure and the contacting substrate form an intimate contact. In order to form intimate contact, especially to rough surfaces, adhesive structures have to be as soft as possible to enhance real contact area. This in turn, however, would increase the elastic deformations and thus the contact reaction force which would lead to a lower pull-off force. In light of the above, it is clear that there must be a trade-off for the adhesive structure between being soft enough, in order to form an intimate contact, and being stiff enough, to allow for the generation of a sufficiently high pull-off force. Solutions for this problem are either (a) in a hierarchical fibrillar design of the adhesive pad (Fig. 3d–f), (b) in a gradient material in nonhierarchical fibrillar structures (Fig. 12), or (c) in the tilted orientation of hairs relatively to the substrate (Fig. 13). In case of the gecko (see above), the complex hierarchical architecture of the adhesive pad made of relatively stiff β-keratin allows for a relatively low effective modulus (Persson 2003) while simultaneously avoiding setal condensation. In geckos, there is a higher number of structural levels made of relatively stiff β-keratin (see above): lamella (the first lowermost level), seta (the second one), branching pattern of setulae (third and sometimes fourth level(s)), and spatulae (the uppermost one). Since many natural surfaces (stones, soil, tree bark, plants) have fractal roughness with several overlapping wavelengths, it is not enough to have adhesive pads optimized for one a particular wavelength of roughness. In order to optimize adhesion on a wide range of natural surfaces, biological fibrillar adhesives employed several hierarchical levels, each responsible for contact formation optimization for a particular range of wavelengths. In general, biological adhesive systems represent a kind of mirror structure of the natural fractal world. Recently, numerous models were developed, in order to explain the role of the hierarchy in the functioning of biological adhesive pads (e.g., Kim and Bhushan 2007; Schargott 2009). While the hierarchical beam model with vertical beams shows the possibility of reducing the effective elastic modulus, the tilted beams are the key to obtaining very soft systems, even when the material employed has a high elastic modulus. In the model combining hierarchy and tilting, elastic and adhesive properties are furthermore enhanced (Schargott 2009). It is shown that when interacting with a rough stochastic surface, the attachment properties are improved with each additional hierarchical layer (Kim and Bhushan 2007; Schargott 2009). Animals with fewer hierarchical levels have evolved different strategies for the same purpose. In the seven-spotted ladybird beetle (Coccinella septempunctata), a material gradient of tanned chitin at the setal base and the rubberlike resilin at the

1656

S. N. Gorb and L. Heepe

Fig. 12 Morphology and material composition of adhesive tarsal setae. Ventral part of the second adhesive pad of a foreleg of a female Coccinella septempunctata, lateral view. (a) Scanning electron micrograph. (b) CLSM maximum intensity projection showing an overlay of the four different autofluorescences indicating the material composition. The arrows indicate the dorsoventral material gradient in exemplary setae. S exemplary spatula-like seta, P exemplary seta with a pointed tip. Scale bars, 25 mm (From Peisker et al. 2013)

setal tip were recently found (Fig. 12; Peisker et al. 2013). Such a gradient in material properties similarly allows for an efficient adaptation to rough surfaces, due to the soft setal tip, while, at the same time, providing sufficient mechanical stability, due to stiffer base, to prevent clusterization and low contact reaction forces. Tilted geometry of hairs is responsible for the bending mode of their deformation in contact rather than for the buckling one. It is clear that hairs are much stiffer in buckling than in bending. That is why tilted hair arrays have a much lower structural modulus of elasticity than a non-tilted one, and the adaptability to the surface profile of the tilted array is much higher. In addition, bending as a failure mode leads to lesser material fatigue than the buckling mode. Additionally, asymmetrical shape of a single contact element in combination with the proper movements may provide the way to switchable adhesives, when contacts formed by fibers are activated and when the structure is sheared in the direction of the slope. The contact can be broken by peeling at high angles, when fibers are sheared in an opposite direction.

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1657

Fig. 13 Slope angle of adhesive setae in the tokay gecko Gekko gecko (a, b) (SEM) and in the fly Calliphora vicina (c) (semithin section, light microscopy). Arrows indicate distal direction of the pad (From Gorb 2011)

55.8

Adhesion Control

Animal locomotion involves simultaneous peeling of many spatulae, at the level of adhesive pads (Fig. 7), at the level of contralateral legs (Fig. 14a, b), and sometimes also at the level of contralateral toe pads within one foot (Fig. 14c). Thus, in climbing animals, the situation is more complex than in Kendall’s classical peeling model and resembles a hierarchical multiple peeling configuration (Pugno and Gorb 2009; Pugno 2011; Heepe et al. 2017a). On the macroscale, however, the peeling configuration of climbing animals may be reduced, for simplicity, to the peeling of two elastic tapes loaded at a common hinge (Fig. 14d). This double-peeling configuration has interesting adhesive properties, if compared to the classical single peeling. It exhibits a self-stabilized optimum peeling angle with a maximum in peeling force, i.e., the optimum configuration is reached irrespective of the initial peeling angle configuration of the system. Below the optimum angle, there is no delamination (peeling) but pure elastic deformation of the system. Above the optimum angle, stable delamination is observed. Figure 15 shows experimental results of the doublepeeling configuration using an elastic adhesive tape (Heepe et al. 2017a). When an animal is adhering to the ceiling, its legs tarsomeres are oriented toward the center of the body mass (Fig. 16a). In this orientation, setae are recruited by bending to their tilted direction, and spatulae are set under shear load. Under these conditions, setae are activated to their adhesive state. If five of the six insect legs are ablated, the animal is not able to adhere to the ceiling anymore, even if the average adhesive force of the single leg (measured for all six legs acting in concert and then normalized by the number of legs) should be sufficient to resist the body weight on the ceiling. However, with two intact legs, an animal can perfectly stay on the ceiling

1658

S. N. Gorb and L. Heepe

Fig. 14 (a) A beetle Gastrophysa viridula in contact with a glass ceiling (lateral view from behind). The arrow indicates weight force W acting at the center of mass. (b) Fly Calliphora vicina in contact with a glass ceiling (view from above through the glass ceiling). Arrows indicate contralateral leg movement. (c) Foot of a gecko Gekko gecko in contact with a glass ceiling (view from above through the glass ceiling). Arrows indicate contralateral toe pad movement. (d) Simplified mechanical model of the ceiling situation (lateral view from behind, similar to (a). Two elastic tapes are in contact with a rigid ceiling and having a common hinge, where a load F is applied. The tapes make angles with respect to the substrate of α1 and α2 (From Heepe et al. 2017)

(Fig. 16b). With three remaining intact legs on one body side, an animal can reorient one leg to the other body side and stay on the ceiling (Fig. 16c). In the latter two cases, the adhesive system can be activated by shear force. In light of the above results, the explanation of this phenomenon is rather simple. The situation of the single leg in contact can be simulated by a piece of sticky tape adhering to the ceiling and holding a certain mass. The peeling angle in such geometry would tend to be 90 , which according to Kendall’s model (Kendall 1975) generates a rather low peeling force. When two or multiple tapes are combined, the peeling angle will be generally lower than 90 , due to the self-stabilization effect of such geometry. High-speed video recordings of a fly foot attaching to the substrate show that the pulvilli are pressed down to the surface and also pulled towards the body. Through this movement, the claws become spread apart (Fig. 17a, b). During such an action, setae and their spatulae will be activated, and the maximal force will be reached. In order to detach, four different movements may occur (Fig. 17c–f) (Niederegger et al. 2001;

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1659

Fig. 15 (a) Schematic of the experimental setup. Principal double-peeling configuration: the tape (dark gray) is in contact with glass substrates (light gray) and loaded with a force F at the hinge in the middle of the tape. The tape makes initial peeling angles with the substrate of α0. (b) Experimental results (open circles) of the double-peeling configuration using elastic adhesive tape. Experiments were started at initial peeling angles 0 and 90 as indicated by the arrows (Adapted from Heepe et al. 2017a)

Fig. 16 (a–c) Posture of the chrysomelid beetle Chrysolina fastuosa on the ceiling. (a) Intact beetles. (b) Two-legged beetles (four legs are removed). Arrows indicate orientation of leg tarsomeres. (c) Three-legged beetles (three legs on one body side are removed). Please note rearrangement of leg positions, in order to generate shear toward center of the body mass (arrows) (From Gorb 2011)

1660

S. N. Gorb and L. Heepe

Fig. 17 Attachment and detachment movements of the fly adhesive organs (diagrams are based on the high-speed video recordings under binocular microscope). (a) Side view of attachment organs (pulvilli) in the fly Calliphora vicina (SEM). (b) Movement at attachment. (c–f) Movements at detachment by using shifting (c), pulling (d), twisting (e), and rotation (c) (From Niederegger et al. 2001; Niederegger and Gorb 2003)

Niederegger and Gorb 2003). (1) In shifting (or local peeling), the foot is shortly moved forward. This compresses the pulvilli, which in turn leads to a smaller contact area and thus to a reduction of attachment forces. Additionally, peeling occurred at the level of the spatulae. (2) In twisting, the leg is turned by up to 90 and pulled backward. This bends the setae and eventually detaches them from the surface. (3) In rotation (or global peeling), the foot is lifted from its back to the front, which causes the claws to be passively pressed to the surface. With this force, the adhesion is interrupted by peeling at the level of the whole pad. (4) In pulling, the foot is simply pulled backward until the attachment devices detach from the surface. This type of detachment possibly requires quite strong force, but it happens in some behavioral situations, especially when the insect is stressed. In geckos, attachment is activated by proximodistal rolling out of the toe, whereas detachment is provided by distoproximal peeling of it (Gao et al. 2005). Observations of the gait pattern in flies walking on a horizontal surface revealed that three opposing legs were in the swing phase (moving), whereas the other three legs remained in the stance phase (motionless) (Fig. 18a). Does the gait pattern of the fly change on the wall and ceiling? Using high-speed camera video recordings of

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1661

Fig. 18 Gait patterns of insects on glass ceiling. (a, b) Gait diagram of the house fly, Musca domestica, on the floor (a) and ceiling (b) (From Niederegger et al. 2001; Gorb 2001, 2005). (c) Relationship between the safety factor (attachment force divided by the body weight) and the gait pattern on the ceiling in different insects. Insets show feet (black dots) that are typically in contact during the walk on smooth ceiling (From Gorb 2011)

flies walking on a vertical surface, it was shown that their gait pattern is not different from the pattern on a horizontal surface. However, the fly walking on the ceiling moved only two legs at once and four remained in the stance phase (Niederegger et al. 2001; Gorb 2001, 2005) (Fig. 18b). This result led us to assume that the gait pattern may contribute to the fly’s ability to adhere dynamically on the ceiling, in providing an optimal relationship between the body weight and the supporting contact area. Experiments with insect species that have different safety factors (attachment force divided by the body weight) show that insects with a safety factor above 10 can employ tripod gait similar to one observed during locomotion on the floor. Insects with a safety factor in the range of 4–5 change their gait compared to that reported above for the fly. Insects with a safety factor lower than 3 change their gait to that, where only one leg is in the swing phase and the other five remain in the stance phase (Fig. 18c). These results show that dynamical adhesion in insects may also strongly rely on a multiple peeling configuration, especially in ceiling locomotion. Thus, it is plausible to assume that biological adhesive systems try to actively use changes in their configuration of their tape-like contact geometry, to control and/or maintain adhesion. To securely stay on the ceiling, animals tend to keep their macroscopic peeling angle below the optimum. In this regime, no peeling occurs. Interestingly, in flies clinging to a glass ceiling, it was observed that their feet were in constant slow motion toward the body center of mass (Wigglesworth 1987), potentially due to the tarsal secretions produced in insect adhesive systems (Gorb 2001). After a certain sliding motion of the feet, they were quickly extended and reattached, thereby maintaining a position with a low macroscopic peeling angle. Quick and

1662

S. N. Gorb and L. Heepe

easy detachment may be achieved by a simple active change of the macroscopic peeling angle above the optimum one by, e.g., proximodistal peeling in the case of flies (Niederegger et al. 2002) and distoproximal peeling or digital hyperextension in the case of geckos (Russel 1975). This may be possible, since their available contact area is restricted to the pad area and, thus, only a short peeling distance is necessary before complete detachment of a pad or a foot occurs.

55.9

Environmental Conditions Affecting Animal Adhesion

All natural surfaces have surface roughness on many different length scales (Persson 2014) which may strongly reduce attachment of animals (e.g., Gorb 2001; Gorb and Gorb 2002; Peressadko and Gorb 2004b; Huber et al. 2007; Voigt et al. 2008; Wolff and Gorb 2012). The influence of surface roughness on insect attachment has been proven experimentally in several studies (Gorb 2001; Peressadko and Gorb 2004b; Voigt et al. 2008; Bullock and Federle 2011). There seems to be a common trend that insects generate much higher forces on either smooth or rough surfaces with an asperity size exceeding 3.0 μm than on those with the roughness ranging from 0.3 to 3.0 μm (Fig. 19). This effect has been explained by the specific geometry of spatulalike terminal elements of insect tenent setae that are able to generate sufficient contact, if the surface irregularities are sufficiently large. Worst attachment has been observed on substrates with a roughness of 0.3 and 1.0 μm (Fig. 19). In this case, the spatula-like terminal elements cannot follow the surface profiles with these small asperities, and thus, the area of real contact between these substrates and the tips of insect setae is very small. Since adhesion force depends on the area of real contact, insects were not able to attach successfully to surfaces with such a microroughness. Interesting examples of natural anti-adhesive surfaces are well known from many plants, which are covered with microscopic wax crystals of different shapes, chemistry, and mechanical properties. To explain the anti-adhesive properties of these plant substrates, four hypotheses were previously proposed (Gorb and Gorb 2002). (i) Wax crystals cause microroughness, which considerably decreases the real contact area between the substrate and setal tips of adhesive pads (roughness hypothesis); see above. (ii) Wax crystals are easily detachable structures that contaminate pads (contamination hypothesis). (iii) Structured wax coverage may absorb the fluid from the setal surface ( fluid absorption hypothesis). (iv) Insect pad secretion may dissolve wax crystals (wax-dissolving hypothesis). This would result in the appearance of a thick layer of fluid, making the substrate slippery. Recently, only the first three hypotheses were tested. Contamination of insect pads by plant wax crystals has been demonstrated for several insects and in a series of plant species (e.g., Gaume et al. 2004; Gorb and Gorb 2006). It has been found that plants differ essentially in their contaminating effects on insect pads (Gorb and Gorb 2006). The degree of contamination depends on the micromorphology of waxes (Borodich et al. 2010). The analysis of the relationship between the contamination ability and geometrical parameters of wax

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1663

Fig. 19 Role of the substrate roughness (asperity diameter) in the attachment of biological fibrillar adhesives. (a) Attachment forces of chrysomelid beetles Gastrophysa viridula measured on the horizontal surface of the drum of the centrifugal force tester on substrates with different roughness. (b) Surface of the substrate in SEM and measured with the profilometer (From Peressadko and Gorb 2004b)

crystals has shown that the contamination is related to both the largest dimension and the aspect ratio of crystals (Borodich et al. 2010). The fluid absorption hypothesis was recently tested in an experimental study, where nanoporous substrates with the same pore diameter but different porosity (area of voids in a material surface, normalized over the total area) were used (Gorb et al. 2010). According to this hypothesis, absorption of the insect pad fluid by the porous substrate results in a reduction in the fluid thickness between the terminal plates of the insect pads and the substrate, and this leads to a reduction of capillary interaction in the contact area. Traction force experiments performed on tethered ladybird beetles walking on different porous substrates showed that forces were significantly reduced. The reduction in insect attachment on nanoporous surfaces may be explained by possible absorption of the secretion fluid from insect adhesive pads by porous media and/or the effect of surface roughness. The wax-dissolving hypothesis has not been experimentally tested yet. A rather long-standing question in animal attachment is which substrate property is most important for the attachment ability: surface roughness or surface chemistry? Since the surface chemistry (here in terms of surface energy) affects the strength of

1664

S. N. Gorb and L. Heepe

either van der Waals or capillary interaction, substrates with low surface energy are expected to reduce the overall attachment ability. In a recent experiment, traction forces of ladybird beetles Coccinella septempunctata were systematically measured on eight types of substrates, each with different chemical and topographical properties (England et al. 2016). The results clearly showed that surface roughness rather than surface chemistry essentially affected insect adhesion. Surface chemistry had no significant effect on the attachment ability of the beetles (England et al. 2016). So far, however, the attachment ability of insects and also that of spiders and geckos has been tested on rigid substrates only, whereas the natural habitats of climbing animals may provide a variety of substrate stiffness ranging from rigid rock surfaces to soft, biofilm-covered substrates. Friction force measurements, using a centrifugal force tester, with male and female beetles Coccinella septempunctata on smooth silicone elastomer substrates with different stiffness between 2 and 0.3 MPa revealed an overall decrease in attachment ability if compared to a rigid substrate (Heepe et al. 2017b). Within the range of measured stiffness, female’s attachment ability was not affected, whereas male’s decreased with decreasing stiffness. This sexual dimorphism in attachment ability is explained by the presence of a specialized, discoidal seta type in males, which is not present in females (Heepe et al. 2017b). Substrate properties are not the only environmental factors shown to influence animal adhesion. Also the ambient humidity affects the attachment ability, as it was shown in the dry adhesive pads of geckos (Huber et al. 2005a; Niewiarowski et al. 2008; Puthoff et al. 2010; Prowse et al. 2011) and spiders (Wolff and Gorb 2011) as well as in the wet adhesive pads of beetles (Heepe et al. 2016). In all animals, reduced attachment ability was observed for very low and very high relative humidities, whereas maximum attachment ability was found for intermediate humidities (Heepe et al. 2016). This is particularly interesting since both types of adhesive systems (wet and dry) are supposed to be based on different physical interactions (capillarity versus van der Waals forces).

55.10 Biomimetic Implications As mentioned above, fibrillar adhesive systems appeared several times in animal evolution and at least three times independently even in insect evolution. This fact may indicate that design principles of biological fibrillar adhesives must have an advantage for adhesion enhancement not only in biological systems but also in artificial surfaces having similar geometry. Geometrical effects discussed above, such as multiple individual contacts, high aspect ratio of single contact structures, peeling spatula-like tips of single contact elements, are responsible not only for generation of a strong pull-off force in such devices but also for other interesting features, such as adhesion reversibility, contamination reduction, etc. (Fig. 20). The physical background of these effects were intensively discussed theoretically in several publications (for early original work, see, e.g., Jagota and Bennison 2002; Arzt et al. 2003; Persson 2003; Persson and Gorb 2003; Chung and Chaudhury 2005; Gao et al. 2005; and for current reviews, Federle 2006; Autumn 2007;

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1665

Fig. 20 Some functional principles (FEATURE) at which biological reversible and biological adhesive systems operate. In the middle of the diagram, the functions (FUNCTION) and their relationship to the features are depicted. The resulting effect that is required for optimization is shown at the right-hand side (arrows indicate enhancement/reduction of the function). Simultaneous implementation of all these features in one artificial system is desirable but hardly possible. However, one principle or a combination of few of them can be implemented depending on particular requirements for material or system (From Gorb et al. 2007b)

Kampermann et al. 2010; Gorb 2011; Jagota and Hui 2011). For fibrillar biological systems, the following key functional properties can be defined (Autumn et al. 2007; Creton and Gorb 2007). 1. Anisotropic attachment: Attachment force variable depending on the setal-spatula orientation with respect to the substrate, normal load, and parallel drag. 2. High adhesion coefficient: Ratio of preload to pull-off force, which represents the strength of adhesion as a function of the compressive load. 3. Low detachment force: Animals detach their feet in a few milliseconds by different kinds of movement, which lead to peeling of adhesive structures at different levels. 4. Universality of adhesion: Van der Waals-based interaction and oil-based capillary interaction with the substrate depend basically on geometry of features (fibrillar shape, spatulae, fluid thickness, etc.) and are less dependent on the substrate chemistry.

1666

S. N. Gorb and L. Heepe

5. Self-cleaning: The feet of geckos are resistant to the contamination by particles corresponding in size to typical environmental dirt particles (diameter 5–100 μm). In insects, secreted fluid may additionally contribute to rinsing of dirt particles. Biological fibrillar adhesive systems possess an anti-contaminating property. 6. Anti-self-adhesion: Hierarchical fibrillar structure avoids the self-adhesion of individual structural elements. 7. Nonsticky default state: Most of the fibrillar structures with an exception of mushroom-shaped setae are nonsticky by default, because only a very small contact fraction is possible without deforming the setal array. Based on the inspiration from biology and using contact theory as a guideline, artificial surfaces were developed with enhanced pull-off forces in contact with the flat surface, if compared to the flat control. These materials have been produced using various micro- and nanofabrication techniques ranging from laser technology and carbon nanotube packaging to various lithography techniques (for early original work, see, e.g., Geim et al. 2003; Sitti and Fearing 2003; Peressadko and Gorb 2004a; Northen and Turner 2005; Yurdumakan et al. 2005; and for current reviews, e.g., Greiner et al. 2009; Sameoto and Menon 2010; Kwak et al. 2011) (Fig. 21). Independently on the pull-off forces achieved by the surface patterning, these materials are strongly limited in the patterned area. Usually, the overall patterned area was in the best case restricted to few square centimeters, which makes their use for industrial applications rather difficult. Also most bioinspired fibrillar adhesives are just based on the fibers with flat punches at their tips or in most elaborate cases sloped punches (Autumn et al. 2007), without implementation of proper terminal elements in the form of spatula or mushroom, as we know them from the majority of biological systems including insects, spiders, and geckos. So far, the best results of mimicking fibrillar adhesive systems have been achieved with mushroomlike surface microstructures (for early original work, see, e.g., Daltorio et al. 2005; Kim and Sitti 2006; Gorb et al. 2007a; Davies et al. 2009; for a current review, see Heepe and Gorb 2014). Some of these developments may be produced in large-scale industrial processes in the form of square meters large foils (Gottlieb Binder and Co 2017) (Fig. 22). Interestingly, such microstructures can be found in attachment pads of male beetles from the family Chrysomelidae. Although both sexes possess adhesive hairs on their tarsi, only males bear hairs with such extreme specialization for adhesion on the smooth surface and most likely to attach to female’s covering wings during pairing (Stork 1980b; Voigt et al. 2008). It has been previously reported that males of the Colorado potato beetle Leptinotarsa decemlineata generate such a strong adhesion on the smooth clean glass that they are not able to break off the contact and that is why they have a hard time walking on this surface (Pelletier and Smilowitz 1987). The hairs responsible for this effect have broad flattened tips and narrowed flexible region below the tips. These features, as well as distribution patterns of pillars, were implemented in the design of the patterned polymer structure described here. The adhesive properties of the industrial micropatterned adhesive foil were characterized using a variety of measurement techniques and compared with those

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1667

Fig. 21 Bioinspired fibrillar adhesives of different dimensions. (a, b) Metallic nanowhiskers. (c, d) Hierarchically branching fiber arrays made of epoxy resin. (e, f) Silicone molds of laser-patterned metal surface. (g, h) Carbon nanotubes. (a–d, g, h) Original micrographs ((a–d, g, h) From Gorb 2010. (e, f) From Peressadko and Gorb 2004a)

of the flat foil made of the same polymer (Daltorio et al. 2005; Gorb et al. 2007a; Varenberg and Gorb 2008a, b, c; Heepe et al. 2011, 2012, 2013, 2014a, b, 2017a; Kovalev et al. 2012; Kizilkan et al. 2013, 2017; Kasem and Varenberg 2013; Dening et al. 2014; Breckwoldt et al. 2015). The foil with the microstructure pattern demonstrates considerably higher pull-off force per unit real contact area and per

1668

S. N. Gorb and L. Heepe

Fig. 22 Patterned insect-inspired polivinylsiloxane surface. (a) Single structures are distributed on the surface according to the hexagonal pattern, in order to reach the highest packaging degree of single pillars (above aspect, SEM image). (b) White light interferometer image of a single pillar head demonstrates almost flat shape of the contacting surface. (c) Side aspect of the pillar array. (d–f) Behavior of structured PVS surfaces in contact with the glass surface (SEM images). Black arrowhead shows a dust particle in contact (From Gorb et al. 2007b)

unit apparent contact area as well. It is also less sensitive to contamination by dust particles, and after being contaminated and washed with soap and water, the adhesive properties can be completely recovered. The foil represents an industrial dry adhesive based on the combination of several principles previously found in biological attachment devices of insects, spiders, and geckos. This material is a promising candidate for the application in dynamical adhesive systems, such as robot soles adapted for locomotion. Robots that could climb smooth and complex inclined terrains, like insects and lizards, would have many

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1669

applications such as exploration, inspection, or cleaning. Walking machines usually use suckers to hold onto vertical surfaces and under a surface. A primary disadvantage of this attachment principle is large energy consumption for vacuum maintenance. The novel biologically inspired materials may enable future robots to walk on smooth surfaces regardless of the direction of gravity. Mini-Whegs™, a small robot (120 g) that uses four-wheel legs for locomotion, was recently converted to a wall-walking robot with compliant, adhesive feet (Daltorio et al. 2005). The robot is capable of ascending vertical smooth glass surfaces using a micropatterned adhesive (Fig. 23a). A new version of this robot (about 22 g) is even able to transverse from the floor to the ceiling without detaching from the substrate (Fig. 23b). Another demonstration of adhesion capability of such material is shown in Fig. 23c, d. A PMMA plate, 20  20 cm, covered by the foil is sufficient to support the weight of a man attached to the glass ceiling (Fig. 23c). Using custom-built gloves for hands and feet, the material, in principle, allows climbing a vertical glass wall (Fig. 23d). Bioinspired fibrillar materials can be additionally applied to the design of microgripper mechanisms with the ability to reversibly adhere and to adapt to a variety of surface profiles and for any kind of reversible adhesive tapes.

55.11 Conclusions 1. Real biological attachment systems usually rely on the combination of different functional principles based on elaborate micro- and nanostructures. 2. Attachment systems in biology can be subdivided according to their structure and basic physical forces responsible for enhancement of contact forces. 3. Reversible attachment devices used for locomotion may be hairy and smooth. Hairy systems consist of fine microstructures, whereas smooth systems are composed of materials with unusual inner structure. Due to the flexibility of the material of the attachment pads or fine surface structures, both mechanisms can maximize the possible contact area with a wide range of substrate profiles. 4. Recent phylogenetic analyses have shown that similar functional solutions for attachment have evolved several times independently in the evolutionary history of animals (e.g., geckos, anoles, spiders, beetles, flies, etc.). 5. In hairy systems, the density of single contact elements usually increases with increasing body weight of the animal group. From the scaling analysis, it has been previously suggested that animal lineages relying on dry adhesion (lizards, spiders) possess a much higher density of terminal contact elements compared to systems using the wet adhesive mechanism (insects). 6. Lizards’ and spiders’ attachment systems are mainly based on van der Waals forces, but also wetting phenomena caused by water absorbed on the surface have an effect on the attachment ability. In insects, mainly attractive capillary and viscous forces mediated by the pad secretion are the basic physical mechanism in their adhesion.

1670

S. N. Gorb and L. Heepe

Fig. 23 (a) Mini-Whegs™ on vertical glass with a microstructured adhesive tape (From Daltorio et al. 2005). (b) Inverted Mini-Whegs demonstrating internal transitions from floor to wall to ceiling (From Breckwoldt et al. 2015). (c) Photograph of a man attached to the glass ceiling by a 20  20 cm PMMA plate covered by microstructured adhesive tape (see SEM image in the inset) (original). (d) Photograph of a man attached to a vertical glass wall by custom-built gloves for hands and feet covered with the microstructured adhesive tape (original)

7. The effective elastic modulus of the fiber arrays is very small, which is of fundamental importance for proper contact formation and adhesion on smooth and rough substrates.

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1671

8. Hierarchical organization of surface features in attachment pads may enhance adaptability to real surfaces, which often have fractal roughness. 9. Asymmetrical shape of single contact element (tilted position relative to the support) in combination with proper movements may provide a way to switchable adhesives. 10. Hairy pads show a clearly cut minimum of adhesion at certain ranges of substrate roughness. Such a critical range of roughness depends on the relationship between the diameter of single contact elements of the pad and the length scale of the roughness. 11. During their evolution, plants have developed structures to prevent adhesion of insects. To explain anti-adhesive properties of plants, structured with crystalline waxes, four hypotheses were proposed: (1) roughness hypothesis, (2) contamination hypothesis, (3) wax-dissolving hypothesis, and (4) fluid absorption hypothesis. 12. Since many functional effects found in biological systems are based on fundamental physical principles which are mostly related to the geometry of the structure, they must hold also for artificial surfaces with similar geometry.

References Alibardi L (1997) Ultrastructural and autoradiographic analysis of setae development in the embryonic pad lamellae of the lizard Anolis lineatopus. Ann Sci Nat Zool Biol Anim 18:51 Arzt E, Gorb SN, Spolenak R (2003) From micro to nano contacts in biological attachment devices. Proc Natl Acad Sci U S A 100:10603 Autumn K, Liang YA, Hsieh ST, Zesch W, Chan WP, Kenny TW, Fearing R (2000) Adhesion force measurements on single gecko setae. Nature 405:681 Autumn K, Sitti M, Liang YA, Peattie AM, Hansen M (2002) Evidence for van der Waals adhesion in gecko setae. Proc Natl Acad Sci U S A 99:12252 Autumn K, Dittmore A, Santos D, Spenko M, Cutkosky M (2006) Frictional adhesion: a new angle on gecko attachment. J Exp Biol 209:3569 Autumn K, Gravish N, Wilkinson M, Santos D, Spenko M, Cutkosky M (2007) Frictional adhesion of natural and synthetic gecko setal arrays. In: Proceedings of 30th annual meeting adhesion society, Inc, The Adhesion Society, Blacksburg, VA Barnes WJP (2006) Whole animal measurements of shear and adhesive forces in adult tree frogs: insights into underlying mechanisms of adhesion obtained from studying the effects of size and scale. J Comp Physiol A 192:1179 Bauchhenss E (1979) Die Pulvillen von Calliphora erythrocephala (Diptera, Brachycera) als Adhäsionsorgane. Zoomorphologie 93:99 Betz O (2010) Adhesive exocrine glands in insects: morphology, ultrastructure, and adhesive secretion. In: Byern J, Grunwald I (eds) Biological adhesive systems. From nature to technical and medical application. Springer, Vienna, pp 111–152 Beutel RG, Gorb SN (2001) Ultrastructure of attachment specializations of hexapods (Arthropoda): evolutionary patterns inferred from a revised ordinal phylogeny. J Zool Syst Evol Res 39:177 Beutel RG, Gorb SN (2006) A revised interpretation of the evolution of attachment structures in Hexapoda with special emphasis on Mantophasmatodea. Arthrop Syst Phylogeny 64(1):3–25 Borodich FM, Gorb EV, Gorb SN (2010) Fracture behaviour of plant epicuticular wax crystals and its role in preventing insect attachment: a theoretical approach. Appl Phys A Mater Sci Process 100:63

1672

S. N. Gorb and L. Heepe

Breckwoldt WA, Daltorio K, Heepe L, Horchler AD, Gorb SN, Quinn R (2015) Walking inverted on ceilings with wheel-legs and micro-structured adhesives. In: Intelligent robots and systems (IROS), IEEE/RSJ international conference on. IEEE, Hamburg, Germany, pp 3308–3313 Bullock JMR, Federle W (2011) The effect of surface roughness on claw and adhesive hair performance in the dock beetle Gastrophysa viridula. Insect Sci 18:298 del Campo A, Greiner C, Arzt E (2007) Contact shape controls adhesion of bioinspired fibrillar surfaces. Langmuir 23:10235 Chung JY, Chaudhury MK (2005) Roles of discontinuities in bio-inspired adhesive pads. J R Soc Interface 2:55 Creton C, Gorb SN (2007) Sticky feet: from animals to materials. MRS Bull 32:466 Daltorio KA, Gorb SN, Peressadko A, Horchler AD, Ritzmann RE, Quinn RD (2005) A robot that climbs walls using micro-structured polymer feet. In: Proceedings of international conference on climbing and walking robots CLAWAR, London, UK, pp 131–138 Davies J, Haq S, Hawke T, Sargent JP (2009) A practical approach to the development of a synthetic Gecko tape. Int J Adhes Adhes 29:380 Dening K, Heepe L, Afferrante L, Carbone G, Gorb SN (2014) Adhesion control by inflation: implications from biology to artificial attachment device. Appl Phys A Mater Sci Process 116:567 Edwards JS, Tarkanian M (1970) The adhesive pads of Heteroptera: a re-examination. Proc Roy Ent Soc Lond A 45:1 Eimüller T, Guttmann P, Gorb SN (2008) Terminal contact elements of insect attachment devices studied by transmission X-ray microscopy. J Exp Biol 211:1958 Eisner T, Aneshansley DJ (2000) Defense by foot adhesion in a beetle (Hemisphaerota cyanea). Proc Natl Acad Sci U S A 97:6568 England MW, Sato T, Yagihashi M, Hozumi A, Gorb SN, Gorb EV (2016) Surface roughness rather than surface chemistry essentially affects insect adhesion. Beistein J Nanotechnol 7:1471 Federle W (2006) Why are so many adhesive pads hairy? J Exp Biol 209:2611 Federle W, Riehle M, Curtis ASG, Full RJ (2002) An integrative study of insect adhesion: Mechanics and wet adhesion of pretarsal pads in ants. Integr Comp Biol 42:1100 Filippov AE, Popov VL, Gorb SN (2011) Shear induced adhesion: Contact mechanics of biological spatula-like attachment devices. J Thero Biol 276:126 Gao H, Wang X, Yao H, Gorb SN, Arzt E (2005) Mechanics of hierarchical adhesion structures of geckos. Mech Mater 37:275 Gaume L, Perret P, Gorb E, Gorb S, Labat J-J, Rowe N (2004) How do plant waxes cause flies to slide? Experimental tests of wax-based trapping mechanisms in three pitfall carnivorous plants. Arth Struct Dev 33:103 Geim AK, Dubonos SV, Grigorieva IV, Novoselov KS, Zhukov AA (2003) Microfabricated adhesive mimicking gecko foot-hair. Nat Mater 2:461 Geiselhardt SF, Geiselhardt S, Peschke K (2009) Comparison of tarsal and cuticular chemistry in the leaf beetle Gastrophysa viridula (Coleoptera: Chrysomelidae) and an evaluation of solidphase microextraction and solvent extraction techniques. Chemoecology 19:185 Geiselhardt SF, Federle W, Prüm B, Geiselhardt S, Lamm S, Peschke K (2010) Impact of chemical manipulation of tarsal liquids on attachment in the Colorado potato beetle, Leptinotarsa decemlineata. J Insect Physiol 56:398 Gladun D, Gorb SN, Frantsevich LI (2009) Alternative tasks of the insect arolium with special reference to hymenoptera. In: Gorb SN (ed) Functional surfaces in biology – adhesion related phenomena, vol 2. Springer, Dordrecht/Heidelberg/London/New York, pp 67–103 Gorb SN (1998) The design of the fly adhesive pad: distal tenent setae are adapted to the delivery of an adhesive secretion. Proc Roy Soc Lond B 265:747 Gorb SN (2000) Biological microtribology: anisotropy in frictional forces of orthopteran attachment pads reflects the ultrastructure of a highly deformable material. Proc Roy Soc Lond B 267:1239

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1673

Gorb SN (2001) Attachment devices of insect cuticle. Springer, New York Gorb SN (2005) Uncovering insect stickiness: structure and properties of hairy attachment devices. Amer Ent 51:31 Gorb SN (2007) Smooth Attachment Devices in Insects: Functional Morphology and Biomechanics. Adv In Insect Phys 34:81 Gorb SN (2009) Adhesion in nature. In: Brockmann W, Geiß PL, Klingen J, Schröder B (eds) Adhesive bonding – materials, applications and technology. Wiley-VCH, Weinheim, pp 346–356 Gorb SN (2010) Biological and biologically inspired attachment systems. In: Bhushan B (ed) Springer handbook of nanotechnology. Springer Verlag, Berlin, pp 1525–1551 Gorb SN (2011) Biological fibrillar adhesives: functional principles and biomimetic applications. In: da Silva LFM, Öchsner A, Adams RD (eds) Handbook of adhesion technology, pp 1409–1436. https://doi.org/10.1007/978-3-642-01169-6_54 Gorb SN, Beutel RG (2001) Evolution of locomotory attachment pads of hexapods. Naturwissenschaften 88:530 Gorb EV, Gorb SN (2002) Attachment ability of the beetle Chrysolina fastuosa on various plant surfaces. Entomol Exp Appl 105:13 Gorb EV, Gorb SN (2006) Do plant waxes make insect attachment structures dirty? Experimental evidence for the contamination hypothesis. In: Herrel A, Speck T, Rowe N (eds) Ecology and biomechanics: a mechanical approach to the ecology of animals and plants. Taylor & Francis, Boca Raton, pp 147–162 Gorb SN, Varenberg M (2007) Mushroom-shaped geometry of contact elements in biological adhesive systems. J Adhes Sci Technol 21:1175 Gorb SN, Varenberg M, Peressadko A, Tuma J (2007a) Biomimetic mushroom-shaped fibrillar adhesive microstructure. J R Soc Interface 4:271 Gorb SN, Sinha M, Peressadko A, Daltorio KA, Quinn RD (2007b) Insects did it first: a micropatterned adhesive tape for robotic applications. Bioinspir Biomim 2:S117 Gorb EV, Hosoda N, Miksch C, Gorb SN (2010) Slippery pores: anti-adhesive effect of nanoporous substrates on the beetle attachment system. J R Soc Interface 7:1571 Gottlieb Binder GmbH & Co KG (2017) http://www.binder.de/en/products/geckonanoplast/ Greiner C, Arzt E, del Campo A (2009) Hierarchical Gecko - Like Adhesives. Adv Mater 21:479 Heepe L, Gorb SN (2014) Biologically inspired mushroom-shaped adhesive microstructures. Annu Rev Mater Res 44:173 Heepe L, Varenberg M, Itovich Y, Gorb SN (2011) Suction component in adhesion of mushroomshaped microstructure. J R Soc Interface 8:585 Heepe L, Kovalev AE, Varenberg M, Tuma J, Gorb SN (2012) First mushroom-shaped adhesive microstructure: A review. Thero Appl Mech Lett 2:014008 Heepe L, Kovalev AE, Filippov AE, Gorb SN (2013) Adhesion failure at 180 000 frames per second: direct observation of the detachment process of a mushroom-shaped adhesive. Phys Rev Lett 111:104301 Heepe L, Carbone G, Pierro E, Kovalev AE, Gorb SN (2014a) Adhesion tilt-tolerance in bioinspired mushroom-shaped adhesive microstructure. Appl Phys Lett 104:011906 Heepe L, Kovalev AE, Gorb SN (2014b) Direct observation of microcavitation in underwater adhesion of mushroom-shaped adhesive microstructure. Beilstein J Nanotechnol 5:903 Heepe L, Wolff JO, Gorb SN (2016) Influence of ambient humidity on the attachment ability of ladybird beetles (Coccinella septempunctata). Beilstein J Nanotechnol 7:1332 Heepe L, Raguseo S, Gorb SN (2017a) An experimental study of double-peeling mechanism inspired by biological adhesive systems. Appl Phys A Mater Sci Process 123:124 Heepe L, Petersen DS, Tölle L, Wolff JO, Gorb SN (2017b) Sexual dimorphism in the attachment ability of the ladybird beetle Coccinella septempunctata on soft substrates. Appl Phys A Mater Sci Process 123:34 Hiller U (1968) Untersuchungen zum Feinbau und zur Funktion der Haftborsten von Reptilien. Z Morphol Tiere 62:307

1674

S. N. Gorb and L. Heepe

Homann H (1957) Haften Spinnen an einer Wasserhaut? Naturwissenschaften 44:318 Huber G, Gorb SN, Spolenak R, Arzt E (2005a) Resolving the nanoscale adhesion of individual gecko spatulae by atomic force microscopy. Biol Lett 1:2 Huber G, Mantz H, Spolenak R, Mecke K, Jacobs K, Gorb SN, Arzt E (2005b) Evidence for capillarity contributions to gecko adhesion from single spatula nanomechanical measurements. Proc Natl Acad Sci U S A 102:16293 Hui CY, Glassmaker NJ, Tang T, Jagota A (2004) Design of biomimetic fibrillar interfaces: 2. Mechanics of enhanced adhesion. J R Soc Interface 1:35 Ishii S (1987) Adhesion of a Leaf Feeding Ladybird Epilachna vigintioctomaculta (Coleoptera: Coccinellidae) on a Virtically Smooth Surface. Appl Entomol Zool 22:222 Israelachvili JN (1992) Intermolecular and surface forces: With Applications to Colloidal and Biological Systems, 2nd edn. Academic, London Jagota A, Bennison SJ (2002) Mechanics of adhesion through a fibrillar microstructure. Integr Comp Biol 42:1140 Jagota A, Hui C-Y (2011) Adhesion, friction, and compliance of bio-mimetic and bio-inspired structured interfaces. Mater Sci Eng R Rep 72:253 Johnson KL, Kendall K, Roberts AD (1971) Surface energy and the contact of elastic solids. Proc R Soc Lond A 324:301 Kampermann M, Kroner E, del Campo A, McMeeking RM, Arzt E (2010) Functional Adhesive Surfaces with “Gecko” Effect: The Concept of Contact Splitting. Adv Eng Mater 12:335 Kasem H, Varenberg M (2013) Effect of counterface roughness on adhesion of mushroom-shaped microstructure. J R Soc Interface 10:20130620 Kendall K (1975) Thin-film peeling-the elastic term. J Phys D Appl Phys 8:1449 Kesel AB, Martin A, Seidl T (2003) Adhesion measurements on the attachment devices of the jumping spider Evarcha arcuata. J Exp Biol 206:2733 Kim TW, Bhushan B (2007) Adhesion analysis of multi-level hierarchical attachment system contacting with a rough surface. J Adhes Sci Technol 21:1 Kim S, Sitti M (2006) Biologically inspired polymer microfibers with spatulate tips as repeatable fibrillar adhesives. Appl Phys Lett 89:26911 Kizilkan E, Heepe L, Gorb SN (2013) Underwater adhesion of mushroom-shaped adhesive microstructure: an air-entrapment effect. In: Biological and Biomimetic Adhesives: Challenges and Opportunities. RCS, Cambridge, pp 65–71 Kizilkan E, Strueben J, Staubitz A, Gorb SN (2017) Bioinspired photocontrollable microstructured transport device. Sci Robotics 2:eaak9454 Kosaki A, Yamaoka R (1996) Chemical composition of footprints and cuticula lipids of three species of lady beetles. Jpn J Appl Entomol Zool 40:47 Kovalev AE, Varenberg M, Gorb SN (2012) Wet versus dry adhesion of biomimetic mushroomshaped microstructures. Soft Matter 8:7560 Kwak MK, Pang C, Jeong HE, Kim HN, Yoon H, Jung HS, Suh KY (2011) Towards the next level of bioinspired dry adhesives: new designs and applications. Adv Funct Mater 21:3606 Langer MG, Ruppersberg JP, Gorb SN (2004) Adhesion forces measured at the level of a terminal plate of the fly’s seta. Proc R Soc Lond B 271:2209 Murphy MP, Aksak B, Sitti M (2007) Adhesion and anisotropic friction enhancements of angled heterogeneous micro-fiber arrays with spherical and spatula tips. J Adhes Sci Tech 21:1281 Niederegger S, Gorb SN (2003) Tarsal movements in flies during leg attachment and detachment on a smooth substrate. J Insect Physiol 49:611 Niederegger S, Gorb SN (2006) Friction and adhesion in the tarsal and metatarsal scopulae of spiders. J Comp Physiol A 192:1223 Niederegger S, Gorb SN, Vötsch W (2001) Fly walking: a compromise between attachment and motion? In: Wisser A, Nachtigall W (eds) Technische Biologie und Bionik. 5. Bionik – Kongress, Dessau 2000. Gustav Fisher Verlag, Stuttgart/Jena/Lübeck/Ulm, pp 327–330 Niewiarowski PH, Lopez S, Ge L, Hagan E, Dhinojwala A (2008) Sticky gecko feet: the role of temperature and humidity. PLoS One 3:e2192

55

Biological Fibrillar Adhesives: Functional Principles and Biomimetic. . .

1675

Northen MT, Turner KL (2005) A batch fabricated biomimetic dry adhesive. Nanotechnology 16:1159 Peattie AM, Full RJ (2007) Phylogenetic analysis of the scaling of wet and dry biological fibrillar adhesives. Proc Natl Acad Sci U S A 104:18595 Peisker H, Gorb SN (2012) Evaporation dynamics of tarsal liquid footprints in flies (Calliphora vicina) and beetles (Coccinella septempunctata). J Exp Biol 215:1266 Peisker H, Michels J, Gorb SN (2013) Evidence for a material gradient in the adhesive tarsal setae of the ladybird beetle Coccinella septempunctata. Nat Commun 4:1661 Peisker H, Heepe L, Kovalev AE, Gorb SN (2014) Comparative study of the fluid viscosity in tarsal hairy attachment systems of flies and beetles. J R Soc Interface 11:20140752 Pelletier Y, Smilowitz Z (1987) Specialized tarsal hairs on adult male Colorado potato beetles, Leptinotarsa decemlineata (Say), hamper its locomotion on smooth surfaces. Can Entomol 119:1139 Peressadko A, Gorb SN (2004a) When less is more: experimental evidence for tenacity enhancement by division of contact area. J Adhes 80:247 Peressadko A, Gorb SN (2004b) Surface profile and friction force generated by insects. In: Fortschritt-Berichte VDI, Boblan I, Bannasch R (eds) Surface profile and friction force generated by insects, vol 249[15]. VDI Verlag, Düsseldorf, pp 257–263 Persson BNJ (2003) On the mechanism of adhesion in biological systems. J Chem Phys 118:7614 Persson BNJ (2014) On the fractal dimension of rough surfaces. Tribol Lett 54:99 Persson BNJ, Gorb SN (2003) The effect of surface roughness on the adhesion of elastic plates with application to biological systems. J Chem Phys 119:11437 Popov VL (2010) Contact mechanics and friction: physical principles and applications. SpringerVerlag, Berlin Prowse MS, Wilkinson M, Puthoff JB, Mayer G, Autumn K (2011) Effects of humidity on the mechanical properties of gecko setae. Acta Biomater 7:733 Pugno NM (2011) The theory of multiple peeling. Int J Fract 171:185 Pugno NM, Gorb SN (2009) Functional mechanism of biological adhesive systems described by multiple peeling approach. In: Proceedings of the 12th international conference on fracture, July 1217, Ottawa Puthoff JB, Prowse MS, Wilkinson M, Autumn K (2010) Changes in materials properties explain the effects of humidity on gecko adhesion. J Exp Biol 213:3699 Richards AG, Richards PA (1979) The cuticular protuberances of insects. Int J Insect Morphol Embryol 8:143 Rizzo NW, Gardner KH, Walls D, Keiper-Hrynko JNM, Ganzke TS, Hallahan DL (2006) Characterization of the structure and composition of gecko adhesive setae. J R Soc Interface 3:441 Röll B (1995) Epidermal fine structure of the toe tips of Sphaerodactylus cinereus (Reptilia, Gekkonidae). J Zool 235:289 Ruibal R, Ernst V (1965) The structure of the digital setae of lizards. J Morphol 117:271 Russell AP (1975) A contribution to the functional analysis of the foot of the Tokay, Gekko gecko (Reptilia: Gekkonidae). J Zool (Lond) 176:437 Sameoto D, Menon C (2010) Recent advances in the fabrication and adhesion testing of biomimetic dry adhesives. Smart Mater Struct 19:103001 Schargott M (2009) A mechanical model of biomimetic adhesive pads with tilted and hierarchical structures. Bioinspir Biomim 4(026002):9 Scherge M, Gorb SN (2001) Biological micro- and nanotribology: nature’s solutions. Springer, Berlin Schleich HH, Kastle W (1986) Ultrastrukturen an Gecko-Zehen (reptilia: sauria: gekkonidae). Amphibia-Reptilia 7:141 Sitti M, Fearing RS (2003) Synthetic gecko foot-hair micro/nano-structures as dry adhesives. J Adhes Sci Technol 17:1055 Smith JM, Barnes WJP, Downie JR, Ruxton GD (2006) Structural correlates of increased adhesive efficiency with adult size in the toe pads of hylid tree frogs. J Comp Physiol A 192:1193

1676

S. N. Gorb and L. Heepe

Spolenak R, Gorb SN, Gao H, Arzt E (2005) Effects of contact shape on the scaling of biological attachments. Proc R Soc Lond A 461:305 Stork NE (1980a) Experimental analysis of adhesion of Chrysolina polita (Chrysomelidae: Coleoptera) on a variety of surfaces. J Exp Biol 88:91 Stork NE (1980b) A scanning electron microscope study of tarsal adhesive setae in the Coleoptera. Zool J Linnean Soc 68:173 Stork NE (1983) A comparison of the adhesive setae on the feet of lizards and arthropods. J Nat Hist 17:829 Tang T, Hui CY (2005) Can a fibrillar interface be stronger and tougher than a non-fibrillar one? J R Soc Interface 2:505 Varenberg M, Gorb SN (2008a) A beetle-inspired solution for underwater adhesion. J R Soc Interface 5:383 Varenberg M, Gorb SN (2008b) Close-up of mushroom-shaped fibrillar adhesive microstructure: contact element behaviour. J R Soc Interface 5:785 Varenberg M, Gorb SN (2008c) Shearing of fibrillar adhesive microstructure: friction and shearrelated changes in pull-off force. J R Soc Interface 4:721 Varenberg M, Pugno NM, Gorb SN (2010) Spatulate structures in biological fibrillar adhesion. Soft Matter 6:3269 Varenberg M, Murarash B, Kligermann Y, Gorb SN (2011) Geometry-controlled adhesion: revisiting the contact splitting hypothesis. Appl Phys A Mater Sci Process 103:933 Voigt D, Schuppert JM, Dattinger S, Gorb SN (2008) Sexual dimorphism in the attachment ability of the Colorado potato beetle Leptinotarsa decemlineata (Coleoptera: Chrysomelidae) to rough substrates. J Insect Physiol 54:765 Vötsch W, Nicholson G, Müller R, Stierhof Y-D, Gorb SN, Schwarz U (2002) Chemical composition of the attachment pad secretion of the locust Locusta migratoria. Insect Biochem Mol Biol 32:1605 Walker G, Yulf AB, Ratcliffe J (1985) The adhesive organ of the blowfly, Calliphora vomitoria: a functional approach (Diptera: Calliphoridae). J Zool (Lond) 205:297 Wigglesworth VB (1987) How does a fly cling to the under surface of a glass sheet? J Exp Biol 129:373 Wolff JO, Gorb SN (2011) The influence of humidity on the attachment ability of the spider Philodromus dispar (Araneae, Philodromidae). Proc R Soc London, Ser B 279:139 Wolff JO, Gorb SN (2012) Surface roughness effects on attachment ability of the spider Philodromus dispar (Araneae, Philodromidae). J Exp Biol 215:179 Wolff JO, Gorb SN (2016) Attachment structures and adhesive secretions in arachnids. Springer, Berlin Yurdumakan B, Raravikar NR, Ajayan PM, Dhinojwala A (2005) Synthetic gecko foot-hairs from multiwalled carbon nanotubes. Chem Commun 16041421:3799

Adhesives with Nanoparticles

56

Ambrose C. Taylor

Contents 56.1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.1.1 Nanotechnology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.1.2 Nanoparticle Reinforcement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2 Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.1 Nanoparticles Definition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.2 Equi-Axed Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.3 Nanotubes and Nanofibers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.4 Plate-like Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.2.5 Other Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3 Manufacturing Using Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.1 Dispersion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.2 Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.3 Sonication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.3.4 Alignment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4 Properties of Nanoparticle-Modified Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.1 Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.2 Functional Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.3 Electrical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.4 Fracture Toughness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.5 Fatigue Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.4.6 Peel and Lap-Shear Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1678 1678 1679 1680 1680 1680 1681 1683 1683 1684 1684 1685 1686 1686 1687 1687 1688 1690 1691 1694 1696 1696 1697

A. C. Taylor (*) Department of Mechanical Engineering, Imperial College London, London, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_55

1677

1678

A. C. Taylor

Abstract

The increased commercial availability and the reduced prices of nanoparticles are leading to their incorporation in polymers and structural adhesives. This chapter outlines the principal types of nanoparticles, and the methods that may be used to disperse the particles in a polymer matrix. It discusses how nanoparticles can alter the mechanical properties (e.g., stiffness), electrical properties (e.g., conductivity), functional properties (e.g., permeability, glass transition temperature), and fracture performance of thermoset polymers. The effect of nanoparticles on joint performance is also discussed.

56.1

Introduction

56.1.1 Nanotechnology It can be argued that nanotechnology is not new, as the use of nanoparticles to alter the properties of materials is not a recent development. For example, the late Roman (fourth century AD) Lycurgus cup is on display in London at the British Museum. This cup is made of cut glass, and looks green in reflected light, but appears red when light is shone through it. This effect is due to the light being scattered by the colloidal dispersion of gold and silver nanoparticles, which are about 70 nm in diameter, in the glass. Another example is carbon black, as used currently in millions of car tires and printer cartridges per year, which is composed of particles of size 20 nm and above. However, this material was first produced 3,500 years ago in China as lamp black. Nanotechnology has become a hot topic in science and engineering in recent years. One of the definitive points in the history of nanotechnology came in a lecture by Richard Feynman, in 1959, entitled “There Is Plenty of Room at the Bottom” (Feynman 1959). In this lecture, Feynman highlighted the potential of working at a micro- or nanoscale, and discussed the problem of manipulating and controlling things on a small scale. He issued a challenge to scientists (and engineers) to work at very small scales, and this challenge is now being taken up. So what is nanotechnology? Nanotechnology could be defined as the combination of existing technologies with the ability to observe and manipulate at the nanometer (10 9 m) scale (Hay and Shaw 2001; Harper 2003). The question is then, why is there so much interest in nanotechnology now? It is linked to (a) the expanding ability to synthesize nanometer-scale materials, and (b) the availability of the tools that enable the imaging of structures at this scale. Thus, it is now possible to characterize how nanoparticles are dispersed and start to understand why the observed effects occur, something that was not possible for the Romans. As the limits of the nanoscale, say at scales of less than 20 nm, are approached then different effects, such as quantum effects, become more significant. A major concern with the use of nanoparticles is their potential toxicity. Care must be taken not to repeat the health problems caused by exposure to asbestos. Stories of potential catastrophe such as Michael Crichton’s book “Prey” and Prince Charles’ “grey goo” speech have not helped the perception of nanoparticles and nanotechnology

56

Adhesives with Nanoparticles

1679

(Radford 2003). Hence, manufacturers are approaching this new technology cautiously. Studies are underway to assess the health effects and some results have been published (e.g., Lam et al. 2004; Warheit et al. 2004; Brown et al. 2007), but as yet there is no real agreement on the real dangers to health due to inhaling nanoparticles. It must be borne in mind that the nanoparticles in cured adhesives are not free in the atmosphere, but are bound into a matrix. Recent studies indicate that even during fracture of a nanoparticle-modified epoxy material, only a few nanoparticles are released into the atmosphere. In contrast, nanoparticles are produced in large quantities by the combustion of fossil fuels, and millions of tonnes of nano- and submicronparticles are already in use by various industries including in paints and toiletries.

56.1.2 Nanoparticle Reinforcement Five basic characteristics of particulate fillers were defined by Rothon and Hancock (1995). These remain true whether the particles are micro- or nanoparticles: 1. 2. 3. 4. 5.

What properties are being sought? What deleterious changes may also occur and can they be tolerated? How easy is the filler to handle and how might it affect processing? Are any special additives needed? What is the true cost of using the filler, is it justifiable and are there more costeffective alternatives?

Rothon and Hancock (1995) observed that it is widely assumed that fillers are cheap and that polymers are expensive. Conversely, for nanoparticles it is widely assumed that nanoparticles are expensive and that polymers are cheap. This is not necessarily the case. As the nanoparticle manufacturing industry expands, nanoparticles are increasingly available in large (i.e., tonne) quantities. The prices of expensive nanoparticles such as carbon nanotubes are also being reduced. The price of nanoparticles varies greatly with the type, as well as with the purity of the material. For example, silica nanoparticles supplied dispersed as a masterbatch in epoxy cost about $20/kg (Nanoresins 2008), and core-shell rubber nanoparticles similarly dispersed cost approximately $12/kg (Kaneka 2008). Nanoclays can cost as little as $7/kg (SigmaAldrich 2008). However, a kilogram of carbon nanotubes cost between $400 and $98,000 in Autumn 2009 (CheapTubes 2009). Nanoparticles are, by definition, very small. It must be noted that the interparticle distance for nanoparticle-modified polymers is also very small, especially at high volume fractions. These factors may affect the interaction between a particle and its neighbors, or between a particle and the polymer surrounding it, or may affect the structure of the polymer itself. Hence, the properties of a particle-modified polymer may vary depending whether nanoparticles or microparticles are used. The addition of nanoparticles can be simply a marketing tool, as the performance benefit that they provide could often be obtained using conventional fillers at a much lower cost. However, some very successful applications that rely on the small size

1680

A. C. Taylor

of nanoparticles have been shown, for example, the addition of ceramic nanoparticles to transparent coatings gives improved scratch resistance without loss of transparency. Nanoparticles have also been used in sporting goods, for example, carbon nanotubes in tennis racket or bicycle frames, and ski poles that use a glass fiber composite with silica nanoparticles.

56.2

Nanoparticles

56.2.1 Nanoparticles Definition The common definition of a nanoparticle is that it has at least one dimension in the nanometer range. Indeed, the term “nano” has virtually replaced “submicron,” even when the latter is more appropriate. A search of the scientific literature published in the last 5 years yields more than 60,000 articles per year that mention “nano,” but less than 3,000 papers per year that mention “submicro” (of which almost a third mention “nano” as well). Nanoparticles are characterized by a large surface area to volume ratio. They can be made from a wide range of materials – metallic (e.g., gold, silver), ceramic (e.g., silica, alumina, layered silicates, silicon carbide), or organic (e.g., carbon black, rubber particles, graphene, carbon nanotubes, and nanofibers). Due to the small size of nanoparticles, the numbers present in nanoparticle-modified materials are enormous. For example, there are approximately 1017 individual particles in 1 kg of 20 nm silica nanoparticles. As many nanoparticles are smaller than the wavelength of light (about 400–700 nm), they appear translucent when added to a translucent polymer. A consequence of this is that they are too small to be seen using optical microscopy, but can be imaged by high-magnification techniques, such as transmission electron microscopy (TEM), atomic force microscopy (AFM), and field emission gun scanning electron microscopy (FEGSEM). The structure of nanoparticles can also be characterized using X-ray diffraction or other techniques (Wang 2001). Because the particles are so small, it is not normally possible to measure their mechanical properties, and hence these are assumed to be the same as for the bulk material, or equal to those of a similar material. For many applications, such as the prediction of the mechanical properties of nanoparticle-modified materials, this appears to be a reasonable assumption. As with micron-sized fillers, the particle shape and aspect ratio are important in determining the properties of nanoparticle-modified materials, such as stiffness, flow characteristics, tensile strength, etc. The aggregation and dispersion of the particles are also important, and may also affect the properties. Nanoparticles can be classified by their shape, as equi-axed (e.g., spherical), rod-like, and plate-like. These classes will be discussed in turn.

56.2.2 Equi-Axed Nanoparticles Most equi-axed nanoparticles are spherical, but irregular particles are also available. The common particles are metallic (e.g., gold, silver), ceramic (e.g., silica, alumina,

56

Adhesives with Nanoparticles

1681

titania), or organic (e.g., rubber particles, fullerenes). The ceramic particles are typically prepared by sol-gel or flame-spraying methods; silica and alumina are commonly used with adhesives, see Fig. 1. Rubber particles are typically coreshell particles, with a soft core and a hard shell of a polymer such as poly(methyl methacrylate).

56.2.3 Nanotubes and Nanofibers The rod-like nanoparticles that are commonly discussed in the scientific literature are carbon nanotubes. Carbon nanofibers (CNFs), which have a larger diameter, are also available. They are cheaper and are available in much larger quantities than carbon nanotubes. Ceramic nanotubes (e.g., zirconia, tungsten disulfide) or whiskers (e.g., silicon nitride, silicon carbide, alumina) can also be used. Carbon nanotubes were first reported by Iijima (1991; Iijima and Ichihashi 1993), and are effectively sheets of graphite rolled into tubes. They are produced by chemical vapor deposition (CVD), electric arc, or laser ablation processes (Harris 1999, 2009). These techniques produce nanotubes with different lengths, straightness, purities, and degrees of entanglement. Hence, the nanotubes require purification to remove the amorphous carbon that is also produced, typically by heating them in a mixture of nitric and sulfuric acid (Gojny et al. 2003). CVD is becoming the most popular method for producing relatively large quantities of nanotubes relatively cheaply. In this technique, the nanotubes are grown on a ceramic substrate,

Fig. 1 Transmission electron micrograph of silica nanoparticles in epoxy (Courtesy S. Sprenger)

1682

A. C. Taylor

rather like the bristles of a brush. The length of the nanotubes can be controlled, and they have a relatively low degree of entanglement. There are two principal types of carbon nanotubes. Firstly, single-walled nanotubes (SWNTs) have a single graphene layer rolled into a tube typically 1–2 nm in diameter, with hemispherical end caps (Iijima and Ichihashi 1993; Tjong 2006). Secondly, multiwalled nanotubes (MWNTs) are comprised of a number of coaxial graphene tubes, each with an end cap. The outer diameter of a MWNT is typically 3–10 nm. Under tension, only the outer layer of a MWNT carries the load as the van der Waals forces between the layers are too weak to transfer stress to the inner layers (Lau et al. 2004). Variants such as double-walled nanotubes (DWNTs) are also available. The length of nanotubes varies significantly, but they can be as long as several millimeters (Chakrabarti et al. 2006). Hence, nanotubes can possess very large aspect ratios. Their modulus has been measured to be up to 1 TPa, with strengths of up to 150 MPa (Demczyk et al. 2002), although they are rather flexible normal to their axis (Saito et al. 1998). However, it is difficult to adhere the tubes to a matrix, and hence surface treatments are used to activate the surface, which can weaken the tubes. The electrical properties of nanotubes vary from metallic to semiconducting depending on their chirality. Vapor-grown carbon nanofibers do not have such a high aspect ratio as carbon nanotubes, and have more defects. However, they are significantly cheaper and available in much larger quantities. The nanofibers are typically 50–200 nm in diameter and 30–100 μm long (Green et al. 2006; Zhou et al. 2007), as shown in Fig. 2. Ceramic whiskers have been produced for many years, as reported by Gordon (1978), though their use in epoxies seems to be limited. However, aluminum borate whiskers have been combined with thermoset polymer matrices by Liang and coworkers (Liang and Hu 2004; Tang et al. 2007). Ceramic nanotubes have been produced by coating carbon nanotubes with silicon carbide (Morisada et al. 2007). Silicon carbide nanofibers, which can be about 40 nm in diameter with lengths of up to several hundred microns, have also been produced (Zhu et al. 2002; Bechelany et al. 2007). Fig. 2 Scanning electron micrograph of carbon nanofibers in an epoxy matrix, showing fractured surface (Courtesy J. H. Lee)

56

Adhesives with Nanoparticles

1683

56.2.4 Plate-like Nanoparticles A range of plate-like nanoparticles have been used with polymer matrices, including layered silicate nanoclays (e.g., montmorillonite), graphite, and α-zirconium phosphate. Generally, these particles are comprised of stacks of platelets that are intercalated or exfoliated by the polymer during processing. The resulting morphology is typically described as “particulate” (or “conventional”), “intercalated” or “exfoliated,” as identified by wide-angle X-ray scattering (WAXS). These microstructures are shown schematically in Fig. 3 for a nanoclay-modified polymer. In an intercalated microstructure, polymer chains enter the galleries between the platelets and increase the measured spacing (Schadler 2003). For an exfoliated nanocomposite, the platelets are pushed further apart, and the spacing becomes too large to measure using WAXS. Note that an exfoliated structure may be described as “ordered’ or “disordered,” as shown in Fig. 3, but the fully disordered exfoliated structure is relatively rare. For a particulate structure, the particles remain unchanged. Nanoclays (e.g., montmorillonite, hectorite) are layered silicates with a structure of stacked platelets similar to that of mica. The platelet thickness is approximately 1 nm and the lateral dimensions of the plates can vary from 30 nm to tens of microns. Hence, the platelets have a high aspect ratio. The layers have a regular van der Waals gap in-between them, called the “interlayer” or the “gallery” (Alexandre and Dubois 2000). These nanoclays are hydrophilic, and hence are typically surface-treated using long chain alkylammonum compounds to make them organophilic. The stacks of platelets can be readily delaminated by organic molecules to form an intercalated or exfoliated structure. Natural graphite has also been used. A combination of chemical and thermal treatments can generate exfoliated graphite platelets with a thickness of 20–100 nm. Here, the graphite is intercalated by an acid treatment, followed by exfoliation by a thermal shock at a temperature of around 600  C. The platelet diameter is in the order of 10 μm, and hence these platelets have a high aspect ratio (Yasmin and Daniel 2004; Yasmin et al. 2006b).

56.2.5 Other Nanoparticles Carbon black is mostly an aciniform (grape-like cluster) particulate, which is produced by the incomplete combustion of oil. The current worldwide production

Epoxy

Silicate Particulate

Intercalated

Fig. 3 Microstructures of nanoclay-modified polymers

Exfoliated (ordered)

Exfoliated (disordered)

1684

A. C. Taylor

is 8.1 million tonnes per year. It is very common as a pigment in inks and toners, and as a functional filler in tires and rubber products (ICBA 2008).

56.3

Manufacturing Using Nanoparticles

56.3.1 Dispersion To obtain the best properties from a nanoparticle-modified polymer, it is considered that the nanoparticles should be well dispersed, and each particle should be properly wetted by the polymer. If agglomerates are formed instead, these can act as defects resulting in a reduction in performance rather than any enhancement. This is a major challenge when preparing adhesive formulations. Dispersing nanoparticles is difficult due to their high surface area and incompatability with the matrix polymer. Generally, a surface treatment or compatabilizer is required, and once agglomeration occurs it is very difficult to break up the agglomerates. Hence, particles that are supplied pre-dispersed in a resin (e.g., Nanopox from Nanoresins, or the MX range of core-shell particles from Kaneka) are popular with formulators. However, there is no guarantee that this good initial dispersion will remain after subsequent processing. For example, the addition of a liquid rubber, which phase-separates during curing, to some systems can cause the nanoparticles to agglomerate (Mohammed 2007; Hsieh et al. 2010). Although it is normally assumed that good dispersion is required, small agglomerates do not necessarily adversely affect the performance of such systems. Assessing the quality of the dispersion of nanoparticles is difficult, as the techniques that can identify particles at the nanoscale such as AFM and TEM involve lengthy specimen preparation. Only a very small volume of the sample can be studied using these techniques, and hence the results may not be representative of the material as a whole. Alternatively, X-ray diffraction can be used, as this technique can analyze a larger volume. However, the data do not characterize how well the particles are dispersed. It is normally suggested that dispersion should be assessed at three size scales – (a) optically, (b) by scanning electron microscopy, and (c) by transmission electron microscopy. If the results from all three techniques indicate a good dispersion of nanoparticles, then it is fairly safe to assume that this is the case. Many authors agree that using optical microscopy is useful to investigate the dispersion of nanoparticles as it highlights the presence of agglomerates (e.g., Hackman and Hollaway 2006; Brooker et al. 2008), although individual nanoparticles will not be visible. Quantitative methods to assess dispersion, using the gray scale of images or by a quadrat method, are being developed (Brooker et al. 2010). It should be noted that even small volume fractions of nanoparticles involve huge numbers of nanoparticles and very small interparticle distances. Shaffer and Kinloch (2004) point out that one major difficulty with small diameter nanotubes is that they become increasingly difficult to wet. By trivial estimation, even a 1 vol% loading of single-walled nanotubes ensures that all of the polymer molecules are within one

56

Adhesives with Nanoparticles

1685

radius of gyration (say 5 nm) of a nanotube. This result implies that complete wetting of high loading fractions of single-walled nanotubes will be difficult, at least by conventional means, and that even more modest concentrations may be brittle and hard to process due to the constraint of the matrix (Shaffer and Kinloch 2004). The chemical functionalization of the nanoparticles’ surface can help improve dispersion, as it can reduce agglomeration, and can also improve the bonding between the particle and the matrix (Gojny et al. 2003; Kathi and Rhee 2008).

56.3.2 Mixing If nanoparticles are supplied well dispersed in a polymer, then simple mixing is often enough to blend the particles into an adhesive formulation (e.g., Kinloch et al. 2005). Low-shear mixing is also sufficient for well-compatabilized systems, as the thermodynamics of these materials can ensure a good dispersion. However, if nanoparticles are supplied as an untreated powder they will generally be agglomerated, and these agglomerates must be broken down during processing. Low-shear mixing alone will not break up these agglomerates. High-power dispersion methods, such as highshear mixing and sonication, are the easiest methods to improve the dispersion of nanoparticles in polymers (Xie et al. 2005). In a multicomponent system such as a structural adhesive, the polymer component into which the nanoparticles are introduced does not generally cause a noticeable difference in the final morphology (Hackman and Hollaway 2006; Brooker et al. 2008). However, the morphology may be dependent on the cure time and temperature; plus the temperature, time, and shear rate used for mixing. For carbon nanotubes, Fiedler et al. (2006) report that the size and shape of the impeller and the mixing speed control the dispersion. Intensive stirring of MWNTs in epoxy resin can achieve a relatively good dispersion. However, reagglomeration occurs due to frictional contacts and elastic interlocking mechanisms (Schmid and Klingenberg 2000). SWNTs have a greater tendency to reagglomerate than MWNTs, and hence higher shear forces are required to achieve a reasonable dispersion. High-shear mixing can be achieved using a range of techniques. Hackman and Hollaway (2006) compared low-shear mixing to the use of a grinding media mill, filled with 2 mm diameter glass beads. Green et al. (2006) compared low-shear mixing to high-shear mixing generated by extrusion of a nanoparticle-modified epoxy resin through a small orifice. In both cases, the high-shear method gave better dispersion. Glass beads have been shown to break down agglomerated nanoclay particles to form a uniform material. However, although neither work evaluated the dispersion at the nanometer-scale using microscopy, differences could be seen at the microscale (Hackman and Hollaway 2006). A three-roll mill (or calender) has been used by Yasmin et al. to disperse nanoclay in epoxy (Yasmin et al. 2003; Yasmin et al. 2006a), and by Gojny et al. to disperse nanotubes in epoxy (Gojny et al. 2004). This method is commonly used to disperse microparticles in polymers (Fiedler et al. 2006). As it is not a continuous process, the material was collected and fed back into the mill after each pass. The authors

1686

A. C. Taylor

assumed that the dispersion was achieved (a) by the shear forces generated between the dispersing of the particles as smaller tactoids, and (b) by the combined shear and diffusion processes facilitating the separation and penetration of polymer between the clay platelets (Yasmin et al. 2003; Yasmin et al. 2006a). This process requires a relatively viscous material, and hence is suitable for higher loadings of nanoparticles. According to Fiedler et al. (2006), a major advantage of this method is that it allows the efficient manufacturing of larger quantities of nanocomposites, as the nanoparticle-modified polymer requires only minutes of processing. By comparison, low-shear methods may require processing times of hours.

56.3.3 Sonication Sonication uses an ultrasonic bath or probe to apply sound energy to a liquid that contains particles. Many authors combine sonication and mixing, especially when using probe sonicators as the power is directed into only a small volume, and the low-shear forces do not generally ensure that all of the polymer passes through this volume. The combination of probe sonication with high-speed mixing has been reported to be a relatively successful way to disperse nanoparticles. This technique has been used with nanoclays in epoxy resin (Zunjarrao et al. 2006) and in polyimide (Gintert et al. 2007). It has also been used with carbon nanotubes in epoxy (Sandler et al. 1999). The very localized introduction of energy from probe sonication leads to a considerable amount of damage, including buckling, bending, dislocations, and rupture of carbon nanotubes (Lu et al. 1996; Fiedler et al. 2006). It will also cause local heating of the sample, and hence if sonication is undertaken with a curing agent or initiator present it may be necessary to cool the sample to prevent premature curing (Lu et al. 1996). The use of an ultrasonic bath reduces the energy density and alleviates some of these problems, but the time taken to achieve a reasonable dispersion is much greater (Brooker et al. 2008). The processing time can affect the size of nanoclay agglomerates (Lam et al. 2005). However, the d-spacing (the distance between the nanoclay platelets) was unaffected by the sonication time, and hence exfoliation of the nanoclay could not be achieved using sonication. Gintert et al. (2007) also found that sonication was an important step in breaking up nanoclay agglomerates.

56.3.4 Alignment The properties of composites are generally improved by the alignment of the fibers, although this does produce an anisotropic material. It is also possible to align nanoparticles. This is generally achieved by physical means, for example, by application of a force, cutting, extrusion, or drawing (e.g., Ajayan et al. 1994; Baik et al. 2005). Alignment can also be achieved using the application of magnetic or electric fields, with carbon nanotubes (e.g., Chen et al. 2001; Martin et al. 2005),

56

Adhesives with Nanoparticles

1687

and with layered silicates (e.g., Koerner et al. 2005). For more information on the alignment of carbon nanotubes, and their dispersion, see the review by Xie et al. (2005).

56.4

Properties of Nanoparticle-Modified Polymers

56.4.1 Mechanical Properties The addition of high modulus particles increases the modulus of a polymeric material, independent of the particle size; although the stiffening effect may be higher for particles with larger aspect ratios. There are many theoretical models available to predict the modulus of particle-modified polymers, such as the HalpinTsai and Mori-Tanaka models (e.g., Ahmed and Jones 1990). These models have been shown to broadly apply for nanoparticle-modified polymers (e.g., Kinloch and Taylor 2006), see Fig. 4, though more accurate predictions can be obtained by finite element modeling (Sheng et al. 2004). The true aspect ratio of the particles, the degree of their alignment, and some idea of the strength of adhesion between the particle and the polymer matrix, are required for accurate predictions. However, the elastic properties of nanoparticles cannot generally be measured, and hence are almost unknown. This includes nanoclays such as montmorillonite (Vanorio et al. 2003). This absence of modulus data is because the small particle size makes it impossible to perform reliable measurements. Hence, most authors assume that the 3.0

Halpin-Tsai parallel

Relative modulus (Ec/Em)

2.5

Mori-Tanaka parallel Mori-Tanaka random

2.0 Halpin-Tsai random

1.5 Modified rule of mixtures

1.0

0.5

0.0 0

2

4

6 8 10 Colume of silicate (%)

12

14

16

Fig. 4 Relative modulus (composite modulus divided by matrix modulus) of epoxy polymer modified with nanoclay. Experimental data are shown as points, predictions using various theoretical models are shown as lines (Adapted from Kinloch and Taylor 2006)

1688

A. C. Taylor

properties (e.g., modulus, density) of the nanoparticles are equal to those of the bulk material, or of a similar material. Examples of the stiffening effect that may be obtained using nanoparticles include the addition of 3% carbon nanofibers (CNF) to an amine-cured epoxy polymer (Zhou et al. 2007), which gave a 19% increase in modulus when added. This was accompanied by an increase in the tensile strength and a reduction in the strain to failure. These changes are typical of nanoparticle-modified epoxies. Generally, the stiffness is increased; the strain to failure and the tensile strength may be increased or remain the same. However, when small diameter particles are used, for example, 20 nm diameter silica nanoparticles, then the tensile strength remains unchanged. By comparison, the addition of hard microparticles tends to reduce both the strain to failure and the tensile strength.

56.4.2 Functional Properties The high aspect ratio and low permeability of plate-like nanoparticles should provide good barrier properties when the particles are aligned and incorporated into thin films, by creating a tortuous path for diffusion (Hackman and Hollaway 2006; Sorrentino et al. 2006). Numerical predictions show that the permeability reduction is governed by the product a
Vf, where a is the platelet aspect ratio and Vf is the volume fraction (Gusev and Lusti 2001). Epoxy films that were highly filled with nanoclay have been shown to have low oxygen permeability, up to three orders of magnitude lower than that of the unmodified epoxy (Triantafyllidis et al. 2006), see Table 1. The rate of diffusion of acetone into epoxy is also reduced by the addition of nanoclay (Chen and Curliss 2001), although these data did not reach saturation and so it is not possible to compare the equilibrium uptakes. The fire resistance of nanoparticle-modified polymers has been studied, but most work has concentrated on thermoplastics. This has generally shown that the addition of nanoclays, and nanotubes, can improve the fire resistance (e.g., Gilman 1999; Gilman et al. 2000b; Beyer 2002). A significant amount of work has been performed using various thermoset polymers modified with nanoclays, including epoxy, vinyl ester, cyanate ester, and polyimide (e.g., Gilman et al. 1999a; Gilman et al. 1999b; Table 1 Oxygen permeability data for films of epoxy and nanoclay-modified epoxy (Adapted from Triantafyllidis et al. 2006) Clay film composition Epoxy Epoxymontmorillonite Epoxymontmorillonite Epoxyfluoroectorite

Clay film thickness (mm) – 0.060

Epoxy-clay film thickness (mm) 0.20 0.11

O2 permeability (cm3 ml/m2 day) 98.9 0.1

0.035

0.10

0.97

0.065

0.14

1.2

56

Adhesives with Nanoparticles

1689

Gilman et al. 2000a). Tests using cone calorimetry showed that the addition of 10% of silicate reduced the peak heat release rate by up to 50%, and increased the mass of char residue formed. However, the onset time was also often reduced, probably due to presence of the surface treatment. The thermal degradation behavior of nanoclay-modified epoxy has been studied, and is closely linked to the fire performance. Nanoclay has little effect on this behavior, and may even reduce the temperature at which degradation starts due to the relative instability of the surface treatment (Brnardic et al. 2008). The addition of carbon nanofibers to an amine-cured epoxy polymer had no effect on the decomposition temperature of epoxy (Zhou et al. 2007), but single-walled nanotubes have been shown to degrade the thermal stability (Puglia et al. 2003). The effect on the glass transition temperature, Tg, of the addition of nanoparticles is dependent on the type of nanoparticles used and the interaction between the particles and the polymer (Becker and Simon 2006). The Tg can shift to a higher temperature than that of the unmodified polymer (e.g., Messersmith and Giannelis 1994; Kinloch and Taylor 2006; Bugnicourt et al. 2007). This may be due to a high interaction between the particles and the polymer, which may also locally change the properties of the polymer network. For example, the Tg of a polyurethane (PU) adhesive was increased from 32  C to 62  C by the addition of 5% of functionalized nanoclay (Dodiuk et al. 2006). Where polymer chains are intercalated between silicate platelets, then the silicate may constrain the polymer, hence increasing Tg (Kinloch and Taylor 2006). Bugnicourt et al. (2007) reported that Tg increased from 161  C to 180  C for an amine-cured epoxy modified with silica nanoparticles. They also noted that the poorer the dispersion, the lower the impact of the addition of silica on the magnitude of the tan δ peak and Tg. However, other tests showed no effect of the addition of silica nanoparticles (Bugnicourt et al. 2007; Johnsen et al. 2007; Hsieh et al. 2010). Reductions in the glass transition temperature have also been reported (e.g., Kornmann 1999). These reductions may be due to the surface treatment degrading during curing (Wang et al. 2000), or by enhanced free volume in the interphase between the particle and the matrix. Carbon nanotubes have been shown to increase the rate of the curing reaction of epoxies (Yin et al. 1993; Puglia et al. 2003). This effect arises from the surface chemistry of the nanotubes (Shaffer and Sandler 2006). However, barium titanate nanoparticles did not affect the curing behavior of epoxy cured using diamino diphenyl methane (DDM) (Chandradass and Bae 2008). The use of nanoparticles with a diameter less than that of the wavelength of light allows transparent particle-modified materials to be produced. In practice, the particles do reduce the optical transparency compared to the unmodified matrix due to differences in the respective refractive indices. For example, the light transmittance of epoxy modified with 25 nm diameter silica particles was greater than that for epoxy modified with particles of 540–1,520 nm (Naganuma and Kagawa 2002). However, using the smaller particles does give transmittance of 80–90% of the unmodified matrix value. It is reported that the addition of nanoclays to epoxy will also reduce the coefficient of thermal expansion (CTE) of the epoxy polymer, as the CTE of the

1690

A. C. Taylor

ceramic particles is less than that of the polymer (Wang et al. 2000). For example, the CTE of epoxy below Tg can be reduced from 77 to 63 μm/m C by the addition of 3% of nanoclay (Chen and Curliss 2001). However, the addition of nanotubes has been shown to have no significant effect on the CTE (Salinas-Ruiz 2009).

56.4.3 Electrical Properties Alternating current (AC) impedance spectroscopy has been used to measure the conductivity of an amine-cured epoxy modified with multiwalled carbon nanotubes (MWNTs) (Sandler et al. 2003). The conductivity increased as a function of the MWNT weight fraction, see Fig. 5. There was an increase in conductivity of about two orders of magnitude, from 10 9 to 10 7 S/m, when 0.001 wt% of nanotubes was added. There was a further increase to above 10 3 S/m for loading fractions of greater than 0.005 wt%. This increase in conductivity suggests that an infinite network of percolated nanotubes starts forming above 0.001 wt%. The authors also state that comparison with the direct current (DC) conductivity of the composites gave identical results. A similar transition in conductivity occurs for carbon black, but more than 1 wt% of carbon black is required (Sandler et al. 1999; Sandler et al. 2003), see Fig. 5.

Fig. 5 Conductivity versus filler content for aligned multi-walled carbon nanotubes, entangled nanotubes, and carbon black particles in epoxy (Courtesy M. Shaffer)

56

Adhesives with Nanoparticles

1691

The addition of layered silicates to epoxy resin improves the surface discharge endurance of the polymer, considerably increasing the time before electrical breakdown occurs (Montanari et al. 2005). The addition of nonconducting silica or alumina nanoparticles will reduce the conductivity of the polymer (Cao et al. 2004). Layered silicates have been shown to increase the electrical breakdown strength and breakdown time (Imai et al. 2006), see Fig. 6. This indicates that these materials may have potential applications for electrical insulation materials. The thermal conductivity of epoxy modified with carbon nanotubes has been investigated (Gojny et al. 2006). The authors reported that the thermal conductivity increased slightly with nanotube content, although the changes are very small compared to the changes in electrical conductivity.

56.4.4 Fracture Toughness It is well established that microparticles can increase the fracture toughness of thermoset polymers in mode I loading (e.g., Moloney et al. 1983; Kinloch and Taylor 2002). Indeed, this is the basis of most structural adhesives, which have a rubber or thermoplastic particulate toughening phase, often combined with co-tougheners. These particles may be preformed, or may be formed by phaseseparation of an initially soluble polymer during curing. Inorganic particles such as glass or mica also provide a toughening effect, though the increases in toughness are generally lower than for rubber-toughening. In addition, hybrid formulations that combine rubber and glass particles have been shown to give high toughness.

Fig. 6 Insulation breakdown time for conventionally filled (50 vol% silica microparticles) epoxy, and this epoxy with added nanoclay under constant AC voltage (10 kV, 1 kHz) (Adapted from Imai et al. 2006)

1692

A. C. Taylor

The addition of nanoclay particles to epoxy has been shown to increase the toughness. The fracture energy of vinylester/epoxy systems containing 5 wt% nanoclay was increased more than two times compared to the unmodified polymer (KargerKocsis et al. 2003). There was no significant effect of the type of the surface treatment used on the nanoclay. Increasing the amount of nanoclay above 5 wt% reduced both the fracture toughness and fracture energy. Kinloch and Taylor compared the toughening effect due to the addition of layered silicate nanoparticles (nanoclays) with that due to similar microparticles, in this case mica (Kinloch and Taylor 2003, 2006). They reported higher fracture toughness values for the mica microparticles than for the nanoclays, see Fig. 7. They also showed that the fracture toughness was dependent on the weight fraction of nanoclay, with maximum toughness being achieved at between 1 and 5 wt% of nanoclay. The fracture toughness was also dependent on type of nanoclay used, and hence on the surface treatment and the resulting morphology. The addition of layered silicates does not always increase, and may reduce the fracture toughness, for example, with thermoset acrylic systems (Tarrant 2004). Spherical nanoparticles have also been used to toughen epoxies. Ragosta et al. increased the fracture toughness, KC, of tetrafunctional epoxy cured with 4,4-diaminodiphenyl sulfone (DDS) from 0.5 to 1.2 MPam1/2 using 10 wt% of silica particles (Ragosta et al. 2005). The particles were between 10 and 15 nm in diameter. It has also been shown that the addition of 13 nm diameter alumina particles increased the fracture toughness of epoxy, cured using a cycloaliphatic amine, from 0.5 to 1.2 MPam1/2 using 11 vol% of alumina (Wetzel et al. 2006). Similarly, the addition of 11 vol% of 300 nm diameter titania particles increased KC to 0.85 MPam1/2.

Fig. 7 Fracture toughness of unmodified epoxy, mica microparticle- and nanoclay-modified epoxy versus content of layered silicate. Note “viscosity limit” indicates maximum silicate content before particle-modified resin becomes too viscous to cast (Adapted from Kinloch and Taylor 2006)

56

Adhesives with Nanoparticles

1693

The addition of silica nanoparticles increased the fracture toughness of an anhydride-cured epoxy from 0.59 to 1.42 MPam1/2 when 20 wt% of nanoparticles were added (Kinloch et al. 2005). These toughness values are equivalent to fracture energy, GC, values of 100 and 460 J/m2 respectively, see Fig. 8. In addition, 4 wt% of silica nanoparticles was sufficient to almost triple the fracture energy of the unmodified epoxy, to 290 J/m 2. The toughening mechanisms involved have been discussed by Johnsen et al. (2007). Zhang et al. (2006) and Rosso et al. (2006) also reported toughness increases due to the addition of similarly sized silica nanoparticles. The toughening mechanisms of silica nanoparticle-modified epoxy have been identified as plastic shear-yielding, and debonding of the polymer from the nanoparticles, followed by plastic void-growth of the epoxy (Hsieh et al. 2010). The toughening mechanisms of crack pinning, crack deflection, and immobilized polymer around the particles have all been discounted (Johnsen et al. 2007). Micron-sized rubber particles toughen polymers by cavitating, hence relieving the constraint in the plastic zone at the crack tip. The matrix is then able to deform, absorbing energy by dilation and shear banding, thus giving an increased toughness. However, nanometer-sized rubber particles are thought not to cavitate, because the triaxial stresses required for cavitation are too high. Theoretical models predict that there is a minimum size below which rubber particles will not cavitate (Lazzeri and Bucknall 1993; Dompas and Groeninckx 1994). This diameter is predicted to be approximately 50 nm (Kody and Lesser 1999). This lack of cavitation of small rubber particles has been observed experimentally. For example, no cavitation was observed with carboxyl-terminated butadiene-acrylonitrile (CTBN) rubber particles,

Fig. 8 Fracture energy versus content of silica nanoparticles in epoxy and rubber-toughened (9% CTBN) epoxy (Adapted from Kinloch et al. 2005)

1694

A. C. Taylor

which were 200 nm in diameter, but cavitation was seen with micron-sized CTBN particles (Chen and Jan 1992). However, cavitation was observed using 200 nm diameter core-shell latex particles comprised of a methacrylated butadiene-styrene copolymer (MBS) with a poly (methyl methacrylate) (PMMA) shell in a piperidine-cured epoxy polymer (Azimi et al. 1996b). Nano-sized core-shell rubber particles have also been reported to increase toughness (Liang and Pearson 2009). Liang and Pearson showed that the addition of 100 nm core-shell particles increased the fracture toughness of a piperidine-cured epoxy to 3.1 MPam1/2. Although no toughness was reported for the unmodified epoxy, similarly cured material has a fracture toughness of between 0.9 and 1.2 MPam1/2 (Oba 1999; Kawaguchi and Pearson 2003). In earlier work, the toughening effect of 200 nm MBS core-shell particles were compared to 1, 10, and 100 μm diameter CTBN particles (Pearson and Yee 1991). The fracture toughness was very dependent on particle size, and small particles are more efficient in producing a toughening effect than large particles. The largest increases in toughness have been reported for so-called hybrid materials, which combine rubber and nanoparticle-modification of the epoxy. For example, the addition of silica nanoparticles to a rubber-toughened epoxy, using 9 wt% of CTBN, increased the toughness from 1.11 to 2.19 MPam1/2 when 15 wt% of nanoparticles were added (Kinloch et al. 2005). These toughness values are equivalent to fracture energy, GC, values of 440 and 1,480 J/m2, respectively, see Fig. 8. Synergistic increases in fracture toughness (Liang and Pearson 2009), and in impact strength (Zeng et al. 2007) have also been reported. Synergistic toughening has previously been reported for rubber-toughened epoxy with micron-sized glass beads (e.g., Kinloch et al. 1985; Azimi et al. 1996a). Increases in toughness have also been observed when nanoclay is added to rubber-modified epoxy, but there is not such a strong synergistic effect in this case (Liu et al. 2004).

56.4.5 Fatigue Performance Nanoparticles have been shown to improve the fatigue performance of thermoset polymers. Wetzel et al. (2006) measured the fatigue crack propagation (FCP) rate, d a/d N, versus the stress-intensity factor range, and showed that this relationship follows the classic Paris-Erdogan law. These fatigue curves are shifted toward higher stress-intensity values as the nanoparticle content rises. At the same time, the gradient decreases. They reported that the gain in FCP resistance of aluminamodified epoxy nanocomposites (10 vol%) is comparable to the reinforcement achieved by rubber particles in epoxy, as was demonstrated using small amounts of rubber (5–10 phr CTBN) (Karger-Kocsis and Friedrich 1992), and is similar to that using 15–20 vol% of glass microspheres (Sautereau et al. 1995). In contrast to such traditional modifiers, the authors point out that the benefits conferred by nanoparticles are neither at the expense of modulus nor of strength. Silica nanoparticles have been shown to significantly improve the cyclic-fatigue behavior of an anhydride-cured epoxy polymer, increasing the range of applied

56

Adhesives with Nanoparticles

1695

stress-intensity factor at threshold, ΔKth, see Fig. 9 (Blackman et al. 2007). The nanoparticles also increased the fracture toughness, KC. The fatigue data followed the modified Paris law. Similarly, the addition of 200 nm diameter rubber particles increases fracture toughness and improves fatigue resistance (Azimi et al. 1996b). The fatigue data followed the classic Paris-Erdogan law. The use of 200 nm MBS particles in place of 1.5 μm CTBN particles resulted in about one order of magnitude improvement in fatigue resistance. A synergistic effect between CTBN rubber particles and silica nanoparticles has also been shown in fatigue, with the maximum increase in the fatigue threshold value being recorded using 9 wt% of rubber and 9 wt% of silica (Sohn Lee 2009). The addition of carbon nanofibers (CNF) to an amine-cured epoxy polymer significantly increased the number of cycles to failure, as measured using fatigue stress versus number of cycles to failure (S-N) curves (Zhou et al. 2007). The fatigue life went through a maximum at 2 wt% CNF, although only two specimens were used for each combination of stress level and CNF loading. A reduction in crack growth rate per cycle has also been shown for an epoxy system with the addition of up to 0.5 wt% of carbon nanotubes (Zhang et al. 2007). The authors attributed this effect to pullout of the nanotubes that bridge across the crack opening. Though there are data available that show how nanoparticles can change the fatigue performance of adhesives, there are no data that quantify the durability (i.e., the resistance to an environment such as water) of these materials when used as adhesives. However, it is likely that nanoparticle-modified adhesives will perform as conventionally filled adhesives.

Fig. 9 Logarithmic crack growth rate per cycle, d a/d N, versus logarithmic range of applied stressintensity factor from cyclic-fatigue tests for unmodified epoxy and silica nanoparticle-modified epoxy (Courtesy J. Sohn Lee)

1696

A. C. Taylor

56.4.6 Peel and Lap-Shear Performance The modification of a polyurethane (PU) adhesive using nanoclays was tested by Dodiuk et al. (2006), using adhesive joints made with aluminum-alloy adherends. The incorporation of functionalized nanoclays into PU improved the lap-shear strength by up to 195%. The functionalized nanoclays also gave higher peel strength than the unmodified PU. Increases of up to 40% were measured, but one of the nanoclays gave almost no change in the peel strength. The lap-shear strength of carbon-fiber composites bonded using an epoxy adhesive was increased by 45% with the addition of 5% of MWNT (Hsiao et al. 2003). However, the locus of failure was also altered – from interfacial for the control specimens to cohesive for the nanotube-modified materials. Climbing drum peel and lap-shear tests were performed using aluminum-alloy substrates, with a model rubber-toughened epoxy film adhesive (Gilbert et al. 2003). The adhesive was modified with 5 or 10 wt% of either 50 nm diameter alumina nanoparticles or alumina nanofibers with a diameter of 2–4 nm and an aspect ratio ranging from tens to hundreds. Both of the nano-modifiers increased the peel and shear strength. However, the more successful modification was the addition of 5% of alumina nanoparticles, which increased the peel strength of the adhesive by 50%, and the lap-shear strength by 15%. The lap-shear and 180 peel test performance of very soft acrylic adhesives modified with nanoclay or silica nanoparticles were improved by the addition of nanoparticles (Patel et al. 2006). The addition of 6 wt% of nanoclay or 50 wt% of silica increased the peel force by about 40%. Sprenger and coworkers used a room-temperature curing rubber-toughened epoxy adhesive with a Tg of 70  C (Kinloch et al. 2003; Sprenger et al. 2003, 2004). The addition of only 4 wt% of 20 nm diameter silica nanoparticles doubled the fracture energy, from 1,200 to 2,300 J/m 2. The addition of the nanoparticles also increased both the lap-shear and peel strength, as shown in Fig. 10.

56.5

Conclusions

Nanoparticle-modified polymers are being extensively researched, and a great deal of work has been done to produce nanoparticles. These nanoparticles are now commercially available in kilogram or tonne quantities, and at reasonable prices ($10s per kilogram). Hence, nanotechnology and nanoparticles are beginning to be applied in commercial products. It is anticipated that nanoparticles will be increasingly used in structural adhesives. Their application will continue to expand as their availability increases, and hence as their cost decreases. The dispersion of nanoparticles in polymer matrices can be difficult, and agglomerates can remain even after aggressive processing. These may have a significant detrimental effect on the properties of the material. For this reason, pre-dispersed nanoparticles, available as a masterbatch in a polymer, are attractive. However, these

56

Adhesives with Nanoparticles

1697

Fig. 10 Lap-shear and roller peel strength for hybrid (nanosilica and rubber-modified) epoxy (Adapted from Kinloch et al. 2003)

particles can agglomerate during processing or curing, so there is no guarantee that a good dispersion will be obtained in the final material. Nanoparticles have been shown to improve the structural and functional properties of thermoset polymers. However, in many cases similar increases can be obtained using micron-sized particles, especially for structural properties. High aspect ratio particles, such as carbon nanotubes or nanoclay, can give significant improvements in functional properties. Examples include electrical percolation at very low particle volume fractions and increases in barrier properties. It is perhaps the synergistic effect of combining nanoparticles with existing technology based on micron-sized particles that is the most exciting area at present. For example, the fracture toughness and peel performance of adhesive joints can be improved by combining silica nanoparticles and rubber microparticles. Hierarchical materials like these, and patterned arrays of nanoparticles, appear to be the future developments in this field.

References Ahmed S, Jones FR (1990) A review of particulate reinforcement theories for polymer composites. J Mater Sci 25:4933 Ajayan PM, Stephan O et al (1994) Aligned carbon nanotube arrays formed by cutting a polymer resin-nanotube composite. Science 265:1212 Alexandre M, Dubois P (2000) Polymer-layered silicate nanocomposites: preparation, properties and uses of a new class of materials. Mater Sci Eng R 28:1 Azimi HR, Pearson RA et al (1996a) Fatigue of hybrid epoxy composites: Epoxies containing rubber and hollow glass spheres. Polym Eng Sci 36:2352

1698

A. C. Taylor

Azimi HR, Pearson RA et al (1996b) Fatigue of rubber-modified epoxies: effect of particle size and volume fraction. J Mater Sci 31:3777 Baik Y, Lee S et al (2005) Unidirectional alignment of carbon nano-sized fiber using drawing process. J Mater Sci 40:6037 Bechelany M, Brioude A et al (2007) Large-scale preparation of faceted Si 3N 4 nanorods from betaSiC nanowires. Nanotechnology 18:335305 Becker O, Simon GP (2006) Epoxy nanocomposites based on layered silicates and other nanostructured fillers. In: Mai YW, Yu ZZ (eds) Polymer nanocomposites. Woodhead, Cambridge, p 594 Beyer G (2002) Carbon nanotubes as flame retardants for polymers. Fire Mater 26:291 Blackman BRK, Kinloch AJ et al (2007) The fracture and fatigue behaviour of nano-modified epoxy polymers. J Mater Sci 42:7049 Brnardic I, Macan J et al (2008) Thermal degradation kinetics of epoxy/organically modified montmorillonite nanocomposites. J Appl Polym Sci 107:1932 Brooker RD, Blackman BRK et al (2008) Nano-reinforcement of epoxy/thermoplastic blends. In: 31st annual meeting of the adhesion society, Austin, USA, Adhesion Society, Blacksburg, USA Brooker RD, Kinloch AJ et al (2010) The morphology and fracture properties of thermoplastictoughened epoxy polymers. J Adhes 86:726 Brown DM, Kinloch IA et al (2007) An in vitro study of the potential of carbon nanotubes and nanofibres to induce inflammatory mediators and frustrated phagocytosis. Carbon 45:1743 Bugnicourt E, Galy J et al (2007) Effect of sub-micron silica fillers on the mechanical performances of epoxy-based composites. Polymer 48:1596 Cao Y, Irwin PC et al (2004) The future of nanodielectrics in the electrical power industry. IEEE Trans Dielectric Electrical Insul 11:797 Chakrabarti S, Nagasaka T et al (2006) Growth of super long aligned brush-like carbon nanotubes. Jpn J Appl Phys Expr Lett 45:L720 Chandradass J, Bae D-S (2008) Preparation and properties of barium titanate nanopowder/epoxy composites. Mater Manuf Process 23:116 CheapTubes (2009.) http://www.cheaptubes.com. Retrieved 4 Nov 2009 Chen C, Curliss D (2001) Resin matrix composites: organoclay-aerospace epoxy nanocomposites, Part II. SAMPE J 37:11 Chen TK, Jan YH (1992) Fracture mechanism of toughened epoxy-resin with bimodal rubberparticle size distribution. J Mater Sci 27:111 Chen XQ, Saito T et al (2001) Aligning single-wall carbon nanotubes with an alternating-current electric field. Appl Phys Lett 78:3714 Demczyk BG, Wang YM et al (2002) Direct mechanical measurement of the tensile strength and elastic modulus of multiwalled carbon nanotubes. Mater Sci Eng A 334:173 Dodiuk H, Belinski I et al (2006) Polyurethane adhesives containing functionalized nanoclays. J Adhes Sci Technol 20:1345 Dompas D, Groeninckx G (1994) Toughening behavior of rubber-modified thermoplastic polymers involving very small rubber particles. 1. A criterion for internal rubber cavitation. Polymer 35:4743 Feynman RP (1959) .There’s plenty of room at the bottom – an invitation to enter a new field of physics http://www.zyvex.com/nanotech/feynman.html. Accessed 9 Apr 2008 Fiedler B, Gojny FH et al (2006) Fundamental aspects of nano-reinforced composites. Compos Sci Technol 66:3115 Gilbert EN, Hayes BS et al (2003) Nano-alumina modified epoxy based film adhesives. Polym Eng Sci 43:1096 Gilman JW (1999) Flammability and thermal stability studies of polymer layered-silicate (clay) nanocomposites. Appl Clay Sci 15:31 Gilman JW, Harris RH et al (1999b) Cyanate ester clay nanocomposites: synthesis and flammability studies. In: Evolving and revolutionary technologies for the new millennium 44th international SAMPE symposium/exhibition, Long beach, CA, USA, Society for the Advancement of Material and Process Engineering (SAMPE)

56

Adhesives with Nanoparticles

1699

Gilman JW, Kashiwagi T et al (1999a) Flammability studies of polymer layered silicate nanocomposites: polyolefin, epoxy and vinyl ester resins. In: Ak-Malaika S, Golovoy A et al (eds) Chemistry and technology of polymer additives. Blackwell Science, Malden, p 249 Gilman JW, Harris RH et al (2000b) Phenolic cyanate ester nanocomposites: effect of ammonium ion structure on flammability and nano-dispersion. In: Spring meeting, division of polymeric materials: science and engineering, American Chemical Society, San Francisco, USA, American Chemical Society Gilman JW, Jackson CL et al (2000a) Flammability properties of polymer-layered-silicate nanocomposites. polypropylene and polystyrene. Nanocomposites Chem Mater 12:1866 Gintert MJ, Jana SC et al (2007) A novel strategy for nanoclay exfoliation in thermoset polyimide nanocomposite systems. Polymer 48:4166 Gojny FH, Nastalczyk J et al (2003) Surface modified multi-walled carbon nanotubes in CNT/ epoxy-composites. Chem Phys Lett 370:820 Gojny FH, Wichmann MHG et al (2004) Carbon nanotube-reinforced epoxy-composites: enhanced stiffness and fracture toughness at low nanotube content. Compos Sci Technol 64:2363 Gojny FH, Wichmann MHG et al (2006) Evaluation and identification of electrical and thermal conduction mechanisms in carbon nanotube/epoxy composites. Polymer 47:2036 Gordon JE (1978) The new science of strong materials or why you don’t fall through the floor. Penguin, Harmondsworth Green K, Dean D et al (2006) Aligned carbon nanofiber/epoxy nanocomposites. Polym Mater Sci Eng 94:53 Gusev AA, Lusti HR (2001) Rational design of nanocomposites for barrier applications. Adv Mater 13:1641 Hackman I, Hollaway L (2006) Epoxy-layered silicate nanocomposites in civil engineering. Composites Pt A 37:1161 Harper T (2003) What is nanotechnology? Nanotechnology 14:U5 Harris PJF (1999) Carbon nanotubes and related structures: new materials for the twenty-first century. Cambridge University Press, Cambridge Harris PJF (2009) Carbon nanotube science: synthesis, properties and applications. Cambridge University Press, Cambridge Hay J, Shaw S (2001) Into the Labyrinth. Chem Br 2001:34 Hsiao K-T, Alms J et al (2003) Use of epoxy/multiwalled carbon nanotubes as adhesives to join graphite fibre reinforced polymer composites. Nanotechnology 14:791 Hsieh TH, Kinloch AJ et al (2010) The toughness of epoxy polymers and fibre composites modified with rubber microparticles and silica nanoparticles. J Mater Sci 45:1193 ICBA (2008) .International Carbon Black Association, http://www.carbon-black.org. Retrieved 9 Apr 2008 Iijima S (1991) Helical microtubules of graphitic carbon. Nature 354:56 Iijima S, Ichihashi T (1993) Single-shell carbon nanotubes of 1-nm diameter. Nature 363:603 Imai T, Sawa F et al (2006) Effects of nano- and micro-filler mixture on electrical insulation properties of epoxy based composites. IEEE Trans Dielectric Electrical Insul 13:319 Johnsen BB, Kinloch AJ et al (2007) Toughening mechanisms of nanoparticle-modified epoxy polymers. Polymer 48:530 Kaneka (2008.) http://www.kaneka.com. Retrieved 9 Apr 2008 Karger-Kocsis J, Friedrich K (1992) Fatigue crack-propagation and related failure in modified, anhydride-cured epoxy-resins. Colloid Polym Sci 270:549 Karger-Kocsis J, Gryshchuk O et al (2003) Interpenetrating vinylester/epoxy resins modified with organophilic layered silicates. Compos Sci Technol 63:2045 Kathi J, Rhee K (2008) Surface modification of multi-walled carbon nanotubes using 3-aminopropyltriethoxysilane. J Mater Sci 43:33 Kawaguchi T, Pearson RA (2003) The effect of particle-matrix adhesion on the mechanical behavior of glass filled epoxies. Part 2. A study on fracture toughness. Polymer 44:4239 Kinloch AJ, Lee JH et al (2003) Toughening structural adhesives via nano- and micro-phase inclusions. J Adhes 79:867

1700

A. C. Taylor

Kinloch AJ, Maxwell DL et al (1985) The fracture of hybrid-particulate composites. J Mater Sci 20:4169 Kinloch AJ, Mohammed RD et al (2005) The effect of silica nano particles and rubber particles on the toughness of multiphase thermosetting epoxy polymers. J Mater Sci 40:5083 Kinloch AJ, Taylor AC (2002) The toughening of cyanate-ester polymers. Part I: physical modification using particles, fibres and woven-mats. J Mater Sci 37:433 Kinloch AJ, Taylor AC (2003) Mechanical and fracture properties of epoxy/inorganic micro- and nano-composites. J Mater Sci Lett 22:1439 Kinloch AJ, Taylor AC (2006) The mechanical properties and fracture behaviour of epoxyinorganic micro- and nano-composites. J Mater Sci 41:3271 Kody RS, Lesser AJ (1999) Yield behavior and energy absorbing characteristics of rubber-modified epoxies subjected to biaxial stress states. Polym Compos 20:250 Koerner H, Hampton E et al (2005) Generating triaxial reinforced epoxy/montmorillonite nanocomposites with uniaxial magnetic fields. Chem Mater 17:1990 Kornmann X (1999) Synthesis and characterisation of thermoset-clay nanocomoposites. Licentiate thesis, materials and manufacturing engineering department, University of Lulea, Lulea, Sweden Lam C-W, James JT et al (2004) Pulmonary toxicity of single-wall carbon nanotubes in mice 7 and 90 days after intratracheal instillation. Toxicol Sci 77:126 Lam C-K, Lau K-T et al (2005) Effect of ultrasound sonication in nanoclay clusters of nanoclay/ epoxy composites. Mater Lett 59:1369 Lau K-T, Gu C et al (2004) Stretching process of single- and multi-walled carbon nanotubes for nanocomposite applications. Carbon 42:426 Lazzeri A, Bucknall CB (1993) Dilatational bands in rubber-toughened polymers. J Mater Sci 28:6799 Liang G, Hu X (2004) Preparation and performance of aluminum borate whisker-reinforced epoxy composites. I. Effect of whiskers on processing, reactivity, and mechanical properties. J Appl Polym Sci 92:1950–1954 Liang YL, Pearson RA (2009) Toughening mechanisms in epoxy-silica nanocomposites (ESNs). Polymer 50:4895 Liu WP, Hoa SV et al (2004) Morphology and performance of epoxy nanocomposites modified with organoclay and rubber. Polym Eng Sci 44:1178 Lu KL, Lago RM et al (1996) Mechanical damage of carbon nanotubes by ultrasound. Carbon 34:814 Martin CA, Sandler JKW et al (2005) Electric field-induced aligned multi-wall carbon nanotube networks in epoxy composites. Polymer 46:877 Messersmith PB, Giannelis EP (1994) Synthesis and characterisation of layered silicate-epoxy nanocomposites. Chem Mater 6:1719 Mohammed RD (2007) Material properties and fracture mechanisms of epoxy nano-composites. PhD thesis, mechanical engineering. Imperial college of science, technology and medicine, London Moloney AC, Kausch HH et al (1983) The fracture of particulate-filled epoxide resins. J Mater Sci 19:1125 Montanari GC, Ciani F et al (2005) Electric strength, space charge and surface discharge characterization of nanostructured epoxy-silicate insulating materials. In: 2005 International symposium on electrical insulating materials, vols 1–3, Tokyo, Institute of Electrical Engineers, Japan Morisada Y, Miyamoto Y et al (2007) Mechanical properties of SiC composites incorporating SiC-coated multi-walled carbon nanotubes. Int J Refract Met Hard Mater 25:322 Naganuma T, Kagawa Y (2002) Effect of particle size on the optically transparent nano meter-order glass particle-dispersed epoxy matrix composites. Compos Sci Technol 62:1187 Nanoresins (2008.) http://www.nanoresins.com. Retrieved 9 Apr 2008 Oba T (1999) The fatigue behaviour of toughened epoxy polymers. PhD thesis, mechanical engineering. Imperial college of science, technology and medicine, London

56

Adhesives with Nanoparticles

1701

Patel S, Bandyopadhyay A et al (2006) Synthesis and properties of nanocomposite adhesives. J Adhes Sci Technol 20:371 Pearson RA, Yee AF (1991) Influence of particle-size and particle-size distribution on toughening mechanisms in rubber-modified epoxies. J Mater Sci 26:3828 Puglia D, Valentini L et al (2003) Analysis of the cure reaction of carbon nanotubes/epoxy resin composites through thermal analysis and Raman spectroscopy. J Appl Polym Sci 88:452 Radford T (2003) Brave new world or miniature menace? Why Charles fears grey goo nightmare. The Guardian, London, Issued 29 April 2003 Ragosta G, Abbate M et al (2005) Epoxy-silica particulate nanocomposites: chemical interactions, reinforcement and fracture toughness. Polymer 46:10506 Rosso P, Ye L et al (2006) A toughened epoxy resin by silica nanoparticle reinforcement. J Appl Polym Sci 100:1849 Rothon RN, Hancock M (1995) General principles guiding selection and use of particulate materials. In: Rothon RN (ed) Particulate-filled polymer composites. Longman Scientific and Technical, Harlow, p 1 Saito R, Dresselhaus G et al (1998) Physical properties of carbon nanotubes. Imperial College, London Salinas-Ruiz MDM (2009) Development of a rubber toughened epoxy adhesive loaded with carbon nanotubes, for aluminium – polymer bonds. PhD thesis, School of applied science, Cranfield University Sandler J, Shaffer MSP et al (1999) Development of a dispersion process for carbon nanotubes in an epoxy matrix and the resulting electrical properties. Polymer 40:5967 Sandler JKW, Kirk JE et al (2003) Ultra-low electrical percolation threshold in carbon-nanotubeepoxy composites. Polymer 44:5893 Sautereau H, Maazouz A et al (1995) Fatigue behavior of glass bead filled epoxy. J Mater Sci 30:1715 Schadler LS (2003) Polymer-based and polymer-filled nanocomposites. In: Ajayan PM, Schadler LS et al (eds) Nanocomposite science and technology. Wiley-VCH, Weinheim, p 380 Schmid CF, Klingenberg DJ (2000) Mechanical flocculation in flowing fiber suspensions. Phys Rev Lett 84:290 Shaffer MSP, Kinloch IA (2004) Prospects for nanotube and nanofibre composites. Compos Sci Technol 64:2281 Shaffer MSP, Sandler JKW (2006) Carbon nanotube/nanofibre polymer composites. In: Advani SG (ed) Processing and properties of nanocomposites. World Scientific, Singapore Sheng N, Boyce MC et al (2004) Multiscale micromechanical modeling of polymer/clay nanocomposites and the effective clay particle. Polymer 45:487 Sigma Aldrich (2008.) http://www.sigmaaldrich.com Retrieved 9 Apr 2008 Sohn Lee J (2009) The fatigue behaviour of nano-modified epoxy adhesives. PhD thesis, Department of mechanical engineering, Imperial college London, London Sorrentino A, Gorrasi G et al (2006) Barrier properties of polymer/clay nanocomposites. In: Mai YW, Yu ZZ (eds) Polymer nanocomposites. Woodhead, Cambridge, p 594 Sprenger S, Eger C et al (2003) Nanoadhesives: toughness and high strength. Adhaesion, Kleben and Dichten 2003:24 Sprenger S, Eger C et al (2004). Nano-modified ambient temperature curing epoxy adhesives. Adhaesion, Kleben and Dichten 2004:1 Tang Y, Liang G et al (2007) Performance of aluminum borate whisker reinforced cyanate ester resin. J Appl Polym Sci 106:4131 Tarrant AE (2004) Thermoset-acrylic/layered-silicate nanocomposites: synthesis and structureproperty relationships. Imperial College London Thesis, London Tjong SC (2006) Structural and mechanical properties of polymer nanocomposites. Mater Sci Eng R 53:73 Triantafyllidis KS, LeBaron PC et al (2006) Epoxy-clay fabric film composites with unprecedented oxygen-barrier properties. Chem Mater 18:4393

1702

A. C. Taylor

Vanorio T, Prasad M et al (2003) Elastic properties of dry clay mineral aggregates, suspensions and sandstones. Geophys J Int 155:319 Wang ZL (ed) (2001) Characterization of nanophase materials. Wiley-VCH, Weinheim Wang Z, Massam J et al (2000) Epoxy-clay nanocomposites. In: Pinnavaia TJ, Beall GW (eds) Polymer-clay nanocomposites. Wiley, Chichester Warheit DB, Laurence BR et al (2004) Comparative pulmonary toxicity assessment of single-wall carbon nanotubes in rats. Toxicol Sci 77:117 Wetzel B, Rosso P et al (2006) Epoxy nanocomposites – fracture and toughening mechanisms. Eng Fract Mech 73:2375 Xie X-L, Mai Y-W et al (2005) Dispersion and alignment of carbon nanotubes in polymer matrix: A review. Mater Sci Eng R 49:89 Yasmin A, Daniel IM (2004) Mechanical and thermal properties of graphite platelet/epoxy composites. Polymer 45:8211 Yasmin A, Abot JL et al (2003) Processing of clay/epoxy nanocomposites by shear mixing. Scr Mater 49:81 Yasmin A, Luo JJ et al (2006a) Mechanical and thermal behavior of clay/epoxy nanocomposites. Compos Sci Technol 66:2415 Yasmin A, Luo JJ et al (2006b) Processing of expanded graphite reinforced polymer nanocomposites. Compos Sci Technol 66:1182 Yin M, Koutsky JA et al (1993) Characterization of carbon microfibers as reinforcement for epoxy resins. Chem Mater 5:1024 Zeng MF, Sun XD et al (2007) Effects of SiO2 nanoparticles on the performance of carboxylrandomized liquid butadiene-acrylonitrile rubber modified epoxy nanocomposites. J Appl Polym Sci 106:1347 Zhang H, Zhang Z et al (2006) Property improvements of in situ epoxy nanocomposites with reduced interparticle distance at high nanosilica content. Acta Mater 54:1833 Zhang W, Picu RC et al (2007) Suppression of fatigue crack growth in carbon nanotube composites. Appl Phys Lett 91:193109 Zhou Y, Pervin F et al (2007) Effect vapor grown carbon nanofiber on thermal and mechanical properties of epoxy. J Mater Sci 42:7544 Zhu YQ, Kroto HW et al (2002) A systematic study of ceramic nanostructures generated by arc discharge. Chem Phys Lett 365:457 Zunjarrao S, Sriraman R et al (2006) Effect of processing parameters and clay volume fraction on the mechanical properties of epoxy-clay nanocomposites. J Mater Sci 41:2219

Adhesive Dentistry

57

John W. Nicholson

Contents 57.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.2 Surfaces for Bonding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.3 Surface Pretreatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.4 Bonding Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.5 Evaluation of Bonding Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.6 Leakage and Bonding Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.7 Composite Resins . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.8 Durability of Bonded Composite Resin Restorations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.9 Biocompatibility of Bonding Agents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.10 Adhesive Dental Restorative Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.11 Bonding of Glass-Ionomer Cements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.12 ART Technique . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.13 Glass-Ionomer Fissure Sealants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.14 Bonding in Orthodontics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57.15 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1704 1706 1708 1710 1713 1713 1714 1715 1716 1717 1720 1723 1723 1724 1725 1726

Abstract

The present chapter describes the adhesive techniques and materials that are widely used in modern clinical dentistry. There are two types of tooth-colored restorative material available to the dentist, the so-called composite resins and the glass-ionomer cements. There are, though, variations on these basic types, as explained in the chapter. In addition, there is the zinc polycarboxylate cement, which was the first adhesive dental restorative material to be developed, and it retains a niche in modern clinical practice. J. W. Nicholson (*) Bluefield Centre for Biomaterials, London, UK Dental Physical Sciences, Institute of Dentistry, Queen Mary University of London, London, UK e-mail: john.nicholson@bluefieldcentre.co.uk; [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_56

1703

1704

J. W. Nicholson

Composite resins do not themselves possess any adhesive properties, but need to be used in conjunction with separate adhesive substances known as bonding agents. Over recent years, the technology of these materials has developed rapidly and a wide variety of systems are currently available to the clinician. Bonding agents have been characterized by “generation,” and currently seventh generation bonding systems are available. However, a more useful classification is based on the number of steps involved in using them, and this is described in detail. Bonding agents are applied to the tooth surface following cavity preparation, a process that involves the formation of a layer of deranged dentin known as the “smear layer.” Modern bonding agents may be either “etch-and-rinse,” in which case the smear layer is removed, or “self-etching,” where the smear layer becomes incorporated into the bonding layer. Whichever approach is used, the result is a treated surface to which the composite resin is applied. Testing of such systems is typically in shear mode, and they have been reported to give high shear bond strengths, even though they involve only a very thin layer between a hydrophilic substrate and a hydrophobic repair material. Glass-ionomers, by contrast, are themselves hydrophilic and inherently adhesive to the tooth surface. It therefore requires only slight pretreatment, typically conditioning with 10% aqueous polyacrylic acid. Glass-ionomers rapidly form durable adhesive bonds to the tooth surface through the development of an ion-exchange layer at the interface. This ability to bond is exploited in both the Atraumatic Restorative Treatment (ART) technique and in bonding of orthodontic brackets, both of which are described in detail in the chapter.

57.1

Introduction

Dentistry is concerned with the preservation, repair, and replacement of teeth. In its contemporary form, it makes increasing use of adhesive materials and techniques. One example is the placement of adhesive materials to fill fissures of newly erupted teeth as a preventive procedure. Without this treatment, decay-inducing bacteria might otherwise lodge in these fissures, and cause caries to develop (Mount and Hume 2005). Other options in restorative dentistry include repair of teeth damaged by caries or trauma with the tooth-colored materials known as composite resins. Use of these materials involves bonding to the tooth by means of bespoke adhesive systems. Alternatively, teeth damaged by caries can be repaired with the other toothcolored option, the glass-ionomer cement. These materials are inherently adhesive to both the dentin and enamel of the tooth, which in some applications is an advantage (Mount 2002). Replacement of missing teeth using bridges typically employs bonding agents and surface pretreatment procedures (Roulet and Vanherle 2004). Tooth-realigning techniques used in the dental specialty known as orthodontics employ brackets and wires to force the teeth into more satisfactory positions. Adhesive materials can be used to hold these brackets in place. Aesthetic procedures, such as the placement of

57

Adhesive Dentistry

1705

ceramic veneers to improve the appearance of the incisors, also typically employ adhesives and surface pretreatments (Mount and Hume 2005). From these comments it is apparent that adhesion and adhesives play important roles in modern dentistry. Current research covers many facets of the adhesion process, including surface pretreatment, formulation of adhesives, durability of bonded structures, and clinical effects of bonding agents and adhesive materials (Roulet and Vanherle 2004; Breschi et al. 2008). The deployment of these substances in the mouth is not entirely straightforward. The mouth is a biological structure, and as such is composed of both hard and soft tissues. The chemicals used as adhesives may have adverse effects, particularly on the soft tissue of the gingiva (gum) and pulp (living cells and blood vessels within the tooth). A theme of the current research is the biocompatibility of adhesive materials in the mouth. Tooth repairs are required to be capable of surviving for a considerable time, typically of the order of 10–15 years. The long-term effects of foreign materials and their possible degradation products need to be considered in selecting repair materials and planning treatments involving their deployment. There are many factors driving the changes to increased use of adhesive techniques in dentistry (Roulet and Vanherle 2004). First, there have been many years of research and development activity directed toward the fabrication of esthetic repair materials, both polymer- and ceramic-based. Second, there is an increasing patient demand for esthetic (i.e., tooth-colored) repairs. This is part of a wider trend toward improved dental esthetics that has seen a rise in the developed world of cosmetic dentistry as a branch of the profession and the increasing use of bleaching techniques to make the teeth whiter. This is associated with looking younger and also compensates for the ravages of strongly colored dietary components, such as coffee and red wine. A third factor contributing to the increase in importance of adhesion in dentistry is that patients are retaining their natural teeth for longer. Whereas earlier generations were prepared to accept some tooth loss and even complete edentulousness as a consequence of ageing, the current generation is less tolerant of these possibilities. There has been a dramatic fall in the incidence of tooth decay in the past 40 years, a phenomenon which is multifactorial, but is closely allied to the increased use of fluoride for preventive purposes. Preventive measures include fluoridation of drinking water, the supply of fluoridated toothpastes, and the application of fluoridated gels by dental professionals (Nicholson and Czarnecka 2008a). As a result, patients retain many of their teeth well into old age, and having done so, are then unwilling to accept radical treatment options, such as extraction or even highly invasive repair with silver-tin amalgam. Increasing levels of affluence, too, make such patients willing to pay for demanding clinical procedures and more expensive materials, which means those based on adhesive techniques and materials. Finally, the major advances in understanding of bond formation and durability to the tooth surface has enhanced the ability of the scientific community to develop reliable and long-lasting bonding agents. The dental literature reflects this, containing as it does many hundreds, possibly thousands, of scientific articles on the topic of bonding and adhesion. These range from fundamental studies of bonding

1706

J. W. Nicholson

mechanisms and surface pretreatments to clinical studies of longevity of repairs and failure mechanisms. One indication of the importance of this subject is that a specialist journal, the Journal of Adhesive Dentistry, was launched in 1999. However this is not the only forum for discussion of research on this topic, and other journals, such as Journal of Dental Research, Dental Materials and Journal of Dentistry also carry numerous research articles every year on adhesive dentistry. The current chapter aims to cover the subject of adhesive dentistry as comprehensively as possible. Beginning with a consideration of the surfaces to which bonding takes place, it goes on to consider the nature of the main groups of esthetic dental repair materials (the composite resins and the glass-ionomer cements) and how their adhesion is brought about and their durability maximized. There is considerable focus both on fundamental understanding of the mechanism of adhesion and on the clinical outcomes that can be achieved using these materials. Finally, other clinical applications that rely on adhesion are described. These are (i) orthodontic bracket fixation, (ii) fissure sealing, and (iii) the so-called Atraumatic Restorative Treatment technique (ART). The latter is an important development in clinical dentistry, particularly aimed at Third World countries, which is possible only with adhesive materials and which marks an innovative approach within dentistry.

57.2

Surfaces for Bonding

Unlike adhesive bonding in most areas of technology, in dentistry adhesion involves a living substrate. This is more variable than most surfaces to which adhesive materials are typically applied. To understand the significance of this, the nature of the substrates that are encountered in dentistry must be considered, which means beginning with a consideration of the anatomy of the tooth. The tooth consists essentially of two types of tissue, the enamel and the dentin. Their compositions are shown in Table 2. From this it can be seen that the mineral phase of the enamel far exceeds that of the dentin, resulting in the enamel being much the harder and more brittle of the two tissues. Enamel and dentin are joined by a distinctive dentino-enamel junction (DEJ). This zone has important mechanical properties, uniting as it does a thin and highly brittle layer (enamel) with a tougher, more compliant underlying structure (dentin). Among the important properties of the DEJ is that it does not allow cracks to cross from the enamel to the dentin. This feature maintains the integrity of the whole tooth, so that even if small cracks appear in the enamel, they do not propagate through the dentin and damage the whole tooth. The mineral phase of the tooth is predominantly hydroxyapatite, which is a calcium phosphate mineral of nominal composition Ca 10(PO 4) 6(OH) 2 and has a Ca:P ratio of 1.67:1. Naturally occurring hydroxyapatite, such as that in teeth and also in bones, tends to be a slightly different composition from this idealized formula. It is typically calcium deficient and has a small proportion of its phosphate groups replaced by carbonates (Table 1). This natural mineral, though, is still very

57

Adhesive Dentistry

Table 1 Composition of tooth

1707

Inorganic phase (mainly hydroxyapatite) Organic phase (mainly collagen) Water Ca:P ratio

Enamel 94–96% 4–5% 1% 1.64

Dentin 70% 20% 10% 1.56

Fig. 1 SEM of smear layer formed by slow-speed cavity preparation, magnification 1000 (Mount 1994)

hard and the details of its composition make little or no difference to its mechanical properties. In order to carry out the various processes involved in dental repair, the tooth generally has to be cut with some sort of instrument. This is usually a rotary bur of the so-called dental drill. The cutting process results in the formation of a layer known as the smear layer on the surface of the tooth (Fig. 1). The smear layer is thin, typically 1–2 μm deep, and is very tenacious. It consists of mineral phase embedded in denatured collagen, and is essentially material with a less ordered structure than uncut enamel or dentin. The detailed composition, morphology, roughness, and tenacity of the smear layer vary according to the precise details of the cutting process (Ogata et al.

1708

J. W. Nicholson

2001). Despite this, the smear layer may be left in place and act as the surface to which bonding is carried out. Dental adhesives need to be able to penetrate this layer, and possibly interact with it, to form a layer capable of being retained for many years. Bond strengths and durability are known to vary according to how this smear layer is created (Koibuchi et al. 2001), and the differences in its properties that result from the different ways it may be formed are responsible for the variations in clinical performance of dental bonding agents. During the formation of the smear layer, cutting debris is usually forced down into the tubules of the dentin. This debris results in the formation of so-called “smear plugs” and they have the effect of reducing the permeability of the dentin, and also of reducing the total surface area available for bonding (Koibuchi et al. 2001). The reduction in permeability may be an advantage, however, as it stops fluid being forced up the dentinal tubules. This, in turn, protects the bond formed by an adhesive placed onto the cut surface. The surface treatment that follows cutting may attempt to modify or remove the smear layer, in order to ensure that a uniform and reliable surface is left behind. It may also remove the smear plugs, thus opening the tubules and allowing the bonding agent to penetrate the tooth surface. This may not be desirable, since it may lead to postoperative sensitivity and discomfort in the patient. There is considerable debate amongst dental practitioners as to which is the best way to carry out tooth preparation, and these varying views show that, in the field of adhesive dentistry, maximizing the bond strength is not the only consideration.

57.3

Surface Pretreatment

The earliest dental restorative materials, gold foil and silver amalgam, have no adhesive properties whatsoever. They needed to be held in place by enlarging the cavity caused by the tooth decay, and forming undercuts. This is still the method used for retaining silver amalgam fillings, gold foil having passed into disuse many years ago. This approach of removing sound, healthy tooth tissue is widely recognized in the dental profession as a major disadvantage of silver amalgams (Mount and Hume 2005). The first critical step in developing adhesion in dentistry occurred in 1955, when Buonocore reported the acid-etch process for pretreatment of enamel (Buonocore 1955). This employed solutions of phosphoric acid at 37% concentration, and when enamel surfaces were pretreated in this way, microscopic surface irregularities were created, and polymeric filling materials were able to flow into them before hardening. Once these materials had hardened, there was an interlocking effect that strongly united the filling material with the tooth. Shear bond strengths of the order of 20–25 MPa were obtained, which were enough for the resulting bonded filling to withstand the forces of biting and chewing, and to show very good retention rates. At the time that this work was carried out, the existence of the smear layer was not known. It was obvious, however, that the acid-etch process was critical to successful bonding and that it involved the development of a surface to which bonding could

57

Adhesive Dentistry

1709

occur. The concentration of the phosphoric acid was critical, too. The apparently arbitrary concentration of 37% lies halfway between 25%, which was too dilute to be effective, and 50%, which is so concentrated that it removes too much tooth material, and does not result in the formation of a satisfactory interlock bond. Acid etching owes part of its success to the nature of the enamel. Enamel contains very little protein or water and can therefore be dried without causing the structure to collapse. Clinically, the procedure leads to the observation of a chalky appearance to the enamel, which can be taken as an indication that etching has been successful and the surface is ready to receive the bonding layer. Bonding to dentin has proved much more challenging. Dentin is wetter than enamel and contains protein in the form of collagen fibers. Acid etching as developed by Buonocore is much less successful when applied to dentin than to enamel, and shear bond strengths are much lower, typically lying in the region 5–10 MPa. This is because etching with 37% phosphoric acid changes the nature of the dentin surface, and it ceases to be a hard, mineralized tissue. Instead, removal of the mineral phase by the acid leaves behind a soft, collagen-rich surface structure, which collapses when dried with air. Unlike etched enamel, the resulting mat of collagen fibers forms only weak bonds with organic monomers, which is why the resulting shear bond strengths are low. In order to form satisfactory bonds to dentin, it is important to replace the mineral component by monomer (generally termed “resin” by researchers in this field). This resin infiltration needs to take place immediately after removal of the mineral phase, or even possibly at the same time as the mineral phase is being removed, so that the collagen fibers do not collapse. This leads to the formation of an artificial structure consisting of resin-impregnated collagen fibers, which is known as the hybrid layer (Nakabayashi et al. 1982; Wang and Spencer 2003). The hybrid layer is strong, and is also sufficiently hydrophobic to allow the polymer-based composite resins to bond to it. Its depth varies with length of the infiltration step, and lies between 2.1 and 4.1 μm (Nakabayashi and Saimi 1996). In addition to hybrid layer formation, pretreatment of the surface may remove the smear layer and, more important, the smear plugs, thus opening up the tubules of the dentine. Uncured monomer (resin) is then able to flow into these open tubules and form rigid tags on polymerization. This provides an additional mechanism of retention. However, studies have shown that it is the formation of the hybrid layer that makes the biggest contribution to the bond strengths and that even resin systems that are too viscous to flow into the tubules and form resin tags are capable of generating high strength adhesive bonds (Nicholson 2000). Pretreatment of the cut dentine surface was, for many years, carried out using weak organic acids, such as citric acid or ethylene diamine tetra-acetic acid (EDTA). This was recommended partly on the grounds of the proximity of the pulp. More recently, however, phosphoric acid has been used to etch the surface of dentin, so that similar pretreatments are now recommended for both types of tooth tissue (Nicholson 2000). This was originally called “total etch,” but is now more generally termed “etch-and-rinse” in order to emphasise that the treated surface must be rinsed before placing any further layer on it. In using this approach it is important to ensure

1710

J. W. Nicholson

that the dentin is intact, and that there is no exposure of the pulp, and also that very short exposure times are used for phosphoric acid on dentin. When the etch-and-rinse technique is used, it is important not to dry the tooth surface with air, in order to avoid the collapse of the collagen phase. This was originally known as the “wet” bonding technique, and was developed in the 1990s (Kanca 1996). It was important in simplifying the overall bonding process, and paved the way for the introduction of “etch-and-rinse” bonding agents, now considered one of the two important classes of bonding agent. These are typically based on organic solvents which are hydrophilic enough to interact with a surface containing reasonable amounts of water, such as acetone or ethanol (Tay et al. 1998). They can also dissolve the necessary organic monomers which form the resin layer in the surface of the cut tooth. The question of how to prepare the tooth surface in order to obtain bonds that are strong and durable remains the subject of research, and there is evidence that the most effective technique varies with the type of bonding agent used. There is also the need to have reliable ways of determining success in bonding (Carvalho et al. 2012). Laboratory data on bond strength gives some information on possible clinical outcomes for bonded restorations, but the correlation is still uncertain. One problem is that laboratory studies typically use non-carious tooth surfaces, and these differ significantly in their structure and morphology from surfaces affected by caries (Manolea et al. 2016). Several studies of etching strategies on bond strength and durability have appeared in recent years. Etching remains popular and is also successful with enamel as the substrate using both etch-and-rinse and universal bonding agents (Takamizawa et al. 2016; Antoniazzi et al. 2016). With dentin, recent studies have suggested that dry-bonded etch-and-rinse systems can give strong bonds on both carious and non-carious dentin, at least over a period of a year in vitro (Lenzi et al. 2016). Pre-etching has also been shown to improve the penetration of universal bonding agents into dentin, though this does not lead to improvements in bond strength (Wagner et al. 2014). These recent laboratory findings have not received clinical endorsement. Generally clinical reports suggest that survival rates for bonded composite restorations are good, as described later in this chapter. However the state of the adhesive bond over time in clinical service is not known. It has been suggested that this subject should be given attention and that teeth repaired with composite resin then extracted for periodontal reasons could be examined carefully to determine the state of the bond, and in this way identify the characteristics of successful bonding agents (Carvalho et al. 2012).

57.4

Bonding Agents

Bonding agent is the term generally used in the field of adhesive dentistry to describe the substances applied to the surface of the tooth for use as an adhesive. Bonding agents are usually applied to the tooth surface by painting or dabbing them on using a

57

Adhesive Dentistry

1711

small brush. Manufacturers have formulated them mainly for application to dentin, and have used a variety of components. Details of formulations are generally not disclosed as they are proprietary, but it is known that the monomer 2-hydroxyethyl methacrylate (HEMA) is widely used in these blended materials. Other monomers known to be used include gluteraldehyde and mono- and dimethacrylates. As we have seen, acetone or ethanol, and also water may be employed as solvents for these systems. The earliest systems consisted of three components, namely, (1) conditioning agent to fully or partially remove the smear layer, (2) primer for pretreating the dentin surface, and (3) bonding agent that was applied immediately before resin filling. Results with these systems were good, and have not so far been surpassed by any other systems. In clinical use, though, the use of such complicated systems was time consuming and not easy to tolerate by the patient, and so alternatives have been sought. Modern formulations may combine primer and adhesive components in a single blend, in order to reduce the number of steps involved in the bonding process and to simplify the procedure for clinicians. Despite the widespread use of this approach, some sort of separate acid-etch process is required as the first step of bonding. This may be as a separate etching step, as in the etch-and-rinse approach, or as a combined etching/priming step using self-etching primers. Etch-and-rinse adhesives are recommended for use with wet-bonding approach, to maintain the integrity of the collagen fibrils. Liquid bonding agent can be rubbed into the dentin surface as an alternative to brushing (Zander-Grande et al. 2011) in order to ensure good infiltration of the bonding agent into the tooth surface and to maintain adequate moisture content. Self-etching primers are formulated differently from etch-and-rinse systems, and are acidic (Tay and Pashley 2001). They both etch and prime the surface in a single step without the need to rinse away the products of etching. Eliminating the rinse step has the advantage of reducing the technique sensitivity of the bonding process. This is because there is less chance of the clinician making mistakes during what is a very sensitive procedure. Self-etching formulations are acidic because they contain acid-functional monomers, such as maleic acid. As a result of their acidity, these self-etching primers incorporate the components of the smear layer, a feature that cannot be avoided, as there is no rinse step to wash away the acid treated smear layer. However, this may not be desirable and there is evidence that retention of the smear layer components compromises the bond strength of these primers (Pashley and Carvalho 1997). The acidity of self-etching primers varies depending on their composition, and these formulations may be classified as either weak or strong (Van Meerbeek et al. 2001). The mild self-etch primers have a pH of around 2, and dissolve the dentine surface only partially. A significant amount of mineral phase remains in the resulting hybrid layer and is bonded by carboxylate or phosphate groups from the acidfunctional monomers in the primer formulation. These have been shown to give good clinical outcomes over periods of time of up to eight years (Peumans et al. 2010). By contrast, the strong self-etch primers have a pH as applied of around 1 or

1712

J. W. Nicholson

below. This gives them a more aggressive interaction with the smear layer and causes them to generate a bonded surface that closely resembles the surface produced by the etch-and-rinse approach. As well as etch-and-rinse and self-etch bonding agents, there are some brands available that can be used in either mode and are described as “universal” (Sezinando 2014). The most extensively reported universal bonding agent is Scotchbond™ Universal Adhesive (3 M Oral Care) and this material has been found to give equivalent micro-tensile bond strengths in both etch-and-rinse and self-etch modes (Hanabusa et al. 2012, Munoz et al. 2013). The moisture content of the dentin substrate did not affect bond strengths, probably because the liquid bonding agent itself contains a high proportion of water (D’Hoore et al. 2001). Clinical outcomes with this bonding agent have also been shown to be good (Perdigao et al. 2014). Within the field of adhesive dentistry, bonding agents have been classified by generation (Dunn 2003). This classification considers that Buonocore’s original approach led to the first generation bonding agent (Buonocore 1955) and that the most recent type is the seventh generation bonding agent. This is a single-layer, one-component self-etching system. More helpfully, bonding agents have recently been classified by number of steps involved (Nicholson and Czarnecka 2016). Since this is based on a fundamental aspect of their use, it is considered more helpful than classification by generations, and is shown in Table 2. Detailed information on the composition of the various components of bonding agents is not available, due to considerations of commercial confidentiality. However, the various formulations are generally made up of appropriate blends of monomers that vary in how hydrophilic or hydrophobic they are. Resulting films tend to allow some passage of dentinal fluid through them (Tay and Pashley 2003) and there is some evidence that this may assist hydrolytic degradation in service. Unfortunately, little data is available on the long-term clinical durability of these

Table 2 Classification of modern bonding agents for dentistry (Nicholson and Czarnecka 2016) Number of steps 4

Number of bottles 3

3

2

2

2

1

1

Type Description Comments Etch-and-rinse Etch, rinse, apply primer, Complicated to use but then bonding agent results in the highest reliable bond strengths. Etch-and-rinse Etch, rinse, then apply Agent for priming and prime & bond in one step bonding is in a single bottle. Gives reasonable bond strengths. Self-etch Apply self-etching Self-etching primer primer, then bonding incorporates smear layer. agent Self-etch A single formulation Easiest to use. Gives comprising blended self- acceptable bond strengths. etching primer and bonding agent

57

Adhesive Dentistry

1713

bonding agents, and most publications deal with short-term bond strengths, so the extent to which this interaction with dentinal fluid promotes degradation not clear. Bonding studies have typically reported data for shear bond strengths, though there have also been a number of studies using tensile bond strengths. Values reported vary somewhat, as has been shown, depending on the nature of the particular bonding agent. Reported results typically lie in the range 15–25 MPa for shear bond strengths (Al-Ehaideb and Mohammed 2000; El Kalla and Garcia-Godoy 1998; May et al. 1997), though higher values have been reported. Tensile bond strengths, by contrast, typically lie around 10 MPa (May et al. 1997).

57.5

Evaluation of Bonding Agents

Clinical trials are the most appropriate tests of the effectiveness of bonding agents. However, they are time consuming and in vitro laboratory tests are more often used instead. Such tests are typically carried out on human teeth that have been extracted for orthodontic reasons or on bovine teeth, and shear bond strength and micro-tensile bond strengths are widely used as indicators of bonding effectiveness. Results from these experiments need to be evaluated with care. For one thing, given the complex nature of the surface pretreatment processes, bond strength values are not simply material properties. Rather, they reflect a complex interaction between the strength of the composite resin bonded with the bonding agent, the nature of the bonding agent, the stress rate, and the sample size and its overall geometry (De Munk et al. 2005). Despite the need for caution in interpreting these results, useful information can be obtained from this type of test. For example, information on durability and on the effect of thermal cycling is best obtained from such tests.

57.6

Leakage and Bonding Agents

A clinical issue with bonding agents is the extent to which they have the ability to create a complete seal at the interface with the tooth. This is important to prevent exposure of the dentin to fluids from the mouth. Such exposure may lead to further decay beneath the restoration and may ultimately damage the pulp. Larger voids may allow bacteria to migrate into the space and this may also lead to further decay and associated damage. The most widely studied of the types of leakage is known as microleakage. It is defined as the clinically undetectable passage of bacteria or fluids between a cavity wall and the restorative material (Kidd 1976). All of the bonding agents currently available for clinical use seem able to prevent this phenomenon and to seal the restorative margins fully. Microleakage has conventionally been studied in vitro using dye penetration techniques, in which an organic dye dissolved in water is used to highlight the path of any imperfections through the interface between the tooth and the bonding agent/composite resin. This is a broadly satisfactory approach, although there can be

1714

J. W. Nicholson

problems with it, as leakage may occur through the dentin rather than through the interface. Even where there are no gaps capable of admitting a dye solution between the bonding agent and the substrate, some leakage of dentinal fluid has been found to occur (Sano et al. 1995; Tay and Pashley 2003). It is known as nano-leakage and is detectable with Ag+ ions in solution, typically as silver nitrate. Silver ions are much smaller than the dye molecules, and they will pass across an interface that is able to exclude dye solutions. Since its discovery as a phenomenon, nano-leakage has been widely studied in bonding agents and shown to take place through regions of enhanced permeability within the matrix of the composite resin. Any voids associated with nano-leakage are too small to admit bacteria and so it is doubtful if they are in themselves sources of further tooth decay or other immediate clinical problems. However, they may well act as sites of hydrolytic degradation of the bonding agent and may thus grow to the point at which bacteria and their metabolic products may be admitted. This then leads to potential damage to the underlying tooth tissue.

57.7

Composite Resins

The materials which are retained in place using the bonding agents described in the preceding sections are the dental composite resins. Although all types of aesthetic dental repair materials are composites as defined in materials science, within dentistry the term is restricted to a particular group of materials (Nicholson 2000). These are based on relatively high molar mass monomers, such as bisGMA, an adduct of bisphenol A and glycidyl methacrylate, or urethane dimethacrylate (UDMA). In formulating composite resins for clinical use, these macromonomers may be diluted with other less viscous components, such as ethylene glycol dimethacrylate, and filled with an unreactive filler, such as silica or a barium silicate glass. Stress transfer between matrix and filler is aided by the use of a coupling agent, typically γ-methacroyloxy propyl trimethoxy silane. Modern composite resins are cured by application of visible light at 470 nm, which activates a photo-sensitive initiator such as camphorquinone. The camphorquinone readily dissociates to form free radicals when exposed to visible light, and these bring about a cure reaction, converting the highly viscous paste to a hard, durable solid material. Typical material properties of these materials are shown in Table 3. The advantage of composite resins is that they are esthetic (i.e., tooth-colored), and they are also durable. Reports of durability suggest that they are capable of surviving between 7 and 10 years in most patients, though this varies with the general state of oral hygiene of the patient. Composite resins have an important disadvantage, though, and that is that, because they set by an addition polymerization mechanism, they are prone to shrink on setting. This may cause problems of leakage around the restoration, and also movement of the cusps of the tooth where bonding is particularly strong. These

57

Adhesive Dentistry

Table 3 Properties of composite resins (from McCabe and Walls 2008)

1715 Property Main monomer Filler Compressive strength Tensile strength Flexural strength Modulus of elasticity

Typical components/values bisGMA or UDMA 50–78 wt% glass or quartz 260–300 MPa 40–50 MPa 80–150 MPa 6–14 GPa

problems are minimized in clinical use by curing the composite resin in increments, gradually building up a full restoration layer by layer. Within the family of composite resins are the polyacid-modified composite resins, so-called compomers (Nicholson 2006). These include components similar to those of glass-ionomer cements, that is, an acid-leachable glass as part of the filler phase, and a small proportion of acid-functional monomer. After setting by addition polymerization, these materials are able to draw in small amounts of moisture in the mouth, and this promotes a secondary acid–base reaction within the matrix. Studies suggest that this is of little practical benefit, though it does help liberate fluoride from the reactive glass filler, and thus allow these materials to release a small amount of fluoride steadily over prolonged periods of time. Like the conventional composite resins, polyacid-modified composites are used in conjunction with bonding agents in order for them to be retained adhesively within the tooth.

57.8

Durability of Bonded Composite Resin Restorations

The durability of bonds formed by bonding agents to enamel and dentin is critical to the clinical success of bonded composite restorations. Many studies have concentrated on short-term results only, typically as short as 24 h, and have neglected the fact that these bonds change with time. This is caused by a variety of factors (Breschi et al. 2008). The first factor to be taken into account in considering bond durability is the ageing of the hybrid layer. This arises for a variety of reasons. There are natural variations in the temperature in the mouth, due to the consumption of hot and cold foods and beverages, and these affect the properties of the bond over time. The bond may also be affected by the consumption of acid foodstuffs, because these affect the hybrid layer by attacking the unprotected collagen fibers chemically. Lastly, the resin itself may undergo degradation in situ, resulting in loss of bond strength. Such degradation is typically hydrolytic in character and occurs due to uptake of water by the bonding layer, followed by reaction of the water with ester groups in the resin. Eventually, this leads to loss of portions of bonding agent, and an opening of the bonding layer to further infiltration and hydrolytic degradation. Since many formulations of bonding agents combine hydrophilic or even ionic monomers with hydrophobic components, cured bonding agents may admit water

1716

J. W. Nicholson

fairly easily. Indeed, they may behave as semipermeable membranes, forming distinct water channels within them in response to the ingress of the surrounding saliva. Water movement through bonding agents can be driven by osmotic pressure gradients arising because there are inorganic ions and hydrophilic monomers present, and in extreme cases this may result in the formation of water blisters over the adhesive layer (Breschi et al. 2008). There is evidence that the new and simpler adhesives are less predictable in their long-term performance than the more established multi-step systems. Various suggestions have been made for improving their performance, of which the most practical appear to be (a) use of a hydrophobic coating and (b) application with extended cure time. This latter step improves the degree of polymerization and reduces the permeability of the bonding layer. It is thus a straightforward way of improving the long-term performance of modern adhesives.

57.9

Biocompatibility of Bonding Agents

Biocompatibility is defined as “the ability to perform with an appropriate host response in a specific location” (Williams 1987). Its study is complex and beyond the scope of this chapter, but it needs to be considered briefly in the context of the use of bonding agents in clinical dentistry. The tissues involved range in sensitivity from the completely inert (enamel) to the extremely delicate (the pulp), and their response to the range of types of chemical involved in bonding, including acids, monomers, and solvents, varies widely. An important feature controlling the biocompatibility of bonding agents is the permeability of the dentin and also its thickness. This will determine whether or not monomers diffuse all the way through to the pulp and whether they arrive at sufficiently high concentrations to cause damage. Of particular concern in these studies has been the presence of 2 hydroxethyl methacrylate (HEMA). This has been shown experimentally to be able to diffuse through the pulp chamber when released from a bonding agent placed beneath a composite resin (Hamid and Hume 1997). Thickness made a difference to the rate of diffusion, but even relatively thick layers of dentin (i.e., in the range 3.4–3.6 mm) were not sufficient to eliminate it completely. Another finding of significance was that diffusion of HEMA was much more rapid in teeth that had been severely affected by caries. In other words, the reason that tooth repair is needed clinically, caries, turns out to be responsible for affecting dentin permeability adversely, and opens the way for potentially harmful monomers to reach the pulp. HEMA and other monomers used in bonding agents have been shown to be cytotoxic (Hume and Gerzina 1996). They inhibit cell proliferation and decrease mitochondrial activity significantly. The latter means that energy transduction processes in the living cells are affected adversely. In a fully healthy tooth, pulp fibroblast cells differentiate into odontoblasts, and these cells contribute to the development of the tooth. Low concentrations of bonding agent monomers have been shown to affect this development. A variety

57

Adhesive Dentistry

1717

of processes can be affected at levels well below toxic concentration, including expression of collagen 1, osteonectin, and dental sialoprotein. This, in turn, affects the formation of mineral structures within the tooth (Bouillaguet et al. 2000). Overall, it is clear that dentin bonding agents show significant levels of cytotoxicity to the cells of the pulp. Clinically, though, there has been little concern about this finding and no reported adverse effects in vivo. Reports of sensitivity, discomfort, or even complete death of the pulp have been notably absent from the clinical literature. It seems that biological testing in vitro may have exaggerated the extent to which bonding agents pose a threat. Release of only small quantities of monomer may mean that only very small amounts reach the pulp itself under clinical conditions. Also, there is fluid circulation within the living pulp, and this may wash away diffusing monomers and potentially reduce the concentrations to below those needed for major cytotoxic damage to occur. Current clinical opinion is strongly supportive of the use of bonding agents in dentistry, despite the volume of data on their possible adverse biological effects.

57.10 Adhesive Dental Restorative Materials In this section, the three naturally dental restorative materials will be considered. These are the zinc polycarboxylate cement and the glass-ionomer cements. The latter come in two broad groups, the conventional and the resin-modified glass-ionomers. In all three cases, they are placed as viscous pastes, pressed into place by the dentist to ensure that any surface irregularities are filled, and allowed to cure. They then harden and are finished by shaping them to the structure of the natural tooth with various carving tools. These materials will each be considered in turn. Zinc polycarboxylate was the first of the adhesive dental restorative materials to be described. It was invented by Smith (1968) who reported that it was fabricated from aqueous polyacrylic acid and deactivated zinc oxide, and that it had excellent adhesion to both dentin and enamel. It was also found to have other desirable properties. For example, it set rapidly with minimal shrinkage, it was bland toward the soft tissues of the mouth, and it was almost completely insoluble in saliva (Smith 1968). Zinc polycarboxylate cement can thus be considered a major development in the field of adhesive dentistry. Typical properties of zinc polycarboxylate cement are shown in Table 4. Sometime after its introduction to the dental profession, formulations of zinc polycarboxylate were made available in which the finely divided zinc oxide is combined with a dried powder of polyacrylic acid. The mixture is activated by the addition of an appropriate amount of water, and the whole mass sets rapidly to a solid, hard mass that is indistinguishable from formulations prepared from solutions of pre-dissolved polyacrylic acid. Set cements are opaque, due to the presence in them of considerable amounts of unreacted zinc oxide powder, which acts as a reinforcing filler. Zinc polycarboxylate cements rapidly found use in clinical restorative dentistry. Applications included cavity lining, luting of crowns, and adhesion of orthodontic

1718 Table 4 Properties of zinc Property polycarboxylate cement Liquid Powder: liquid ratio Setting time Compressive strength

J. W. Nicholson Typical components/values 40–50% polyacrylic acid 2.5–3:1 2.4–4 min 80–100 MPa

brackets (see Sect. 14). They remain in use for these applications, but have been displaced to an extent by glass-ionomer cements. The zinc oxide powder for these cements is modified, partly by mixing with small quantities of magnesium oxide, and partly by heat treatment at temperatures of between 1100  C and 1200  C. This reduces the reactivity toward polyacrylic acid so that the cement does not set too rapidly. Heat treatment causes a slight yellowing of the zinc oxide, due to the loss of a very small proportion of the oxygen. The resulting material is very slightly non-stoichiometric, with a formula corresponding to ZnO (1 x), where x is 70 ppm or less (Greenwood and Earnshaw 1984). The bonding strengths of zinc polycarboxylates have been found to be adequate but not outstanding. Tensile bond strength values of the order of 4–6.5 MPa have been found for them to enamel in in vitro studies (Brantley and Eliades 2001). Tensile bond strengths to dentin are lower, suggesting that bonding takes place primarily to the mineral phase. Actual bond strengths may be higher than these reported figures, because failure is typically clearly cohesive within the cement, rather than adhesive in nature. Studies have shown that bond strength varies with thickness of zinc polycarboxylate cement. One study showed that the adhesive layer thickness was maximum at 205 μm (Akinmade and Nicholson 1995) when bonding stainless steel strips together and determining bond strength in shear. Values ranged from 2.5 to 4 MPa. Zinc polycarboxylate cements not only form adequate bonds to enamel and dentin, they will also bond to base metals used in dentistry, and also to gold. However, there is almost no adhesive bond formed on the surface of porcelain (McCabe and Walls 2008). The next adhesive repair material to become available to the dental profession was the glass-ionomer cement, which was first described by Wilson and Kent in 1972 (Wilson and Kent 1972). These materials are now used in clinical dentistry for a variety of aspects of the repair of teeth damaged by dental decay. Depending on how they are formulated, they may be used as liners and bases or as full restorations, and they may also be formulated for other clinical uses, as shown in Table 5. They are esthetic in that they match the natural color of the tooth and, like the zinc polycarboxylate cement, they are able to bond directly to the surface of the tooth. They have the additional feature of being able to release small amounts of fluoride. This is generally considered beneficial preventing further decay of the repaired tooth, though this point has not been confirmed by clinical studies.

57

Adhesive Dentistry

Table 5 Typical properties Property of conventional glassLiquid ionomer cements Powder: liquid ratio Setting time Compressive strength

1719 Typical components/values 40–50% polyacrylic acid 3:1 2–4 min 150–220 MPa

Glass-ionomers are usually made from polyacrylic acid, although acrylic acid/ maleic acid copolymer is used in some brands as an alternative. Typically, an aqueous solution of this polymer is mixed with a special ion-leachable glass of complex composition. The glass is basic, and consists of an alumino-silicate structure augmented with phosphate and also containing fluoride species, such as CaF2 or cryolite, Na3AlF6. Setting occurs by neutralization, and involves the polymeric acid molecules being cross-linked with multivalent counter-ions (Ca2+, Sr2+, Al3+) after they have been leached from the glass. As indicated, the glasses are complex and often contain two phases, either completely or partially separated. The setting reaction is, therefore, also complicated and involves not only the formation of ionically cross-linked polymer chains but also hydration processes, the details of which are not yet clear. There is also evidence that an inorganic network forming species from the glass, consisting of phosphate moieties, contributes to the strengthening processes that occur in the cement matrix. Whatever the detail of the setting reactions, they take place fairly rapidly so that typically after about 5 min the cement hardens and can then be finished by the dentist. Subsequently, some slow reactions take place that bring about maturation. This maturation causes gradual increases in translucency and compressive strength and takes place over several weeks following the initial hardening. The final product is a rigid material that has a porcelain-like appearance. As well as the original glass-ionomer cement described above, there is another variation known as the resin-modified glass-ionomer cement (RMGIC). Like the conventional glass-ionomer, RMGIC contains a basic ion-leachable glass powder, and also a water-soluble polymeric acid. However, this type of material also contains water-soluble organic monomer, usually 2-hydroxyethyl methacrylate (HEMA). An initiator system is also required to bring about additional polymerization of the HEMA (Mitra 1991a). In most RMGICs the initiator system is photosensitive, and most RMGICs are light cured. Initiators used are sensitive to blue light of wavelength 470 nm, which is the main light emitted by conventional dental curing lamps. Some brands of RMGIC contain modified polymers. Typically, these are based on poly (acrylic acid) but contain a minority of branches that are terminated in vinyl groups. These undergo copolymerization with the HEMA on irradiation, which means that the final material contains organic as well as ionic cross-links when cured. Typical properties are given in Table 6. Like conventional glass-ionomers, RMGICs form strong adhesive bonds to enamel and dentine (Mitra 1991b) and also release fluoride. The presence of

1720

J. W. Nicholson

Table 6 Typical properties of resin-modified glass-ionomer cement Property Liquid Powder:liquid ratio Setting time Compressive strength

Typical components/values 40–50% polyacrylic acid, plus 2-hydroxyethyl methacrylate (HEMA) 3:1 30 s (on application of curing lamp) 150–180 MPa

HEMA causes the set RMGIC to have some hydrogel character, which means they are capable of swelling in water. In saliva, any such swelling occurs only slightly, if at all, so this does not pose any clinical problems. RMGICs were originally developed for use as liners and bases in dentistry (Wilson 1990), but have since been formulated for use as complete restoratives. They can also be used for core build-up and for luting, though for the latter application, they are not photocured. Rather, they are cured with a two-part initiator that forms free radicals on mixing, so that they are able to set in the dark. They are widely used, and are reported to be reliable materials capable of giving good results in service. They have been particularly used in pediatric dentistry as alternatives to amalgam. Based on observations from HEMA-based dental bonding agents, the presence of HEMA has been suggested to compromise the biocompatibility of RMGICs (Nicholson and Czarnecka 2008b). HEMA is released from RMGICs especially in the first 24 h. As already mentioned, it is able to diffuse through the dentin and affect the pulp, to which it has been shown to exhibit cytotoxicity. It may also have systemic effects on dental personnel, including contact dermatitis (Geukins and Goossens 2001) and sensitization (Katsumo et al. 1996). Safety precautions have been recommended for dental personnel using these materials, mainly to ensure good ventilation of the work place and to avoid breathing HEMA vapor from the unset RMGIC.

57.11 Bonding of Glass-Ionomer Cements Glass-ionomers of all types are able to form adhesive bonds to the surface of the tooth. In order to do so under clinical conditions, the newly cut enamel and/or dentin surfaces are generally treated with a dilute solution of polyacrylic acid, typically at 10% concentration. This is a much milder treatment than that used with current bonding agents for composite resins and is more akin to pretreatments used with third generation bonding agents. Conditioning with polyacrylic acid results in removal of the smear layer and of the smear plugs, and opens up the dentinal tubules (Fig. 2). It does not, however, lead to loss of much, if any, intact tooth mineral from the surface. The glass-ionomer cement is then applied and allowed to cure. For conventional glass-ionomers, this occurs simply through the acid–base setting reaction already described; for RMGICs, it is generally brought about by exposure to a curing lamp.

57

Adhesive Dentistry

1721

Fig. 2 SEM of dentin conditioned with 10–15 s application of 10% aqueous polyacrylic acid, magnification 900 (Mount 1994)

Restorations of RMGIC are often completed in increments, to offset the polymerization shrinkage, as is the case for composite resins. There have been many published studies of the bond strength of freshly placed glass-ionomers to tooth surfaces. Results obtained are generally relatively low, typically in the range 3–7 MPa when tested in shear. However, this has been shown, in reality, to be a measure of the tensile strength of the glass-ionomer, rather than the shear strength of the bond, as failure tends to be cohesive in nature (Mount 1991). Measured shear bond strengths for RMGICs tend to be slightly higher than for conventional glass-ionomers, a reflection of their higher tensile strengths (Mitra 1991b). Despite the relatively low shear bond strengths, durability of glass-ionomer cements in clinical service is good. Studies have shown that survival rates may exceed those of bonded composite resins (Burgess et al. 1994), though there are problems about undertaking comparisons with the most recently developed materials and systems, since survival data and information on durability take several years to obtain. Studies of marginal leakage with glass-ionomers, both conventional and resin modified, have given variable results. It is therefore not possible to draw conclusions about whether conventional or resin-modified glass-ionomers perform better in this

1722

J. W. Nicholson

regard. However, it is possible to say that both types may exhibit marginal leakage and that neither is completely reliable in sealing enamel margins. Conversely there have been no satisfactory reports of the development of secondary caries below glass-ionomer restorations, which suggests that even if they leak they are sufficiently anticariogenic to prevent further tooth decay. This anticariogenic character may be attributed to the fluoride release in situ. Glass-ionomers have been found to show distinct interactions with the tooth surface over the longer term. In an important paper, Ngo et al. (1997) showed that there was a pronounced zone formed at the interface between glass-ionomer cements and the tooth (Fig. 3). This study involved low-temperature Scanning Electron Microscopy examination of freshly extracted teeth that had been filled with a strontium-based conventional glass-ionomer and allowed to mature for 4 days. The interaction zone that formed within this time was not only clearly visible, but shown in subsequent studies to contain both calcium and strontium. Since there was originally no calcium in the cement and little or no strontium in the surface of the dentin, this result can only mean that there has been some sort of chemical reaction between the cement and the dentin to create this structure. Furthermore, it shows that adhesion of glass-ionomer cements involves an ion-exchange process. Developments in tooth-colored restorative materials have caused some confusion within the dental profession. First, resin-modified glass-ionomers were launched and their composition, in terms of monomer and light-activated initiator, gave them some similarities with composite resins. Soon afterward, the compomers became available. They had a modest but distinct acid-base character which was responsible for maturation reactions within the material. This group of composites thus had some clear glass-ionomer character. The situation was clarified in a paper published in the mid-1990s (McLean et al. 1994), and a system of nomenclature was proposed that has subsequently been widely adopted.

Fig. 3 Low temperature SEM of interaction zone between glass-ionomer cement (top) and dentin (beneath), magnification 4317 (Ngo et al. 1997)

57

Adhesive Dentistry

1723

Recently, this field has been reviewed (Mount et al. 2009) and the conclusion drawn that recent research shows it is the mechanism of bonding that is the most important distinction between the tooth-colored restorative materials. The ionexchange mechanism can only occur in materials with a substantial acid-base setting mechanism, and this only occurs with glass-ionomer cements. No amount of modification can make systems that are fundamentally hydrophobic, and which set exclusive by polymerization, that is, composite resins, capable of undergoing ionexchange bonding. These materials will always need bespoke and discrete bonding agents, of the type previously described.

57.12 ART Technique The clinical technique known as Atraumatic Restorative Treatment (ART) has been developed since the mid-1980s in response to dental clinical needs in Third World countries. It is an important application of adhesive dental materials, specifically conventional glass-ionomer cements, and would not be possible without this type of adhesive material (Frencken et al. 2004). ART involves the removal of carious tooth tissue with hand instruments rather than conventional dental drills. These hand-held instruments take the form of spoon-shaped excavators, and come in graded sizes. Following excavation of the soft carious matter, the cavity is cleaned and filled with freshly mixed glass-ionomer cement. Associated pits and fissures within the tooth are filled at the same time. This technique does not require electricity or clean water, so is particularly suitable for settings in the developing world where these facilities are either absent or unreliable. Adhesive materials are essential for this technique because the tooth can only be excavated to remove soft matter, not drilled to extend the cavities with undercuts. This sort of mechanical retention is required by silver amalgam fillings, and cannot be deployed in regions where the drill cannot be used. The hand-operated excavators used in ART are not capable of removing undamaged enamel or dentin, so cannot be used to prepare undercut cavities. Clinical reports covering 25 years experience have shown that ART is an effective treatment in a variety of countries of the world (Frencken et al. 2012). Modern conventional glass-ionomer cements for use with this technique have been formulated with optimized particle size and size-distribution glass powders, and they are found to perform very well in patients. Long survival times are being recorded, with almost no problems of debonding in clinical service.

57.13 Glass-Ionomer Fissure Sealants Fissure sealing is a preventive technique carried out on newly erupted teeth. Deep fissures in such teeth may act as sites into which plaque and pellicle may become occluded, and it cannot be removed from these sites by brushing. Left undisturbed, this is likely to result in the development of caries.

1724

J. W. Nicholson

So-called sealing of these fissures involves minor clinical procedures. First, the plaque and the pellicle are removed using a dental bur, after which the fissure is widened slightly but only within the enamel. To complete the process, some sort of filling material is placed in the newly opened fissure. For some time, the material of choice was considered to be an unfilled resin, essentially a low viscosity blend of the type of monomers used in conventional composite resins. Such fissure sealants have good retention, which is obviously highly desirable. However, there is beginning to be a shift in clinical opinion that favors glassionomers for this application. Initial results with these materials were not promising, for a variety of reasons. Their early bond strengths are lower than those of resin systems, and they had inferior wear resistance. This meant that they appeared to be lost fairly rapidly. Careful study shows, though, that this does not mean that the treated teeth are no longer protected. Teeth treated in this way do not develop caries and protection may last for at least 6 years (Pardi et al. 2003). This surprising result occurs for two reasons. First, glass-ionomer fissure sealants release fluoride in amounts that are protective to the surrounding tooth tissue. Second, not all of a glass-ionomer sealant is actually lost. Deep within the fissures, glass-ionomer sealants are retained and not only retained, but also transformed over time. Changes which occur include increasing translucency, improvement in smoothness and in hardness, and elevation in the calcium and phosphate ion concentrations in the surface layer. These alterations take of the order of 2–3 years to develop and seem to be akin to the development of artificial enamel (Van Duinen et al. 2004). It must be emphasized that this structural change in the glass-ionomer fissure sealant only occurs in the base of the fissures. The rest of the glass-ionomer typically disappears within a few months. The modified glass-ionomer tends to show integration with the adjacent enamel, with excellent sealing. There is no occurrence of caries and every appearance of strong adhesion, though no bond strength data have yet been reported. These findings are similar to those previously discussed for glass-ionomer restorative materials, which also develop strong adhesive bonds with time and show elevated calcium and phosphate levels in the surface following prolonged contact with saliva.

57.14 Bonding in Orthodontics Orthodontics is that branch of clinical dentistry concerned with correctly aligning crooked teeth. This is typically carried out during the teenage years, and involves applying brackets to several of the teeth, and connecting them with high-tensile strength wires. Successive visits to the orthodontist are made, and in each the tension in the wires is increased, thus applying a force that causes the tooth to move within the socket and to be retained there permanently. This is due to changes in the shape of the supporting bone around the socket. Brackets are typically retained on the tooth, either mechanically, or by bonding.

57

Adhesive Dentistry

1725

The earliest attempts to bond orthodontic brackets used epoxy resins. However, these adhesives were difficult to use, and so this approach did not find much support within the dental profession. In modern orthodontic practice, bonding agents similar to those used in other parts of dentistry are used. This means the deployment of either bonded composite resins or glass-ionomer cements. Direct bonding of orthodontic brackets uses the acid-etch technique, with phosphoric acid gel to etch the surface and deliver the initial pretreatment (Kitayama et al. 2007). This is followed by the application of bonding agents or of lightly filled or completely unfilled resins to the surface. These form attachments to the surface mechanically by flowing into the irregularities created in the etching step. A clinical problem with this technique is that brackets bonded by composite resins in this way have a tendency to develop regions of demineralization around them. This leads to the formation of white spots on the tooth surface. Typically where white spot lesions develop, they affect several teeth of a patient. Another problem is that, when the time comes to remove the brackets, damage may occur to the surface of the tooth. This is because the bond strength generated by this technique is so high that the enamel shears in preference to debonding. To overcome the problems of “white spot” lesions, and also to exploit the lower inherent bond strength, glass-ionomer cements are increasingly being used for this application. Their ability to release fluoride ensures that there are no sites of demineralization, hence no white spot formation. There is also much less damage when the brackets are removed (Charles 1998).

57.15 Conclusions This chapter describes the principal uses of adhesives in clinical dentistry, and covers both the materials and the clinical techniques involved. The subject is of enormous importance in modern dentistry and is connected with the growing interest in the use of tooth-colored repair materials. The two types of tooth-colored material, namely, the composite resin and the glass-ionomer cement, are described, together with their current variations. Composite resins are bonded with specific bonding agents, and these substances may be used in either etch-and-rinse of self-etch modes. This technology continues to evolve rapidly, and recent advances in the field are covered in detail. Surface pretreatment is critical for the success with bonding and current understanding of this topic is reviewed as well. Glass-ionomers are naturally adhesive to both dentin and enamel, though some sort of surface pretreatment is necessary prior to their use. Typically this consists of conditioning the freshly cut tooth surface with dilute aqueous polyacrylic acid, followed by rinsing away the products. Glass-ionomers then bond firmly to the tooth, partly through the development of a mechanically strong ion-exchange layer at the interface that develops with time. This bonding ability is exploited in the important technique of ART, created for use in Third World countries to treat the

1726

J. W. Nicholson

growing problem of tooth decay. Glass-ionomers are also used as bonding agents for orthodontic brackets. Zinc polycarboxylate cement, the first adhesive dental restorative material, is also described in this chapter, though it now has only niche applications in dentistry.

References Akinmade A, Nicholson JW (1995) Effect of adhesive layer thickness on the bond strength of a zinc polycarboxylate dental cement. Biomaterials 15:149 Al-Ehaideb A, Mohammed H (2000) Hear bond strength of “one-bottle” dentin adhesives. J Prosthet Dent 84:408 Antoniazzi BF, Nicolos GF, Lenzi TL, Soares FZM, Rocha RO (2016) Selective acid etching improves the bond strength of universal adhesive to sound and demineralized enamel of primary teeth. J Adhes Dent 18:311 Brantley W, Eliades T (2001) Orthodontic materials, scientific and clinical aspects. Thieme, Stuttgart Breschi L, Mazzoni A, Ruggeri A, Cadenaro M, Lenarda R, De Stafano DE (2008) Dental adhesion review: aging and stability of bonded surfaces. Dent Mater 24:90 Bouillaguet S, Wataha JG, Virgillito M, Gonzalez L, Rakich DR, Meyer JM (2000) Effect of sublethal concentrations of HEMA (2-hydroxyethyl methacrylate) on THP-1 human monocyte macrophages in vitro. Dent Mater 16:213 Buonocore MG (1955) A simple method of increasing the adhesion of acrylic filling materials to enamel surfaces. J Dent Res 34:849 Burgess JO, Norling B, Summit J (1994) Resin ionomer restorative materials: the new generation. J Esthet Dent 6:207 Carvalho RM, Manso AP, Geraldeli S, Tay FR, Pashley DH (2012) Durability of bonds and clinical success of adhesive restorations. Dent Mater 28:72 Charles C (1998) Bonding orthodontic brackets with glass-ionomer cement. Biomaterials 19:589 D’Hoore WRA, Gonthier S, Degrange M, Dejou J (2001) Reliability of in vitro microleakage test: a literature review. J Adhes Dent 3:295 De Munk J, Van Landuyt K, Peumans M, Poitevin A, Lambrechts P, Braem M, Van Meerbeek B (2005) A critical review of the durability of adhesion to tooth tissue: methods and results. J Dent Res 118:132 Dunn JR (2003) iBond ™: the seventh generation, one-bottle dental bonding agent. Compen Cont Dental Educ 24(Suppl 2):14 El Kalla IH, Garcia-Godoy F (1998) Bond strength and interfacial micromorphology of four adhesive systems in primary and permanent molars. J Dent Child 65:169 Frencken JE, van’t Hof MA, van Amerongen WE, Holmgren CJ (2004) Effectiveness of singlesurface ART restorations in the permanent dentition: a meta-analysis. J Dent Res 83:120 Frencken JE, Leal SC, Navarro MF (2012) Twenty-five-year atraumatic restorative treatment (ART) approach: a comprehensive overview. Clin Oral Invest 16:1337 Geukins S, Goossens A (2001) Occupational contact allergy to (meth)acrylates. Contact Dermatol 44:153 Greenwood NN, Earnshaw A (1984) The chemistry of the elements. Pergamon, Oxford Hamid A, Hume WR (1997) Diffusion of resin monomers through human carious dentin in vitro. J Oral Rehabil 24:20 Hanabusa M, Mine A, Kuboki T, Momoi Y, Van Ende A, Van Meerbeek B, De Munck J (2012) Bonding effectiveness of a new multi-mode adhesive to enamel and dentine. J Dent 40:475 Hume WR, Gerzina TM (1996) Bioavailability of components of resin-based materials which are applied to teeth. Crit Rev Oral Biol Med 7:172

57

Adhesive Dentistry

1727

Kanca J (1996) Wet bonding: effect of drying time and distance. Am J Dent 9:273 Katsumo K, Manabe A, Itoh K, Nakamura Y, Wakumoto S, Hisamitsu H, Yoshida T (1996) Contact dermatitis caused by 2-HEMA and GM dentin primer solutions applied to guinea pigs and humans. Dent Mater J 15:22 Kidd EAM (1976) Microleakage: a review. J Dent 4:199 Kitayama S, Nikaido T, Ikeda M, Foxton RM, Tagami J (2007) Enamel bonding of self-etch and phosphoric acid-etch orthodontic adhesive systems. Dent Mater J 26:135 Koibuchi H, Yasuda N, Nakabayashi N (2001) Bonding to dentin with a self-etching primer: the effect of smear layers. Dent Mater 17:122 Lenzi TL, Soares FZM, Raggio DP, Pereira GKR, Rocha RO (2016) Dry-bond etch-and-rinse strategy improves bond longevity of a universal adhesive to sound and artificially-induced caries-affected primary dentin. J Adhes Dent 18:475 Manolea H, Antoniac J, Miculescu M, Rica R, Platon A, Melnicenco R (2016) Variability of the composite resins adhesion with the dental substrate penetration and the used adhesive type. J Adhes Sci Technol 16:1754 May KN, Swift JR, Bayne SC (1997) Bond strengths of a new dentin adhesive system. Am J Dent 10:195 McCabe JF, Walls AWG (2008) Applied dental materials, 9th edn. Blackwell, Oxford McLean JW, Nicholson JW, Wilson AD (1994) Proposed nomenclature for glass-ionomer cements and related materials. Quintessence Int 25:587 Mitra SB (1991a) Adhesion to dentin and physical properties of a light-cured glass-ionomer liner/base. J Dent Res 70:72 Mitra SB (1991b) In vitro fluoride release from a light-cured glass-ionomer liner/base. J Dent Res 70:75 Mount GJ (1991) Making the most of glass ionomer cements. Dent Update 18:276 Mount GJ (1994) A color atlas of glass ionomer cements, 2nd edn. Dunitz, London Mount GJ (2002) A color atlas of glass ionomer cements, 3rd edn. Dunitz, London Mount GJ, Hume WR (2005) Preservation and restoration of teeth, 2nd edn. Knowledge Books and Software, Middlesbrough Mount GJ, Tyas MJ, Ferracane JL, Nicholson JW, Berg JH, Simonsen RJ, Ngo H (2009) A revised classification for direct tooth-colored restorative materials. Quintessence Int 40:691 Munoz M, Sezinando A, Luque-Martinez I, Szesz A, Reis A, Loguercio AD, Bombarda NH (2013) Immediate bonding properties of universal adhesives to dentine. J Dent 41:404 Nakabayashi N, Kojima K, Masuhara E (1982) The promotion of adhesion by the infiltration of monomers into tooth substrates. J Biomed Mater Res 16:265 Nakabayashi N, Saimi Y (1996) Bonding to intact dentin. J Dent Res 75:1706 Ngo H, Mount GJ, Peters MRCB (1997) A study of glass-ionomer cement and its interface with the enamel and dentin using a low-temperature, high-resolution scanning electron microscope technique. Quintessence Int 28:63 Nicholson JW (2000) Adhesive dental materials and their durability. Int J Adhes Adhes 20:11 Nicholson JW (2006) Polyacid modified composite resins (“compomers”) and their use in clinical dentistry. Dent Mater 23:615 Nicholson JW, Czarnecka B (2008a) Fluoride in dentistry and dental restoratives. In: Tressaud A, Haufe G (eds) Fluorine and health. Elsevier, Amsterdam/Boston Nicholson JW, Czarnecka B (2008b) Biocompatibility of resin-modified glass-ionomer dental cement. Dent Mater 24:1702 Nicholson JW, Czarnecka B (2016) Dental adhesives. In: Materials for the Direct Restoration of Teeth. Woodhead Publishing, Duxford Ogata M, Harada N, Yamaguchi S, Nakajima M, Periera PN, Tagami J (2001) Effect of different burs on bond strengths of self-etching primer systems. Oper Dent 26:375 Pardi V, Periera AC, Mialhe FL, Meneghim MC, Ambrossana GM (2003) A 5-year evaluation of two glass-ionomer cements used as fissure sealants. Community Dent Oral Epidemiol 31:386 Pashley DH, Carvalho RM (1997) Dentine permeability and dentine adhesive. J Dent 25:355

1728

J. W. Nicholson

Perdigao J, Kose C, Mena-Serrano AP, De Paula EA, Tay LY, Reis A, Loguercio AD (2014) A new universal simplified adhesive: 18-month clinical evaluation. Oper Dent 39:113 Peumans M, De Munck J, Van Landuyt KL, Poitevin A, Lambrechts P, Van Meerbeek B (2010) Eight-year clinical evaluation of a 2-step self-etch adhesive with and without selective enamel etching. Dent Mater 26:1176 Roulet J-F, Vanherle G (2004) Adhesive technology for restorative dentistry. Quintessence, Chicago Sano H, Takatsu T, Ciucchi B, Horner JA, Matthews JA, Pashley DH (1995) Nanoleakage: leakage within the hybrid layer. Oper Dent 20:18 Sezinando A (2014) Looking for the ideal adhesive – a review. Rev Port Estomatol Med Dent Cir Maxillofac 55:194 Smith DC (1968) A new dental cement. Br Dent J 125:381 Takamizawa T, Barkmeier WW, Tsujimoto A, Endo H, Tsuchiya K, Erikson RL, Latta MA, Miyazaki M (2016) Influence of pre-etching times on fatigue strength of self-etch adhesives to enamel. J Adhes Dent 18:501 Tay FR, Gwinnett AJ, Wei SHI (1998) Relation between water content in acetone/alcohol-based primer and interfacial ultrastructure. J Dent 26:147 Tay FR, Pashley DH (2001) Aggressiveness of contemporary self-etching systems. I: Depth of penetration beyond dentin smear layers. Dent Mater 17:296 Tay FR, Pashley DH (2003) Water treeing: a potential mechanism for degradation of dentin adhesives. Am J Dent 16:6 Van Duinen RNB, Davidson CL, De Gee AJ, Feilzer AJ (2004) In situ transformation of glassionomer into an enamel-like material. Am J Dent 17:223 Van Meerbeek B, Vargas M, Inoue S, Yoshida Y, Peumans M, Lambrechts P, Vanherle G (2001) Adhesives and cements to promote preservation dentistry. Oper Dent 26:S119 Wagner A, Wendler M, Petschelt A, Belli R, Lohbauer U (2014) Bonding performance of universal adhesives in different etching modes. J Dent 42:800 Wang Y, Spencer P (2003) Hybridization efficiency of the adhesive/dentin interface with wet bonding. J Dent Res 82:141 Williams DF (1987) Definitions in biomaterials. Elsevier, Amsterdam Wilson AD (1990) Resin-modified glass-ionomer cements. Int J Prosthodont 3:425 Wilson AD, Kent BE (1972) A new translucent cement for dentistry. The glass-ionomer Cement. Br Dent J 132:133 Zander-Grande C, Ferreira SQ, da Costa TRF, Loguercio RA (2011) Application of etch-and-rinse adhesives on dry and rewet dentin under rubbing action: a 24-month clinical evaluation. J Am Dent Assoc 142:828

Adhesion in Medicine

58

Robin A. Chivers

Contents 58.1 58.2 58.3

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mechanical Fastening . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesives Technologies for Internal Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.2 Requirements of an Adhesive for Internal Use . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.3 Categories of Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.4 Synthetic Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.5 Biological Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.3.6 Laser Welding and Soldering . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.4 Adhesives for External Application to the Skin . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1730 1730 1731 1731 1732 1734 1734 1739 1744 1744 1747 1747

Abstract

Adhesives are increasingly being used in medicine for repairing cuts and tears in the body as an alternative to mechanical fixation such as sutures. There are very strict regulations controlling the application of adhesives within the body and this means that there are only a limited number of chemistries which are approved for clinical use. The principal internal adhesives in current use are those based on fibrin, gelatin, and poly(ethylene glycol) (PEG) hydrogels and function largely as sealants, though do bond to the tissue surfaces. Cyanoacrylates are approved, mostly only for application to the surface of the skin for wound closure, but are quite widely used in a range of other applications. There are many requirements of an adhesive for internal use and this has led to large number of further systems being proposed. A number of those currently being evaluated for internal R. A. Chivers (*) York, North Yorkshire, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_57

1729

1730

R. A. Chivers

indications are described. Some are based on synthetic chemistry, some from biological sources such as marine creatures, and some are akin to welding or soldering. Pressure-sensitive adhesives are widely used for securing dressings and devices to the skin surface. Most of these are now based on acrylic or silicone chemistries.

58.1

Introduction

In simplest terms, the human body is made up of organs and tissues held together in a way which maintains their mechanical integrity and ability to function correctly. However, there are occasions when, due to accident, surgery or wear, some of these tissues or organs fail. In many cases, the body is capable of repairing naturally but sometimes needs assistance with this. Surgery has developed to such an extent that it is frequently possible to refasten or repair the failure such that the tissue can repair itself and regain its normal function, albeit in a scarred state. Initially, these surgical procedures consisted of provision of a mechanical fixation, but more recently a range of adhesive products and techniques have been developed to replace or augment these, analogously to the substitution of adhesives for nails, screws, bolts and rivets in engineering. This chapter aims to summarize the current state of this technology – adhesion and adhesives for repairing the human body. These may be adhesives for repair of internal organs, or for topical application to the skin, alone or on dressings. They are distinguished as adhesives for contact with body tissue, applied for the most part by medical professionals, though with one exception: dentistry is not included. The use of adhesives is well established in dentistry and will not be covered further here as it is the subject of ▶ Chap. 57, “Adhesive Dentistry.” There are a number of other aspects of adhesion in medicine which will only be touched on here. Adhesives are used widely in the manufacture of medical devices. Some of these applications require the adhesive to be in contact with body material, such as recirculating blood, and are therefore subject to some of the strict regulations which are imposed on products with body contact (Tavakoli et al. 2005). Other aspects of adhesion include the aggregation of cells and proteins onto foreign surfaces within the body, forming biofilms, the desired adhesion of cells in culture onto scaffolds for tissue engineering in the regeneration of organs, and the undesirable adhesion, known as “adhesions” between scarred tissues which can produce serious complications after surgery.

58.2

Mechanical Fastening

Traditionally, repair of damage to the body has been performed by surgeons using mechanical means. The most familiar of these is the suture, a thread of resorbable or non-resorbable polymer, which is stitched through soft tissue using a specially designed needle. Pulling the tissue edges together encourages the formation of a

58

Adhesion in Medicine

1731

bond between them as the tissue regrows. When repair is complete, the surgeon may return to remove the sutures or they may degrade and eventually be excreted by the body. Suturing is usually fairly straightforward and medical professionals are trained in its use. It can also be easily removed and reapplied should there be a problem. It has the disadvantage, though, of only holding the tissue at discrete points, leaving less well-secured gaps between these which may be susceptible to leakage (see, e.g., Fig. 1a). More recent developments in soft tissue fixation include metallic staples and barbed polymeric darts, and when fixation is required to hard tissue, there are ranges of interference screws available. Securing pieces of bone mechanically is usually done using plates, held in place with screws, or by the use of long intramedullary nails which can be inserted down the medullary canal in the long bones, and then held fast by transverse screws at the ends. Metallic repair means may be surgically removed after the tissue has healed, but are frequently left in place if they do not cause a problem for the patient. Externally, the skin may be sutured or bandaged to give support during natural healing. However, crude adhesives, such as tar, have long been used to protect sites of severe damage.

58.3

Adhesives Technologies for Internal Applications

58.3.1 Overview From the time of the introduction of the first adhesive used in an internal application, fibrin sealant, some clinicians have been seeking adhesive solutions to problems in the internal repair of the body, aiming to achieve excellent performance without the problems related to the use of the traditional mechanical means of repair such as localized fixation and the need for removal. Chemists and biologists have been keen to support them in this quest and have produced a wide range of different materials

Fig. 1 Surgical repair of the dura. (a) Mechanical fastening, suturing, does not result in a watertight closure. (b) Adjunctive use of a patch to the closure seals the gap (Images provided courtesy of Tissuemed Ltd.)

1732

R. A. Chivers

with the ability to bond one or more tissue types. Initial experimentation is performed on ex vivo tissue and many materials show some promise in that environment. The ability to bond is, naturally, essential, but is not the only requirement, and many systems which showed initial promise proved not to work when applied in vivo or gave too many undesirable side effects to be permitted for human use. Beyond that, other systems may be successful technically but not be of wideenough application, or appeal to enough clinical users and purchasers, to be a commercial success and so may be withdrawn from the market. The technical, techno-commercial, and medical literature has many examples of initially promising materials which are not available for clinical use. This section aims to describe some of the requirements for a tissue adhesive for internal application and then to give further details on the principal materials which are commercially available and clinically used (though not necessarily permitted in all countries). All materials are described generically as chemical species, without brand names and manufacturers’ details as these can frequently change and may also differ between countries, potentially leading to confusion. In addition, some of the more promising and distinct developments of potential novel surgical adhesives are described, though usually in less detail. There can be no guarantee that any of these will ever be successful, but it is to be hoped that some will and that further inventors may be inspired to come forward and test new ideas as the subject is still very much alive and there are still many unmet needs for this technology within the clinical community.

58.3.2 Requirements of an Adhesive for Internal Use The body provides a very distinct environment in which to apply adhesives. There are therefore many special requirements placed on any material considered for this application. Internal use requires stricter provisions than for adhesives which are purely in contact with intact skin. First and foremost they have to be safe. Materials sticking to the body cannot contain chemicals which are toxic to the body or to the cells in their vicinity. In addition, the material must not cause sensitization either for the patient or for the user. Nor must it produce an immune response – rejection – from the body. Safety testing must be comprehensively performed before new chemicals are permitted in these applications. As with any device or material intended for use in medicine, medical adhesives cannot be sold for clinical use without regulatory clearance from the relevant national authority. This is to ensure, as far as possible, the safety of the product for the designated indication or indications. Different countries impose different requirements on products, which also depend on the nature of the device or material and its intended use. These frequently require extensive testing, both preclinically on animals and clinically in carefully controlled studies on humans, as well as the provision of other safety data. Because of the different regulations, products may be found on sale in one country but restricted in others.

58

Adhesion in Medicine

1733

The material must not introduce organisms which might be harmful for the body or for the repair. It must therefore be intrinsically sterile (as produced) or capable of being adequately sterilized. Sterilization is performed by heat (autoclaving), chemical and heat (ethylene oxide), or ionizing radiation (gamma rays), all of which involve the application of much energy. This energy will often be sufficient to initiate chemical reactions, such as the curing of adhesives, so developers of medical adhesives have to include means of ensuring that the adhesives are protected from any possibility of this. Clearly, a medical adhesive must be capable of reliable secure bonding of the tissue that is to be bonded. Not all tissue is the same. An early misapprehension in the development of medical adhesives was that one adhesive would work everywhere. Subsequent experience has proved that this is not the case. Different formulations have been developed for different tissue types. The moisture content, the fat content, and the protein and polysaccharide content differ, and bone also has mineral content (Duck 1990). It has been found that the strongest bonds occur when there is true chemical bonding to the substrate material (Wilson et al. 2005). Almost universally, the area to be bonded is wet, which is a challenge for many conventional adhesives, and even if it can be temporarily dried at the time of application, the bond will soon be wet again. The bond must form rapidly as it is unsafe and inefficient for surgery to take longer than strictly necessary. In most cases, a surgeon will opt for a quick action rather than a prolonged surgery. However, it is possible for bonding to occur too rapidly, leading to fixation of an incorrect placement. Related is the concern over adhesive spills or leakage from the joint. It should ideally be possible to see and remove safely any adhesive which gets in the wrong place, though the nature of the adhesive and means of dispensing it should be designed so as to avoid this happening at all. Adhesive products are sometimes colored (usually with a shade of blue), so that they may be distinguished from the surrounding tissue. Figure 5 shows an example of this. The purpose of most tissue adhesives is to hold the tissue together so that fluids (e.g., blood) do not leak and it is able to heal naturally. Healing may be a slow process, taking several days, so the bond must last as long as the process, ultimately degrading as the tissue is able to regain the necessary strength. Healing, though, will only occur if the adhesive has not formed a barrier to its progress. Several medical adhesives are therefore applied as “patches” or “sleeves,” bridging the gap in tissue, while not penetrating between the tissue surfaces (Fig. 1b). The breakdown products of the adhesive must be safe to the patient and not harmful to the repair process. In general, this requirement is more likely to be met by adhesives of a biological origin, but synthetic and semisynthetic adhesives may also be safe, or the small quantities of material and slow release rate may lead to breakdown products being present in such small concentrations as not to be a problem. In a few cases, it may be necessary for the medical adhesive to provide a permanent fixation, and not to degrade away, if the tissue is not capable of regeneration. As has already been indicated, packaging and delivery of the adhesive is of key importance. The packaging must keep the material sterile and, for multicomponent

1734

R. A. Chivers

materials, separate. It must be easy to use, and, to remove the possibility of crossinfection by multiple uses, contain the volumes needed to achieve a single repair on one patient, although a second application may be needed in some cases. Application must be foolproof, simple (surgeons’ hands may be required to perform many functions), and permit accurate delivery with minimal spillage. Where two components are concerned, the way in which they are mixed, for example, through a mixing nozzle, multi-jet spray, or by sequential delivery, must be reliable and not lead to problems, such as premature setting, clogging the delivery.

58.3.3 Categories of Adhesives Medical adhesives may be categorized in various ways. Clinicians would probably choose listing by the indications for which they may be used. In this chapter, adhesives have been divided into different groups depending on the similarities of the chemistry and technology. There are separate sections below for adhesives which are produced entirely by synthetic chemistry and for those of biological origin. Naturally, there is some crossover where synthetic molecules are used with biological products, and some “biological” materials are being produced by identification of the functional components of a natural product and then reproducing these by a synthetic route. Laser welding and soldering are somewhat different means of producing adhesion between tissues and are covered in a separate section.

58.3.4 Synthetic Adhesives Cyanoacrylates Cyanoacrylates are widely used in a range of engineering applications where instant strong bonding is required when using a minimum of adhesive material. The adhesive only requires a small amount of water on the surfaces in order to set off the polymerization reaction that forms the bond. These features are clearly attractive in a medical context, so it is not surprising that the use of cyanoacrylates in surgery (of small blood vessels) was attempted before 1960. A number of problems were then discovered which have led to only slow adoption over the intervening years, though recently this has gathered pace, particularly after 1998, when the US Food and Drug Administration (FDA) first approved cyanoacrylate products for specific indications, initially external to the body and, more recently, for a few very specific internal applications. The fundamental structure of a cyanoacrylate is given in Fig. 2. The R group can be one of many different species, though usually alkyl, giving rise to a large family of different molecules with different properties. The adhesive is a low viscosity liquid which cures by anionic polymerization, and the process is exothermic. The reaction can be set off by hydroxyl ions, or similar chemistry, such as hydroxyl or amine groups present in proteins in the tissue surface. Thus, cyanoacrylate is capable of bonding covalently to the tissue surface (Wilson et al. 2005) which contributes, along with its

58

Adhesion in Medicine

1735 H

CN C

H

C COOR

Fig. 2 The structure of a cyanoacrylate monomer. R can be selected from many different species; see text for details

ease of penetration into the structure of the surface before curing, to the very strong bonds which it can make to most tissue types (Chivers and Wolowacz 1997). The choice of alkyl group in the molecule is very important. Domestic and industrial “superglues” are usually methyl or ethyl cyanoacrylates. These were initially tried in medical applications and found to be unsatisfactory. The curing reaction was too fast, leading to tissue necrosis due to overheating and the rate of subsequent degradation was also too rapid (Woodward et al. 1965). Larger alkyl groups were found to be more successful and the most commonly used medical cyanoacrylates are currently n-butyl2-cyanoacrylate and octyl-2-cyanoacrylate. In addition to these molecules, formulations include stabilizers to permit sterilization and to reduce the likelihood of polymerization during storage, and thickeners, usually poly(methyl methacrylate) (PMMA) or polymerized cyanoacrylate. The hardness and inflexibility of cured cyanoacrylate can be reduced by the use of other additives. Low curing rates of the octyl- monomer, which otherwise has the advantage of giving a more flexible polymer, can be improved by adding accelerators at the point of delivery. In the early days of their use, medical cyanoacrylate adhesives received bad publicity due to tissue damage at the site of use. This may have been due to the exothermic curing reaction but may also have been due to the degradation products of the polymer (Leonard et al. 1966). Cyanoacrylates degrade hydrolytically by a reverse Knoevenagel reaction to formaldehyde and alkyl cyanoacetate. Concerns about formaldehyde released into the body have led to many considering it as the major cause of the problem, and several routes have been attempted to reduce the likelihood of its release by altering the chemistry or adding agents to capture it (e.g., Leung and Clark 1994). However, as formaldehyde release is only a potential problem within the body, cyanoacrylates have become widely used and approved as a topical means of skin closure in accident and emergency situations, where the adhesive is delivered across the laceration (not in between opposing edges) using careful control of the flow. This process is shown in Fig. 3. The result can be good secure holding of the wound, excellent sealing to provide a barrier to microbial ingress, and also a good cosmesis after the adhesive has sloughed off the surface and the wound is healed (Quinn et al. 1997). Figure 4 shows a comparison of wounds from laparoscopic surgery closed using sutures and using cyanoacrylate. It is only recently that the US FDA has approved this application as a Class III device, for topical use only, initially in 1998 for octyl-2-cyanoacrylate and later (in 2002) for nbutyl-2-cyanoacrylate (Mattamal 2005). Cyanoacrylates have been widely used experimentally in a great number of surgical procedures, particularly in eye and ear surgery, which are largely external to the body

1736

R. A. Chivers

Fig. 3 Application of a cyanoacrylate adhesive to close a skin wound (Published with permission of MedLogic Global Ltd.)

Fig. 4 Comparison of wounds post laparoscopic surgery closed (a) using sutures and (b) with cyanoacrylate (Published with permission of MedLogic Global Ltd.)

(Quinn 2005). One internal application has also been approved by the US FDA. This is for a formulation of which the major ingredient is n-butyl-2-cyanoacrylate. The application is for neurologic embolization in the skull (Mattamal 2005). A cyanoacrylate adhesive, with a special delivery device, has recently been launched in several countries as a fixation means for polypropylene hernia mesh. It is believed that cyanoacrylate adhesives are used off-label for a range of other internal conditions.

58

Adhesion in Medicine

1737

Modified Gelatin An early commercial medical adhesive was made from gelatin, resorcinol and formaldehyde and is therefore known as GRF (Braunwald et al. 1966). The gelatin is a bioresorbable polymer which can be cross-linked with formaldehyde, and this may also crosslink to the tissue surfaces. Resorcinol can also be cross-linked by formaldehyde, making a stronger bond. However, on enzymatic degradation, the formaldehyde is released. The issues with formaldehyde which are perceived with cyanoacrylates are therefore also present here, but this has not stopped the material from being widely used, though not in the USA, particularly in aortic dissection repairs and also in liver and kidney surgery. The gelatin-resorcinol mixture is a viscous paste which has to be heated to 45 C before application and then mixed with an aqueous solution of formaldehyde, usually containing some glutaraldehyde as well, as it is delivered to the tissue. Concerns about the formaldehyde have led to several modifications of this system. In a prominent commercial product, it is replaced by a mixture of glutardialdehyde (pentane-1,5-dial) and glyoxal (ethanedial) (Ennker et al. 1994a, b). This formulation is also used for aortic dissections. A further alternative is to replace the gelatin with another protein, albumin, and an adhesive sealant is commercially available containing 45% albumin solution which is mixed with 10% glutaraldehyde solution as it is delivered to the repair site. This is successful particularly for hemostasis in cardiac and vascular surgery (Chao and Torchiana 2003). Polyurethanes Polyurethanes are formed from the reaction of diisocyanates and diols to make a prepolymer, usually a paste, which can then further polymerize in moist conditions to give a cross-linked strong material. As the body provides a moist environment for bonding, it is not surprising that many groups have attempted to make polyurethane adhesives for medical applications. Indeed, these were some of the earliest surgical adhesives to be explored and were considered particularly for bonding bone. Obviously, the materials have to be able to degrade to safe products, and therefore lactides and caprolactones have featured in the formulations. However, these have not made a commercial success, because of adverse reactions and because of difficulties in obtaining suitable reaction rates. Most are found to cure too slowly. Recently, several new urethane systems have been developed with chain extenders derived from lysine or other hydrolyzable, bioresorbable molecules (Gilbert et al. 2008). These are now approved for sale in Europe and the USA for the approximation of tissue layers where subcutaneous dead space exists between the tissue planes in abdominoplasty. Synthetic Hydrogels Much body tissue is effectively a hydrogel. Any adhesive used in the body has to be compatible with the high water content and mechanical properties of tissue, so it is not surprising that hydrogels have been widely used for tissue bonding and

1738

R. A. Chivers

sealing. Many of these are based on poly(ethylene glycol) (PEG), which may be cross-linked by a number of means after derivatization to make these reactive, for example, with tetra-succinimidyl and tetra-thiol residues (Wallace et al. 2001), to give the necessary cohesive strength. Two-component systems have been commercialized for sealant applications, such as for incisions in the dura mater, lungs, and blood vessels (Bennett et al. 2003). The two components, PEG and a crosslinker, trilysine, both with reactive end groups and in separate solutions each with very low viscosity, are delivered together, often by spraying. The reaction is very rapid, producing a film which seals over the defect and bonds mechanically to the tissue surface by penetration of the precursors into its crevices, followed by curing. Figure 5 shows the result of application of such a system during lung surgery. The material breaks down over time in the body and the components are excreted. Hydrogels may alternatively be cross-linked by derivatization with photocurable monomers. This has the advantage of curing only when the light is shone and the disadvantage of working only in “line of sight” of the light source (Sawhney et al. 1993). A further series of experimental systems use an aminated star PEG cross-linked with dextran aldehydes. The aldehyde not only cross-links the PEG but also bonds chemically to the tissue surface, so varying the aldehyde content can give a range of bond strengths. It is claimed that this chemistry can thereby be tailored to create adhesives for different tissue types (Artzi et al. 2009). Several workers have been attempting to mimic the performance of fibrins without the need to use blood-based materials with their attendant risks. These formulations include gelatin cross-linked with a calcium-independent microbial transglutaminase (McDermott et al. 2004) and other gelatin-based materials.

Fig. 5 Use of a sprayed two-component hydrogel sealant during lung surgery. The sealant (in the lower middle of the picture) contains blue dye to enable the surgeon to see where it is and to judge the thickness of the application (Copyright # 2010 Covidien. All rights reserved. Used with the permission of Covidien)

58

Adhesion in Medicine

1739

Other Synthetic Adhesive Materials Most of the adhesives discussed here for surgical use are delivered in the form of one or more fluids which cure (react) on the tissue surface to produce a bond and also cross-link to develop cohesive strength. An alternative approach is to make a dry film which already has mechanical integrity and which is coated with materials which react on contact with the wet tissue surface to give a strong bond (Thompson 2009). This therefore has some similarities with the wound dressings which are discussed later in this chapter, but there are numerous differences as these commercial “patches” are intended for internal application for sealing leaks of fluid, for example, air leaks in the lungs, blood leaks and cerebrospinal fluid leaks through the dura, but not for holding the tissue in place. An example of its application to seal the dura is shown in Fig. 1b where the sutures holding the tissue together may also be clearly seen through the film. A resorbable biocompatible film of poly(DL-lactideco-glycolide) is coated with a layer of a terpolymer of polyvinylpyrrolidone (PVP), polyacrylic acid (PAA) and N-hydroxysuccinimide (NHS). This layer becomes rapidly hydrated on making contact with the tissue surface due to the presence of PVP. The PAA gives tack to the film, producing rapid adherence to the tissue, while the NHS is capable of covalent bonding to the tissue surface. A patch like this will degrade in the body over time with the degradation products being excreted, so leaving no residue. Bone contains inorganic material, mostly calcium-containing minerals such as hydroxyapatite. Bone defects are often surgically filled with cements made of a variety of calcium phosphates which are found to enhance the healing of the bone. The bonding of the filler to the bone is mostly by mechanical interlocking, and therefore not very strong, but it has been found that some magnesium-containing mineral compounds can be used as adhesives, making stronger bonds to help with the initial fixation of the bone. These have been used, for example, in the fixation of bone-tendon grafts for the reconstruction of the anterior cruciate ligament (ACL) (Gulotta et al. 2008). Glass ionomer cements have been used for some years in dentistry for bonding restorations to teeth. It has been found in certain circumstances that these can also be used for bonding bone, though as the current formulations contain acrylic polymers, these are not resorbable and can only be used in situations where the bond is intended to be permanent and not replaced by regrowth of the tissue. One such commercial application is for repair of the tiny bones in the ear (Tysome and Harcourt 2005).

58.3.5 Biological Adhesives Fibrin Fibrin adhesives have been used for longer than any other tissue adhesive, being first applied in the mid-twentieth century, although they were not sold commercially until 1978. Frequently referred to as fibrin sealants as their ability to seal leaks is more significant than their mechanical strength, they are widely used in a variety of surgical procedures, and more so in recent years after the approval by the US FDA

1740

R. A. Chivers

of one product in 1998 for certain indications and a second product in 2003 (Spotnitz et al. 2005). Fibrin adhesives contain several constituents. These are fibrinogen, thrombin, factor XIII and calcium ions (usually as calcium chloride) as well as fibrinolytic inhibitors, such as aprotinin, tranexamic acid, or aminocaproic acid (Sierra 1993). The fibrinogen and factor XIII are typically prepared from human blood sources, usually pooled from carefully screened donors and then treated to inactivate any viruses present (MacPhee 1996) either by cryoprecipitation and freeze drying or by solvent detergent cleaning, heat pasteurization and nanofiltration (Spotnitz et al. 2005). The thrombin and aprotinin are sometimes from human and sometimes from bovine sources. Recombinant processes are being sought for these materials. Some products contain the components as lyophilized powders and solutions that have to be mixed just prior to application, while others are supplied as a liquid to be stored at or below 18  C and then gently thawed before use. Fibrinogen may also be obtained by extraction from the patient’s own blood. This is known as autologous fibrin, and is prepared during surgery by taking blood before the operation and centrifuging it to extract fibrinogen. This is naturally somewhat time consuming but has the advantage of eliminating any chance of virus transmission. The viscosity of the preparation and low strength of the resulting adhesive are limitations of material from this source. The chemistry of the process of formation of the adhesive material from the ingredients is given in Fig. 6 and resembles the final stages of the coagulation (clotting) of blood. Essentially, the thrombin cleaves fibrinopeptides A and B from the fibrinogen molecule to leave fibrin. This polymerizes to form a soft clot, which is cross-linked by the transglutaminase factor XIIIa, which is produced by the action of thrombin on factor XIII in the presence of calcium ions (Marx 1996; Webster and West 2001). The initial soft clot formation is relatively quick but produces a weak material. The cross-linking is a slower process, taking typically about 2 h, giving rise to a considerable increase in strength by formation of covalent bonds and possibly contributing some bonding to collagen in the tissue surfaces (Marx 1996;

Fibrinogen

Thrombin

Fibrin monomers

Weak gel +

Factor XIII

Thrombin

Factor XIII’

Ca2+

Factor XIIIa

Crosslinked fibrin gel

Fig. 6 Schematic of the formation of fibrin adhesive from its ingredients. This occurs in-situ at the time of use

58

Adhesion in Medicine

1741

Donkerwolke et al. 1998). In addition, the cross-linking stabilizes the clot to proteolytic degradation by plasmin. It has been found that different concentrations of the different components can affect the time to cross-link and the strength of the resulting bond (Marx 1996). However, as the materials are of natural origin, it is not always possible to control the ratios precisely. As a pure biological adhesive resembling blood clots, the fibrin adhesive is able to break down naturally in the body on a timescale resembling that for the tissue to heal (Lontz et al. 1996). Degradation typically occurs in about 2 weeks (Lauto et al. 2008). Since the bond strength of fibrin adhesives to tissue is not very high (Chivers and Wolowacz 1997), the material is usually used as an adjunct to mechanical fastening such as sutures, hence frequently being referred to as a sealant. It has the advantage of excellent hemostasis and is hence very useful in applications where blood leakage is a problem. In the USA, regulatory approval has been given for several applications where its primary purpose is for hemostasis during surgery on internal organs such as the spleen and liver. It is, however, also used for controlling blood loss during many types of surgery, including orthopedic, and is excellent for sealing blood vessels after suturing and before the blood is readmitted. It may be used for sealing other vessels which do not contain blood, such as lymphatic vessels and the dura containing cerebrospinal fluid, and can be applied to seal an anastomosis (reconnection) of the colon and also air leaks in the lung, although the low strength of fibrin means that special care has to be taken in the latter case when preparing and monitoring the seal. As an adhesive, fibrin has found application in plastic surgery, where it can be used to attach skin grafts to the underlying tissue, though care has to be taken to ensure that it does not act as a barrier to tissue integration under the graft. In these various applications, the bond is most effective if the tissue surfaces are as dry as possible when the fibrin is used, and care must be taken that the bond not be disrupted while cross-linking as pre-cross-linked fibrin can be effective as an antiadhesive (Spotnitz et al. 2005). One of the first applications of fibrin adhesive, one which is still used, is in nerve anastomosis. An additional application of fibrin is as a delivery means for a variety of pharmaceuticals such as antibiotics, because of its ability to adhere to tissue and so deliver the drug locally, as well as its safe degradation pathways (MacPhee et al. 1996; Webster and West 2001).

Modified Fibrin Attempts have been made by several workers to overcome the limitations of fibrin adhesives while maintaining the basic fibrin chemistry. These retain the basic components, unlike those synthetic mimics which have been mentioned above. One way of doing this is to mix fibrillar type I collagen into the fibrinogen/factor XIII solution. When formulated correctly, this enhances the viscosity of the initial solution and therefore its ability to stay at the site where needed, though still permitting easy delivery. In addition, the stiffness and strength of the cross-linked adhesive are enhanced, giving the material greater toughness and ability to withstand loads in use. Effectively it is a fibre-reinforced composite (Sierra 1996). This

1742

R. A. Chivers

technology has been commercialized in a number of formats including: collagen and thrombin to mix with the autologous fibrinogen thus enhancing the otherwise poor properties of autologous fibrin adhesives, collagen fleece coated with lyophilized fibrin adhesive components ready to soak to activate as a patch, and collagen (gelatin) with thrombin which acts as a hemostat with the patient’s own blood (Oz et al. 2003).

Adhesives from Shellfish The ability of shellfish, such as mussels and barnacles, to stick to underwater structures and remain stuck despite considerable mechanical disturbance, such as due to waves and tides, has long attracted scientists looking for candidate adhesives suitable for medical application where the substrates are wet, the temperature may be up to or over 37  C, and there is mechanical stress. Since many synthetic adhesives suffer from an inability to stick underwater, an adhesive material which clearly works in wet conditions becomes particularly attractive. Many shellfish-based systems have been assessed, but the most studied has been the byssus, or attachment thread, of several species of mussel, including Mytilus edulis. This thread of solidified adhesive ends in an adhesive plaque complex which forms a very strong bond to the substrate surface (Waite 1987; Silverman and Roberto 2007; Forooshani and Lee 2017). Adhesives (known as mussel adhesive proteins or MAPs) can be extracted from the mussels, but it takes a huge number to produce sufficient material for commercial use. Various groups have therefore worked on the identification and synthesis of the essential components of the natural protein in an attempt to be able to produce a MAP without resorting to mussels. It has been found that one of the protein residues essential for bonding (initial chemisorption to the surface and then cross-linking of the adhesive) is 3,4-dihydroxy-L-phenylalanine (DOPA). Any material which resembles MAP will contain this. Genetic engineering may be a route to large-scale production of suitable proteins, but so far this has only succeeded in producing DOPA-containing peptides, not the full sequences of native MAPs. Synthetic analogs are being developed by several groups, for example Yu and Deming 1998 prepared simple copolypeptides of DOPA with L-lysine which, when mixed with certain oxidizing agents, gave good bond strengths to several nonbiological surfaces. An alternative approach has been developed (Lee et al. 2006; Burke et al. 2007), in which the DOPA residues are chemically coupled to PEG molecules and crosslink these rapidly when in the correct oxidation state. The PEG makes up about 80–95% of the formulation which can lead to undesirable side-effects as it swells considerably in an aqueous environment. Further developments include a DOPA-containing block copolymer gel which exhibits a sol-gel transition between room and body temperature and a DOPA-containing monomer (with an acrylate group) which can be photopolymerized (Lee et al. 2006). Recent work has concentrated on putting a wide variety of functional groups onto the catechol to modify the crosslinking and hence control swelling and the mechanical robustness of the structure while altering its ability to bond to surfaces. Different oxidation states of catechol have been found to be needed for bonding to organic tissues from those required when the substrate, such as bone, contains mineral particles (Forooshani and Lee 2017).

58

Adhesion in Medicine

1743

Other Adhesive Technologies The use of biological molecules has a number of potential advantages in resorption and biocompatibility, though not necessarily in immunogenicity. There is therefore an interest in finding and developing new adhesive systems from biological sources. Many of these are still in the research stage, far from commercialization, and are summarized below to give an idea of the range of ideas being considered. Attempts are being made to produce proteins (resembling silk and elastin) using recombinant technology from synthetic genes. In this way, specific molecular structures can be constructed to have the desired biological effects, such as good elasticity and bonding to tissue surfaces in the body, while, it is claimed, not having the potentially undesirable side effects which can be shown by wholly natural materials (Cappello 1991). A number of organisms produce adhesive materials that serve them in a variety of ways, including defense, attack, and to help create protection. As they are naturally produced, they are biocompatible and usually naturally degrade to safe molecules. They are also able to stick in wet environments. Several of these materials have been assessed for use as medical adhesives with promising results. The sticky secretion from the back of frogs from the genus Notaden has been explored by workers in Australia (Graham et al. 2006). This exudate was extracted by stimulating the backs of the frogs and tested in peeling on ex vivo meniscal tissue. Bond strengths exceeded those produced by GRF and fibrin adhesives but not n-butyl-2-cyanoacrylate, and it was found that the adhesive functioned primarily as a pressure-sensitive adhesive (PSA) in that the strength was largely unaffected on breaking and reforming the joint. Shellfish are not the only organisms using adhesives to ensure that they remain in place despite battery from waves and tides. Red and brown algae also produce adhesives, based on polyphenols, which can bond nonspecifically to both hydrophobic and hydrophilic surfaces under water. These have also been investigated, and synthetic analogues have been made which can successfully mimic this performance and are postulated as a soft tissue adhesive (Bitton and Bianco-Peled 2008). A challenge with any of these is the ability to obtain sufficient for commercial use. These materials are usually made by the organism in very small quantities and harvesting in bulk would be difficult. Preferred approaches are therefore either to identify the essential active molecules and synthesize these (as with mussel adhesion) or to find a way of producing these by using genetic engineering processes. The mechanism of reversible adhesion used by geckos to walk on walls and ceilings is of interest to several research groups which would like to develop applications using synthetic mimics of this. Unfortunately, gecko adhesion, which is the result of van der Waals interactions between large number of nanoscale pillars and a surface, has been found not to be very effective in the wet, reducing its potential in medical applications. A possible way of overcoming this is being developed in which the nanoscale pillars, made of poly(dimethyl siloxane) (PDMS), are coated with a synthetic mimic of mussel adhesive: poly(dopamine methacrylamide-co-methoxyethyl acrylate) (Lee et al. 2007). The inventors believe that this could have applications as a temporary adhesive for use in the wet in medical settings.

1744

R. A. Chivers

58.3.6 Laser Welding and Soldering The use of light to activate adhesives has been mentioned above offering the advantage that the time of activation is precisely controlled, but the disadvantage of only working in the beam of light. Alternative applications of “light” (including infrared and ultraviolet radiation) have been developed for the clinical bonding of tissue. These are known as welding when light alone is used and soldering when a material is cured with the light (Bass et al. 1996; Lauto et al. 2008). When tissue is heated, the proteins in it coagulate to form a hard sticky mass. This is similar to the cooking of egg-white. A laser can provide a precise beam of infrared radiation to the edges of two pieces of soft tissue and can weld these together. This is a quick procedure and does not produce the damage that sutures can, but a hazard of this is that the tissue may become too hot and cells could be damaged. This can be reduced by applying a chromophore, which can preferentially absorb the radiation, converting it to heat. For the usual medical infrared laser wavelength of 808 nm, indocyanine green (ICG) is an approved, safe molecule with a maximum absorbance at the same wavelength. Clinicians have generally found it preferable to apply a solder, a viscous solution of a protein such as albumin or fibronectin (Bass et al. 1996; Chivers 2000; Lauto et al. 2008). When mixed with ICG, this solder is applied over the tissue edges to be bonded and the area is irradiated with an infrared laser beam which is usually pulsed to reduce the net power delivered to the tissue. The enhanced absorption of the ICG enables the solder to reach the coagulation temperature of 65–70  C well before the tissue has become warm, and the proteins are believed to intermingle with the collagen in the tissue to give a relatively strong repair which can degrade naturally over a clinically acceptable timescale. Concerns over the difficulty of applying a runny liquid and keeping it in place and the possibility of it being diluted by water at the surgical site have led to the development of solid solders, strips and patches of high (>53%) albumin concentration, which can be laid on the tissue surface or placed as a short tube over a cut vessel to be resealed and then irradiated to create the bond. As an alternative to proteins, a similar material has recently been created using chitosan, a polysaccharide (Lauto et al. 2008). Other, solder-free welding techniques have been reported. One uses Rose Bengal dye activated by laser light at a wavelength of 514 nm, and another a dye made of 1,8 naphthalimide which can be activated by blue light and bonds to tissue seemingly without any heating.

58.4

Adhesives for External Application to the Skin

Although cyanoacrylates are now being widely used in the clinic for the closure of skin incisions, as described above, the principal adhesives used on the external surface of the body are pressure-sensitive adhesives. These are the only significant adhesives used on the body in which the application is

58

Adhesion in Medicine

1745

frequently not made by a clinician, as they are on first aid dressings and other products for home use. Pressure-sensitive adhesives (PSAs) have long been used to hold tapes, dressings (Fig. 7), and other devices (e.g., electrodes and ostomy pouches) onto the skin. A further application of PSAs for skin contact is to deliver pharmaceutical agents by a transdermal route. Unlike most other applications of adhesives mentioned here, these are purely for topical use and are normally only in contact with intact skin. This greatly simplifies the regulatory approval process and permits a wide range of materials to be used in comparative safety. There are nevertheless some people who are sensitive to chemicals which may be used in adhesives and care must be taken to watch out for this in clinical practice. Early PSAs were usually rubber based, containing natural rubber which needs to be modified by the addition of tackifiers, lanolin, and mineral filler to optimize performance. A traditional tackifier is colophony resin (gum rosin), another natural material. A disadvantage of these materials is that they are very poor at handling moisture. These adhesives have to be physically modified, usually by perforation, so that there is not an excessive moisture build-up under a dressing coated with rubberresin adhesive. Other disadvantages include the lack of resistance to shear, and a tendency to fail in the adhesive mass (cohesive failure), resulting in adhesive

Fig. 7 High performance wound dressings can be held securely in place on the skin using a pressure sensitive adhesive (Image copyright by Smith & Nephew April 2010)

1746

R. A. Chivers

residues being left on the skin around a dressing while being worn, and after removal. A variety of materials have since been used to make more “skin-friendly” adhesives with better mechanical and moisture-handling capabilities. These include, principally, acrylics and silicones, but there are also poly(vinyl ether)s and poly (vinyl pyrrolidone)s for certain specialist applications (Webster and West 2001). Formulations are continually being developed to attempt to improve these further. The process of assessing a PSA for use on skin is extremely complex and relies a lot on subjective assessment. Many properties must be considered and have been widely discussed in the literature (Satas and Satas 1989; Chivers 2001). These include the ability to stick securely and yet be easily and safely removed, the ability to stay in place for several days despite the movement of the underlying skin, and the ability to handle moisture and to “breathe” with the skin. Naturally the adhesive is not used alone and must perform successfully when coated on a suitable backing material, usually a film or fabric. The ability to stick is obvious, as patient health, wound protection and cleanliness can be compromised by a lifting dressing. In contrast, a skin PSA must be able to be removed easily, minimizing discomfort to the patient and risk of damage to skin which may be fragile, as well as ensuring that the adhesive residue is not left on the skin. While it may be thought that the pain experienced when a dressing is removed would be related to the force required to peel the dressing from the skin, it is by no means clear that this is the case. Peeling frequently removes a layer of dead skin cells from the skin surface. It is not known whether this contributes to the discomfort of dressing removal, but it can compromise the ability of the dressing to be repositioned as it can mask the adhesive. Frequent removal of dressings from the same site can cause reddening and even damage to the skin. For that reason, adhesives and systems have been developed which make removal easier (Chivers 2001). Some skinfriendly solvent products are available to help with this, particularly in difficult cases such as ostomy. Shear of the adhesive is also undesirable as it can lead to dressings sliding from the site of attachment and leaving sticky residues exposed on the skin. These can fasten to clothing or bedding. As already mentioned, the moisture handling is important, since human skin loses moisture in the form of water vapor at, typically, 200–500 g/m2/24 h and inability of any skin covering to cope with this leads to maceration, overhydration, of the skin. In extreme cases the dressing can come off if too wet, which it may also become if immersed in water, since the adhesives cannot stick in very wet conditions. Various ways of handling moisture are available – the formulation of the adhesive can help, but how it is spread on the backing is also important. Gaps may be deliberately produced in the adhesive film – by pattern spreading or perforations. Developers of new formulations for medical PSAs have a range of in vitro testing methods which they use to help to evaluate these requirements. Ultimately, though, the performance of PSAs on skin is so subtle and subjective that only testing using the correct backing and on humans from the target population will give a true assessment of the acceptability of a new material.

58

Adhesion in Medicine

58.5

1747

Conclusions

Adhesives technology has two main applications in medicine as considered in this chapter: for internal fixation of tissues usually after surgery, and for use on the skin, primarily to hold dressings in place. There are many differences between the requirements of the technology for these applications. Internal application is very carefully regulated as there is a need for extreme caution on safety grounds. Adhesives for this have to pass a series of tests before approval, and this may not be universal. Many chemical and biological systems have been tried and several are now in widespread clinical use. The appeal of an adhesive over the current mechanical means of fastening tissue has led many developers to attempt to create new materials from a range of sources. Some of these may be successful, but there is a long way to go in many cases. Adhesive technology for external application is, by contrast, mature and many successful systems are available as the regulatory requirements are fewer. This has not stopped developers working to improve the existing products, and indeed to modify them for use on further dressing types and for other applications, such as for transdermal drug delivery.

References Artzi N, Shazly T, Baker AB, Bon A, Edelman ER (2009) Aldehyde-amine chemistry enables modulated biosealants with tissue-specific adhesion. Adv Mater 21:3399–3403 Bass LS, Libutti SK, Kayton ML, Nowygrod R, Treat MR (1996) Soldering is a superior alternative to fibrin sealant. In: Sierra DH, Saltz R (eds) Surgical adhesives and sealants current technology and applications. Lancaster, Technomic, pp 41–46 Bennett SL, Melanson DA, Torchiana DF, Wiseman DM, Sawhney AS (2003) Next-generation hydrogel films as tissue sealants and adhesion barriers. J Card Surg 18:494–499 Bitton R, Bianco-Peled H (2008) Novel biomimetic adhesives based on algae glue. Macromol Biosci 8:393–400 Braunwald NS, Gay W, Tatooles CJ (1966) Evaluation of crosslinked gelatin as a tissue adhesive and hemostatic agent: an experimental study. Surgery 60:857–861 Burke SA, Ritter-Jones M, Lee BP, Messersmith PB (2007) Thermal gelation and tissue adhesion of biomimetic hydrogels. Biomed Mater 2:203–210 Cappello J (1991) Protein engineering for biomaterials applications. Curr Opin Struct Biol 2:582–586 Chao H-H, Torchiana DF (2003) BioGlue: albumin/glutaraldehyde sealant in cardiac surgery. J Card Surg 18:500–503 Chivers RA (2000) In-vitro tissue welding using albumin solder: bond strengths and bonding temperatures Int. J Adhes Adhes 20:179–187 Chivers RA (2001) Easy removal of pressure sensitive adhesives for skin applications Int. J Adhes Adhes 21:381–388 Chivers RA, Wolowacz RG (1997) The strength of adhesive-bonded tissue joints. Int J Adhes Adhes 17:127–132 Donkerwolke M, Burny F, Muster D (1998) Tissues and bone adhesives – historical aspects. Biomaterials 19:1461–1466 Duck FA (1990) Physical properties of tissue – a comprehensive reference book. Academic, London

1748

R. A. Chivers

Ennker J, Ennker IC, Schoon D, Schoon HA, Dörge S, Meissler M, Rimpler M, Herzler R (1994a) The impact of gelatin-resorcinol glue on aortic tissue: a histomorphologic evaluation. J Vasc Surg 20:34–43 Ennker J, Ennker IC, Schoon D, Schoon HA, Rimpler M, Herzler R (1994b) Formaldehyde-free collagen glue in experimental lung gluing. Ann Thorac Surg 57:1622–1627 Forooshani PK, Lee BP (2017) Recent approaches in designing bioadhesive materials inspired by mussel adhesive protein. J Polym Sci A Polym Chem 55:9–33 Gilbert TW, Badylak SF, Gusenoff J, Beckman EJ, Clower DM, Daly P, Rubin JP (2008) Lysinederived urethane surgical adhesive prevents seroma formation in a canine abdominoplasty model. Plast Reconstr Surg 122:95–102 Graham LD, Glattauer V, Peng YY, Vaughan PR, Werkmeister JA, Tyler MJ, Ramshaw JAM (2006) An adhesive secreted by Australian frogs of the genus Notaden. In: Smith AM, Callow JA (eds) Biological adhesives. Springer, Berlin/Heidelberg, pp 207–223 Lauto A, Mawad D, Foster JR (2008) Adhesive biomaterials for tissue reconstruction. J Chem Technol Biotechnol 83:464–472 Lee BP, Dalsin JL, Messersmith PB (2006) Biomimetic adhesive polymers based on mussel adhesive proteins. In: Smith AM, Callow JA (eds) Biological adhesives. Springer, Berlin/ Heidelberg, pp 257–278 Lee H, Lee BP, Messersmith PB (2007) A reversible wet/dry adhesive inspired by mussels and geckos. Nature 448:338–341 Leonard F, Kulkarni RK, Brandes G, Nelson J, Cameron JJ (1966) Synthesis and degradation of poly(alkyl α-cyanoacrylates). J Appl Polym Sci 10:259–272 Leung JV, Clark JC (1994) Biocompatible monomer and polymer compositions. US Patent 5328687 Lontz JF, Verderamo JM, Camac J, Arikan I, Arikan D, Lemole GM (1996) Assessment of restored tissue elasticity in prolonged in vivo animal tissue healing: comparing fibrin sealant to suturing. In: Sierra DH, Saltz R (eds) Surgical adhesives and sealants: current technology and applications. Lancaster, Technomic, pp 79–90 MacPhee MJ (1996) Commercial pooled-source fibrin sealant. In: Sierra DH, Saltz R (eds) Surgical adhesives and sealants: current technology and applications. Lancaster, Technomic, pp 13–18 MacPhee MJ, Singh MP, Brady R Jr, Akhyani N, Liau G, Lasa C Jr, Hue C, Best A, Drohan W (1996) Fibrin sealant: a versatile delivery vehicle for drugs and biologics. In: Sierra DH, Saltz R (eds) Surgical adhesives and sealants: current technology and applications. Lancaster, Technomic, pp 109–120 Marx G (1996) Kinetic and mechanical parameters of fibrin glue. In: Sierra DH, Saltz R (eds) Surgical adhesives and sealants: current technology and applications. Lancaster, Technomic, pp 49–59 Mattamal GJ (2005) US Food and Drug Administration perspective on class I, II and III cyanoacrylate medical devices. In: Quinn JV (ed) Tissue adhesives in clinical medicine, 2nd edn. BC Decker, Hamilton, pp 159–168 McDermott MK, Chen T, Williams CM, Markley KM, Payne GF (2004) Mechanical properties of biomimetic tissue adhesive based on the microbial transglutaminase-catalyzed crosslinking of gelatin. Biomacromolecules 5:1270–1279 Oz MC, Rondinone JF, Shargill NS (2003) Floseal matrix: new generation topical hemostatic sealant. J Card Surg 18:486–493 Quinn JV (2005) Clinical approaches to the use of cyanoacrylate tissue adhesives. In: Quinn JV (ed) Tissue adhesives in clinical medicine, 2nd edn. BC Decker, Hamilton, pp 27–76 Quinn J, Wells G, Sutcliffe T, Jarmuske M, Maw J, Steill I, Johns P (1997) A randomised trial comparing octyl-cyanoacrylate tissue adhesive and sutures in the management of lacerations. JAMA 277:1527–1530 Satas D, Satas AM (1989) Hospital and first aid products. In: Satas D (ed) Handbook of pressure sensitive adhesive technology, 2nd edn. Van Nostrand Reinhold, New York, pp 627–642

58

Adhesion in Medicine

1749

Sawhney AS, Pathak CP, Hubbell JA (1993) Bioerodible hydrogels based on photopolymerised poly(ethylene glycol)-co-poly(α-hydroxy acid) diacrylate macromers. Macromolecules 26:581–587 Sierra DH (1993) Fibrin sealant adhesive systems: a review of their chemistry, material properties and clinical applications. J Biomater Appl 7:309–352 Sierra DH (1996) Fibrin-collagen composite tissue adhesive. In: Sierra DH, Saltz R (eds) Surgical adhesives and sealants: current technology and applications. Technomic, Lancaster, pp 29–39 Silverman HG, Roberto FF (2007) Understanding marine mussel adhesion. Mar Biotechnol 9:661–681 Spotnitz WD, Burks SG, Prabhu R (2005) Fibrin-based adhesives and hemostatic agents. In: Quinn JV (ed) Tissue adhesives in clinical medicine, 2nd edn. BC Decker, Hamilton, pp 77–112 Tavakoli SM, Pullen DA, Dunkerton SB (2005) A review of adhesive bonding techniques for joining medical materials. Assem Autom 25:100–105 Thompson I (2009) Technical bulletins. Tissuemed Ltd. Available at http://www.tissuemed.com Tysome JR, Harcourt J (2005) How we do it: ionomeric cement to attach the stapes prosthesis to the long process of the incus. Clin Otolaryngol 30:458–460 Waite JH (1987) Nature’s underwater adhesive specialist. Int J Adhes Adhes 7:9–14 Wallace DG, Cruise GM, Rhee WM, Schroeder JA, Prior JJ, Ju J, Maroney M, Duronio J, Ngo MH, Estridge T, Cocker GC (2001) A tissue sealant based on reactive multifunctional polyethylene glycol. J Biomed Mater Res (Appl Biomater) 58:545–555 Webster I, West PJ (2001) Adhesives for medical applications. In: Dumitriu S (ed) Polymeric biomaterials, 2nd edn. Marcel Dekker, New York, pp 703–737 Wilson DJ, Chenery DH, Bowring HK, Wilson K, Turner R, Maughan J, West PJ, Ansell CW (2005) Physical and biological properties of a novel siloxane adhesive for soft tissue applications. J Biomater Sci Polym Ed 16:449–472 Woodward SC, Herrmann JB, Cameron JL, Brandes G, Pulaski EJ, Leonard F (1965) Histotoxicity of cyanoacrylate tissue adhesive in the rat. Ann Surg 162:113–122 Yu M, Deming TJ (1998) Synthetic polypeptide mimics of marine adhesives. Macromolecules 31:4739–4745

Recycling and Environmental Aspects

59

Chiaki Sato

Contents 59.1 59.2

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Impact of Adhesion Bonding in the Environmental Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.2.1 Categories of Environmental Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.2.2 Greenhouse Gas Emission . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.2.3 Use of Finite Resources . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.2.4 Basic Strategies to Meet the Challenge for Environmental Issues . . . . . . . . . . . 59.3 Types, Characteristics, and Applications of Dismantlable Adhesives . . . . . . . . . . . . . . . . . 59.3.1 Thermoplastic Adhesives and hot-Melt Adhesives . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.2 Adhesives Including Blowing Agents or Expansion Agents . . . . . . . . . . . . . . . . . 59.3.3 Adhesives Including Chemically Active Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.4 Adhesives to Which an Electrochemical Reaction on Interfaces Is Applied . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.5 Miscellaneous Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.6 Recent Advances . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.3.7 Future Seeds . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1752 1753 1753 1754 1754 1754 1757 1757 1758 1764 1764 1766 1768 1771 1773 1773

Abstract

Recycling and environmental aspects of adhesion technology are discussed in this chapter. Adhesively bonded adherends should be often separated before they can be recycled. For this purpose, dismantlable adhesives, which can be separated with stimulations, have been developed recently. There are lots of technical processes to realize such adhesives. For example, softening of adhesive, expansion force due to blowing agents or thermally expandable microcapsules, chemical degradation, electrochemical reaction, and taking advantage of interfacial C. Sato (*) Precision and Intelligence Laboratory, Tokyo Institute of Technology, Midori-ku, Yokohama, Japan e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_58

1751

1752

C. Sato

phenomena can be applied to dismantlable adhesives. These adhesives have many applications such as temporary joints of work materials, shoe making, integrated circuits (IC) chip fabrication, electronics, housing construction, car fabrication, and so on. The most used trigger to start the separation of these adhesives is heating. Many kinds of methods including infrared, induction, and microwave heating are recently available. In addition, novel dismantlable adhesives, which have anisotropy in strength or are triggered by electric currents, have been presented.

59.1

Introduction

Since adhesive bonding has many advantages such as low cost, light weight, easy application, etc., it is increasingly being used by modern industries. Adhesive bonding is more desirable than the other joining methods in terms of energy saving, so that its environmental load is small. However, the consumption of adhesives has been increasing drastically due to the industry globalization and the recent fastgrowing of developing countries, and the total environmental load of adhesive bonding is also increasing. To meet environmental issues, the total environmental loads, although they are small, have to be reduced by all possible means. Adhesive bonding has many aspects related to the environmental issue. One of them is energy consumption in some processes such as production, transportation, application, and curing of adhesives. Another one is recycling of adhesively bonded materials. Strictly speaking, there are other aspects such as recycling of adhesive itself and environmental contamination including vaporized organic chemical compounds (VOC). This chapter treats the recycling issue of adhesively bonded materials. The recycling of the adhesive itself is not dealt with because it is very difficult so far. Adhesively bonded joints are not easy to separate if they are bonded with strong adhesives or the bonded area is large. Therefore, dismantlability of the joints is not required contrarily to the other joining methods such as mechanical fastening. The fact leads to scarce use of adhesion for substrates that have to be separated. However, a new type of adhesive called a “dismantlable adhesive” has been recently required to meet the demands to separate the adherend for recycling. For instance, a dismantlable design for a product is essential because the materials bonded together become bad wastes difficult to be recycled. However, if the materials can be separated, they become resources. The separation of parts is important even if they are bonded adhesively. Thus, dismantlable adhesives, which can be separated when and as we like, are increasingly required. An advantage of adhesive bonding is the ability to join dissimilar materials. However, it becomes a disadvantage if such materials cannot be recycled. Dismantlable adhesives are also appropriate to solve this problem. Dismantlable adhesives, however, are thought difficult to realize because a good adhesive has been apprehended as strong and not being able to separate so far. The

59

Recycling and Environmental Aspects

1753

effort to realize a good adhesive has been concentrated into this point. The capability to separate when we like is a totally different concept from the past efforts. As a matter of fact, dismantlable adhesives, if they are weak in strength, have been present since fairly early times. Hot-melt adhesives, for instance, can be separated by heating. However, strong adhesives, especially structural adhesives, present a high technical hurdle to be dismantlable. The trend has been changing for the last decade and many dismantlable adhesives have been proposed as a kind of novel functional adhesives. Although their applications are mainly temporary joints of works or product reworks so far, adherend recycling has become an issue recently. In addition, the adhesives are applied to make separation mechanisms for aircrafts, space crafts, or robots. This chapter deals with dismantlable adhesives, whose development has proceeded very fast recently, and shows the results of recent research and technical trends.

59.2

Impact of Adhesion Bonding in the Environmental Issues

59.2.1 Categories of Environmental Issues The main environmental issues can be categorized in three following items: 1. Hazardous material emission 2. Greenhouse gas emission 3. Use of finite resources In this chapter, hazardous material emission is not treated (see ▶ Chap. 39, “Environment and Safety”). However, these items are not independent from each other and it is difficult to discuss them separately. Thus, the first item will be shortly referred in this chapter too. Global warming is the most serious and important issue among environmental issues. There is a lot of theory or controversy on the effect brought by the problem and the time limit that we have for preventing vital results. However, it is certain that the final result is devastating to the global environment unless we can reduce, or sustain in the worst case, greenhouse gas emission. In this case, irreversible warming must occur within some decades. Although it is very rare for adhesives to emit greenhouse gases directly, the gases, mainly carbon dioxide, are possibly emitted during their production, application, curing, or disposal processes. The third item, use of finite resources, may seem different from environmental issues. However, this should be included in environmental issues because the most important finite resource is fossil fuels such as petrol, and their combustion causes carbon dioxide. In addition, avoiding the use of finite resources without their recycling is indispensable to realize sustainable society. Form this viewpoint, this item is also discussed in this chapter.

1754

C. Sato

59.2.2 Greenhouse Gas Emission The main greenhouse gas emitted during the processes of adhesive production, use, and disposal is carbon dioxide. In the production process of adhesives, energy due to fossil fuel may be used and this causes carbon dioxide. Energy should be used in the curing process of adhesives too. Since the raw materials of adhesives are mainly made from petrol, carbon dioxide should also be produced by incineration in the disposal process. In addition, the other greenhouse gases such as hydrocarbons or organic solvents may be emitted to the atmosphere during the production or application process of adhesives. Such greenhouse gases have higher warming effects per unit than carbon dioxide. Therefore, even their total amount of emission is small, attention should be paid.

59.2.3 Use of Finite Resources Some raw materials of adhesives are derived from petrol. Therefore, if petrol resources are depleted, the production of adhesive cannot be maintained. This depletion is an ultimate situation, but even the price rise of petrol can give serious and negative impact to adhesive industries. An alternative option is the use of other fossil carbon resources such as coal. Coal is more abundant than petrol in terms of reserve. However, its carbon dioxide emission per unit is larger than that of petrol. Anyway, true sustainable societies cannot be realized while finite fossil resources such as petrol or coal are widely and often used. Energy consumed in the production process of adhesive and its curing process may be derived from fossil fuels too.

59.2.4 Basic Strategies to Meet the Challenge for Environmental Issues The basic strategies to solve the environmental problems can be categorized as follows: 1. 2. 3. 4. 5.

Use of eco-conscious materials Provide reworkability or dismantlability to adhesives Recycling of adhesives itself Optimization of the bonding process Design optimization and lifetime cycle assessment of adhesively bonded joints In this section, they are explained respectively.

Use of Eco-Conscious Materials Selection of raw materials for adhesives is very important to reduce environmental loads. At first, the raw materials should be environment friendly. Synthetic polymers

59

Recycling and Environmental Aspects

1755

have relatively larger energy consumption in this production process than other materials such as fillers, including alumina powders, silica aerogels, calcium carbonate or talc consisting of minerals. Therefore, the composition of an adhesive is influential to the total environmental loads. In addition, if an adhesive is burned in the disposal process of the product bonded with it, carbon dioxide is emitted. As seen above, the presence of fillers can help to reduce the carbon dioxide emission because the portion of polymer in the adhesive decreases. In order to reduce more effectively the environmental loads, use of recycled materials or carbon-neutral natural materials should be considered. Reclaimed rubbers are often used as the raw material of Pressure Sensitive Adhesive (PSA). Polystyrene foam used as shipping supplies can be solved with solvents and used as a raw material for adhesives. As such, recycled materials can be used to produce adhesives for the purposes of reducing environmental impacts and costs. Use of carbon-neutral natural materials is another option. Since they absorb carbon dioxide from the air during their glowing, the total amount of carbon dioxide in the atmosphere is not changed even if these materials are burned. Historically, natural materials are mainly used as raw materials for adhesives. For instance, glue, lacquer, and casein have been widely used to make adhesives (Fay 2005). They are natural materials and carbon-neutral. The other example of natural material use is the case of PSA. To make PSA, natural rubbers for the main component and rosins for tackifiers, which are derived from plants, are often used. They are also carbonneutral. Recently, other natural materials have been investigated for raw materials of adhesives. For example, polylactide resin can be applied to the main components of adhesives (Viljanmaa et al. 2002). Lignophenolic resins derived from lumber (Kadota et al. 2004) and chitosan resin from exoskeleton of crustacean are other examples (Yamada et al. 2000; Umemura et al. 2003). Soy beans oils can be epoxidized and used for adhesive main components (Ratna and Banthia 2000). The powder of lumbers such as cork is promising as fillers for adhesives because it can give ductility to the cured bulk of adhesive (Jos and Leite 2007).

Reworkability or Dismantlability of Adhesives The key word “3R” is often referred when environmental issues are discussed. This word means “reduce,” “reuse,” and “recycle.” In order to reuse or recycle adhesively bonded materials, joint separation is often required. To meet this demand, a new type of adhesive called “dismantlable adhesive,” which can be separated on demand, has been invented. Defective products due to malfunctioned parts bonded adhesively or to inappropriate adhesion are very bad in terms of environment protection because they become a waste without any use. Therefore, dismantlable adhesive can be used to avoid the situation because bonded parts or inappropriate bonded areas can be separated in order to be repaired. This process is called “reworking” of defect products and involved in the processes of “reducing” wastes and “reusing” other nondefective parts in terms of 3R. This topic will be discussed in more detail in Sect. 2.2.

1756

C. Sato

Recycling of the Adhesive Itself Actual cases of adhesive recycling are still very rare because the amount of adhesive used in a product is much smaller than that of the adherend materials. In addition, adhesives consist usually of various kinds of materials. On the other hand, removing an adhesive layer from substrate is very difficult. Resolving obtained adhesive wastes by any solvent is also not easy because they are fully cured and crosslinked. However, the needs of sealant recycling are recently becoming important because the total amount of its use is huge. Thus, we should start to consider seriously the method to recycle sealant wastes. Optimization of the Bonding Process If the bonding process is not optimized, large environmental loads occur. For instance, although the surfaces of adherends often need pretreatment before bonding, there is room to reduce environmental loads. The use of oil accommodating adhesives is effective for the purpose because degrease and chemical treatments of adherends are not necessary. Otherwise, if solvents are used for degreasing, they are hazardous and their atmospheric release is undesirable. Collection systems of solvent vapor are useful for reducing the atmospheric release. The collected solvent vapor can be used again after condensation or thermally recycled, although carbon dioxide is emitted in this case. It is also effective to reduce sub materials used in adhesion processes such as wiping cloths, gloves, masking tapes, removing agents, mixing cups, brushes, and static mixing nozzles. Since most of them are disposable, much waste is made unless the bonding process is optimized. In addition, release sheets of PSA become a waste too. Energy consumption to cure adhesives is also a problem. Capabilities of ambient temperature curing or relatively low temperature curing are desirable to be installed for adhesives. Some acrylic adhesives can be cured quickly at room temperatures and they are strong and tough enough to be used for structural applications. A promising epoxy adhesive for structural use, which is ambient temperature curable and heat resistant due to nano-fillers, has been recently presented (Sprenger et al. 2004). Design Optimization and Life Cycle Assessment of Adhesively Bonded Joints Design optimization of adhesively bonded joints can contribute to environmental issues in a broad sense because it reduces the used materials and the weight of the joints. Precise knowledge of adhesives strength, modulus, and stress distribution in joints is helpful for designers, and it leads to rational design of the joints to minimize environmental loads. This point of view is likely to be missed, but very important. Adhesive making companies ought to provide positively the data to design joints appropriately. This is important not only for the optimal design, but also for preventing the duplication of basic tests that may be carried out by the adhesive maker and each user.

59

Recycling and Environmental Aspects

1757

The other obligation of adhesive makers is disclosure of information for Life Cycle Assessment (LCA) of adhesively bonded joints. If these data are provided, the users of an adhesive can calculate the environmental impact associated with equivalent energy consumption or equivalent carbon dioxide emission. The later is called “carbon footprint” whose requests by users is recently increasing, and adhesives are not an exception. The users can evaluate the environmental impacts by LCA, and make rational decision to employ adhesive joining or not considering the total tradeoff of the whole life-cycle environmental loads of their products.

59.3

Types, Characteristics, and Applications of Dismantlable Adhesives

The main purpose of dismantlable adhesives is reworking of parts or recycling of materials in a product. Temporary joining of works also needs dismantlable adhesives. Reworking of parts is a process in which defective parts are removed and exchanged by normally functioning parts. If the parts are vitally important, defects of the parts have a huge influence on the product value. If the parts are joined with a dismantlable adhesive, they can be separated and exchanged easily, and that leads to an increased yield ratio of the product. Electronic devices have integrated circuit (IC) chips on their circuit boards bonded with epoxy adhesives called under fillers. Recently, reworkable underfillers, which are softened by heating and can be separated easily, are commercially available. Such reworkable adhesives are very promising because all devices are getting gradually complicated and the probability of error is increasing. Dismantlable adhesives can be categorized as follows: 1. 2. 3. 4.

Thermoplastic adhesives, i.e., hot-melt adhesives Adhesives including blowing agents or expansion agents Adhesives including chemically active materials Adhesives to which an electrochemical reaction on interfaces is applied

59.3.1 Thermoplastic Adhesives and hot-Melt Adhesives Thermoplastic adhesives, i.e., hot-melt adhesives can be separated. They are softened by heating. For instance, solid wax has been used for temporary joints of work materials. The work materials are joined and fixed on the bed of a machine tool with the melt solid wax, and separated by heating. The wax is usually used for many materials such as glass, ceramics, and silicones which are nonmagnetic and cannot be fixed with a magnetic chuck. The amount of the wax used is increasing, although it is a relatively old material. In electronics applications, several types of dismantlable adhesives are used for making IC chips. At first, silicone ingots are fixed with an adhesive and cut into wafers with a dicing saw. The adhesive can be

1758

C. Sato

softened by heating. The residue of the adhesive on the wafers can be washed off using a particular solvent or an alkali solution. Next, the wafer is fixed with another type of adhesive which can be separated mechanically using a scraper. The residue of adhesive on the wafer’s surface can also be washed off with an alkali solution. Hot-melt adhesives began to be used for housing applications in order to recycle the materials. Figure 1 shows a bonding method of wall boards or ceiling boards on beams using a hot-melt tape and an induction heating machine. The method is called the “Allover Method” and has been utilized already by Japanese construction companies (Sekine et al. 2009). A key point is that the tape has a conductor layer made of aluminum alloy in it. To bond the boards on the beam, the tape is inserted between them. The process is easy because the tape has a weak tackiness. After that, the layer can be heated using the induction heating machine, and the board and beam are joined together. They can also be separated by heating because it is totally reversible. Combining thermosetting plastics and thermoplastics, reworkable adhesives having high strength, stability, and good utility, can be realized. Epoxy adhesives, which are thermosetting resins, are used often for electronics packaging. The introduction of a thermoplastic property to the epoxy adhesive leads to a novel function where the adhesive can be cured and melted by heating. This is called “thermally molten epoxy resin” and has become an important object for research recently (Nishida and Hirayama 2006). The resin is used as an under-fill adhesive to bond IC chips on circuit boards. If the IC chips are out of order, they can be removed and replaced by other chips. Such chips are usually bonded with an adhesive and solder, and both of them can be soften by heating. Therefore, the chips can be separated from the boards easily. The process is called “IC chip rework” (Fig. 2) and it is not frequent but occasional.

59.3.2 Adhesives Including Blowing Agents or Expansion Agents Mixture of blowing agents or expansion agents into an adhesive is frequently used as a method to provide dismantlability. For instance, thermally expandable microcapsules (TEMs) are used often, as shown in Fig. 3. The TEMs expand by heating and an internal stress occurs in the matrix resin of the adhesive. The stress causes an Fig. 1 Principle of the Allover method using induction heating (IH) and hotmelt adhesives (Sekine et al. 2009)

Heating coil High frequency power source

Interior finishing material

Beam or stud

Hot-melt adhesive having conductor layer

59

Recycling and Environmental Aspects

Fig. 2 Reworkable underfillers used for integrated circuit (IC) chip bonding

Reworkable underfiller

1759 IC chip with malfunction Solder bump

Circuit board Heating IC chip

Removing IC chip

Soldering of new IC chip and bonding with underfiller

Fig. 3 Dismantlable adhesive including thermally expandable microcapsules (TEMs)

Adhesive including thermally expandable microcapsules

Heating

Adherend

1760

C. Sato

interfacial separation between the adhesive and the adherend. The type of the matrix resin is important because the internal stress depends on its modulus and strength. Thermoplastic resins are suitable for the adhesive. They can deform much at high temperatures. The expanding force of TEMs induces large deformation in the matrix resin and that leads to interfacial fracture between the adhesive and the adherend (Ishikawa et al. 2005). In addition, the separated surface of the adhesive becomes rough because of the TEMs’ expansion, and the roughness becomes a barrier for the adhesive to stick again to the adherend. Therefore, the separated interface cannot be joined again even after the temperature decreases. In the case of a thermoplastic adhesive, the resin is softened over its glass transition temperature, and a similar phenomenon happens also in the case of thermosetting adhesives. Epoxy resins, for example, can be applied as a matrix of a dismantlable system, and they can be used for a wide variety of applications because of the high strength and heat resistance. As an example of application of this technique, a commercially available tape including TEMs can be shown, which has been used for temporary joints of IC chips die-cutting. The tape has double sided coats of a pressure sensitive adhesive (PSA). The bonding strength and modulus of the PSA are low, so that the tape can expand much with a small amount of TEM inclusion, and the expansion leads to an interfacial separation between the IC chips and the tape. Similar methods are applied to other applications such as adhesives or sealants for housing. A vinyl emulsion adhesive including TEMs is used to join fiber-reinforced plastics (FRP) or decorative steel sheets to plaster boards for fabrication of unit-bath structures in Fig. 4. A silicone sealant including TEMs is used for sealing of kitchen sinks, etc., as shown in Fig. 5. In both the cases above, the joint can be dismantled by heating (Ishikawa et al. 2004). The technique has been applied to strong adhesives too. A project called “ECODISM,” which is the abbreviation of “Ecological and economical development of innovative strategy and process for clean maintenance and dismantling further recycling,” was carried out in Europe aiming at the bonding of car structures using dismantlable adhesives or primers (Papon et al. 2008; Alcorta et al. 2004; Bain Heating Befor heating

Bath room

Bath wall

After heating

Surface layer Back of Steel plate

Substrate Plaster board

Fig. 4 Wall panels for a unit-bath room consisting of plaster boards and steel plates bonded with a dismantlable adhesive of vinyl emulsion including TEMs (Ishikawa et al. 2004)

59

Recycling and Environmental Aspects

Befor heating (back side)

1761 Heating

After heating

Fig. 5 Bonded and dismantled parts of a kitchen sink bonded with a dismantlable adhesive of silicone resin including TEMs (Ishikawa et al. 2004) Fig. 6 Application of a dismantlable adhesive with thermally expandable microcapsules and blowing agents in the car industry, the ECODISM project (Papon et al. 2008)

and Manfre 2000). A urethane adhesive including TEMs and blowing agents (pTSH) was applied to joints between a windscreen and the body structure, as shown in Fig. 6. A rear hatch having polymer-alloy panels and a glass was also fabricated with the dismantlable adhesive. Infrared (IR) lamps were used to heat the structure in order to separate it, and the structure could be dismantled within 150 s. Epoxy adhesives are also targets for separation using TEMs, although these adhesives are very strong. For instance, a type of TEMs, shown in Fig. 7, is mixed into an epoxy resin to be used as a strong dismantlable adhesive (Nishiyama et al. 2003). The microcapsules have a structure with a core and a shell. The shell consists of polyvinylidene, and the core is filled by isobutene. The average diameter of the TEMs is about 20 μm. The TEMs can expand due to temperature increases because the shell becomes soft and the inner pressure of the TEMs increases. The expansion starting temperature of the TEMs depends on the material characteristics of the shell and the filling hydrocarbon. There are a lot of commercially available grades having different temperatures of expansion beginning and maximum expansion volumes. The TEMs, shown in Fig. 7, have an expansion starting temperature of 80
C and a maximum expansion volume of 70 times. Therefore, the temperature of dismantlement of an adhesive is selected as a function of the TEMs. Generally speaking, a grade of TEMs having a high expansion starting temperature has a smaller expansion volume than that

1762

C. Sato

Shell: Thermoplastics

Heating Gas

Core: Liquid hydrocarbon

Expansion

30 µm

30 µm

Fig. 7 Structure of thermally expandable microcapsules (TEMs) (Nishiyama et al. 2003)

of low temperature expansion TEMs. Therefore, an increase of the dismantlable temperature of an adhesive may induce a decrease of dismantlability because the expansion volume is vitally important to separate the bonded joints. As mentioned before, since the stresses caused by the TEMs expansion is the main driving force for joint separation, the expansion characteristics should be known to improve the performance of a dismantlable adhesive. Several experiments have been carried out to investigate the expansion performance of a type of TEM (Nishiyama and Sato 2005). For the purpose, Pressure-Volume-Temperature tests (PVT tests) were conducted using an experimental set up called “PVT measurement equipments.” In these tests, TEMs are contained in a flexible vessel under hydrostatic pressure with silicone oil under different temperature conditions. The hydrostatic pressure is caused using a pump and the temperature increases using a heater. The volume change of the TEMs was measured using a differential transformer. Therefore, the volume change of TEMs under different conditions of pressure and temperature could be measured and the P- V- T relation could be determined by the tests. For instance, the TEM mixed in the epoxy resin can expand 8 times in volume at 100
C. The pressure, P, caused by the TEM for a volume expansion of 800% and T = 100
C could be determined as 1–2 MPa. The result is strange because the pressure is much lower than the strength of the epoxy resin. In other words, the pressure is too small to deform the epoxy matrix resin enough for dismantlement. However, the resin becomes soft enough to deform at the temperature of the TEMs expansion. Thus, the combination of the resin softening and the TEM expansion is the key for the adhesive’s large deformation and internal stress occurrence. In order

59

Recycling and Environmental Aspects

Adherend

Adhesive

1763

Pit

Adherend

Adhesive

Microcapsule

Crack propagation

Two types of interfacial fracture LED light

Dark part (Debonded area) Bright part (Bonded area)

Glass adherend Adhesive Adhesively bonded glass joint to observe debonding

Before heating

Bonded area

Debonded area

Microphotograph of crack tip region

70°C

73°C

Fig. 8 Dismantling experiments to observe the interfacial debonding caused by TEMs in an epoxy adhesive; debonded regions in the joint can be seen as the dark parts in the images at the bottom of the figure, and they are expanding with respect to temperature

to provide dismantlability to an adhesive by mixing TEMs, not only the expansion force of the TEMs is important, but also the low strength and modulus of the matrix resin at high temperatures. A resin, whose Tg is close to the volume expansion temperature of the TEM, whose modulus and strength decreases drastically around the Tg, and which is soft enough not to suppress the expansion of the TEM in its rubber plateau, has to be selected as the matrix for a dismantlable adhesive. Two mechanical models of the interfacial debonding of dismantlable adhesives are proposed in Fig. 8. One of them is the local crack generation model, in which the expansion of a TEM causes a small interfacial crack nearby the contact point between the TEM and the adherend. The crack expands to the neighbor cracks, and the cracks join and lead to the total fracture of the joint. The other model is the global stress

1764

C. Sato

generation model, in which the expansion of TEMs induces volume increase of the adhesive layer, leading to joint fracture due to the global residual stress concentrating around the tip of the adhesive layer or interfacial cracks. An experimental verification of the models has been carried out using a glass joint bonded by an adhesive, including TEMs. In this test, optical and microscopic observation was conducted through the glass adherend. No gas generation was observed. Small circumferential cracks around each TEM were observed as explained by the local crack generation model. However, global cracks occurred at the edges of the joint and propagated inside the joint. This implies that the global stress generation model is also valid. Thus, the results support both models, which happen simultaneously, although the contribution ratio of each effect is still unknown.

59.3.3 Adhesives Including Chemically Active Materials Mixture of chemically active materials is an effective method to dismantle adhesives. For example, expandable graphite and aluminum hydroxide can be used for the purpose. Expandable graphite is a flaky carbon having a layer structure of planar crystals. Chemicals such as acids are intercalated between the crystal layers. The acid is vaporized by heating, and water included in the acid is also vaporized. Therefore, the vaporized acid and the vapor can expand the layer structure as an accordion being pulled, increasing the volume drastically, as shown in Fig. 9. Since the vaporized acid and vapor are hot, they can attack the resin. Aluminum hydroxide consists in a fine powder that is transformed into aluminum oxide and vapor at high temperature. The hot vapor can not only expand the resin, but also attack the resin in the same manner as the expandable graphite does. Water becomes superheated vapor over 100
C in the atmospheric pressure. The superheated vapor can cause chemical decomposition of the resin. An application of the technique is double-side PSA tapes that include aluminum hydroxide and which used for the joining heat sinks to largesize plasma display panels. They have to be dismantled in their recycling process because they are dissimilar materials. The joint can be separated applying a weak force because the PSA tape can be degraded by heating and the bonding strength decreases very much. Other types of chemicals, oxidizing agents such as ammonium perchlorate, can be used to attack the resin too.

59.3.4 Adhesives to Which an Electrochemical Reaction on Interfaces Is Applied A novel adhesive: Electrelease (EIC Laboratories, MA, USA), which can be separated electrically, is already commercially available (Welsh et al. 2003). The interface between the adhesive and the anodic adherend becomes so weak by applying a DC current that the joint can be broken by hand, as shown in Fig. 10. The adhesive was developed to bond sensors on the wing of aircrafts, and has been tested to be applied to a separation mechanism for space applications. The application of this

59

Recycling and Environmental Aspects

1765

Layered graphite Intercalation (H2So4, HNO3)

Heating Expandable Graphite

Expansion

Fig. 9 Expandable graphite

technique weakens the joint, but not completely. Therefore, some external force has to be applied to separate the joint even after the electric voltage is imposed. For instance, if a weak load is applied continuously using bias springs to the joint, applying only an electric voltage causes the spontaneous separation of the joint. As just described, a one-time separation mechanism triggered electrically can be realized easily using this adhesive as shown, in Fig. 11 (Shiote et al. 2009). The reason why the adhesive can be weakened by electric currents is not completely understood. It is supposed that an electric current transmitted to the adhesive layer causes electrochemical reactions on the surface of the anodic adherend, and this leads to erosion of the metal-oxide layer of the adherend. A similar mechanism of delamination induced by electric currents is known as “cathodic delamination,” in which

1766 Fig. 10 An electrically dismantlable adhesive (Electrelease ®, EIC Laboratories) where PEG and PDMS denote polyethyleneglycol and polydimethylsiloxane, respectively (Shiote et al. 2009)

C. Sato Anode Adhesvie

PEG PDMS

Cathode

Epoxy

Fig. 11 One-time separation mechanism using an electrically dismantlable adhesive (Shiote et al. 2009)

Flat-plate T-plate Adhesive

–

+

Spring

–

+

alkali materials precipitate at the interface and attack the resin (Horner and Boerio 1990). These delamination mechanisms are quite novel and very promising in terms of expanding the variety of dismantlable adhesives.

59.3.5 Miscellaneous Methods PSAs are promising materials to realize dismantlable joints because separation capabilities are built in already. Recently, strong PSAs have been commercially available, and they have enough strength to be applied for semi-structural uses.

59

Recycling and Environmental Aspects

1767

However, it becomes a problem in terms of dismantlability if the strength is too high. To avoid this problem, new types of PSAs, which are strong enough but can be separated easily, have been developed. Command Tab ® (3M, MN, U.S.A.) and PowerStrip ® (TESA, Hamburg, Germany) are examples. Tackiness of PSAs depends on the modulus. If the modulus is out of a particular range known as Chang’s “viscoelastic window,” the PSA loses the tackiness or the bonding strength (Chang 1991). For instance, the PSA of PowerStrip comprises block-copolymer molecular chains having soft and hard segments, and the modulus depends on the deformation (Krawinkel 2003). If the PSA is pulled enough, the modulus increases and the tackiness decreases. Therefore, the PowerStrip tape can be removed easily, as shown in Fig. 12. This theory of the viscoelastic window can be applied to other materials. Mixture of cross-linkers is another way by which PSAs can be hardened by heating or UV irradiation and the tackiness decreased. Recently, many kinds of tapes that can be separated by this method are used for dicing of IC chips. Brittle adhesives have high shear and low peel strength. It is possible to take advantage of this aspect to make a kind of anisotropic adhesive. For example, a temporary joint adhesive, Lock n0 Pop ® (Illinois Tool Works (ITW), WA, U.S.A.) is commercially available. The adhesive is strong enough to fix cardboard boxes due to the high shear strength during their transportation, but is easy to separate by applying vertical forces because of its low peel strength, as shown in Fig. 13.

Load

Hard segment

Soft segment Elongation

Stress concentration (modulus increasing)

Fig. 12 High strength strippable PSA tape (PowerStrip ®, TESA)

1768

C. Sato

Fig. 13 Anisotropic strength adhesive for a temporary joint of cardboard boxes (Lock n’ Pop ®, Illinois Tool Works (ITW))

Strong for shear loads

Fig. 14 A dismantlable adhesive application for shoe industry (Mori and Harano 2009)

Weak for tensile loads

Upper

Mid sole

Sole

Outer sole

59.3.6 Recent Advances The use of adhesives including TEMs has been expanding in many applications. For example, an adhesive has been introduced for shoe fabrication to bond uppers and soles (mid soles and outer soles) as shown in Fig. 14 (Mori and Harano 2009). The adhesive includes TEMs as expansion agents and diethylene glycol or ethylene glycol as heat-generation agents. Since diethylene glycol is a high-dielectric material, the temperature can be risen by microwave irradiation heating. The temperature rise can induce the expansion of TEMs and finally leads to the separation of the uppers and soles. Microwave irradiation has advantages such that low energy loss, and the necessary equipments have become inexpensive recently. For instance, even a home microwave oven can be used to dismantle the shoes. Requirements for adhesives used in sport shoes are many and severe. Bond strength, flexibility, and heat resistance are indispensable. To meet these requirements, a formulation, which includes an urethane resin, 5 wt% of TEMs, and over 20 wt% of high-dielectric materials, is typically used. At the moment, the adhesive is only applied to expensive shoes for top athletes because they want to reuse the uppers when the soles have to be replaced. However, the adhesive is ready to be used in ordinary shoe markets in terms of performance and price. Shoe industries have huge markets in which adhesives have to be used essentially because dissimilar materials are bonded. Since the total mass of production is enormous and has been increasing, the use of recyclable joining techniques is of significance, especially in the future.

59

Recycling and Environmental Aspects

1769

The use of TEMs is convenient to provide adhesives with dismantlability, but it is not versatile because of some disadvantages. One of them is the small expansion forces produced by TEMs, by which a weak adhesive can be separated but strong ones such as rubber modified ductile epoxy adhesives cannot be separated because they have high strength and ductility suppressing interfacial fracture even at a higher temperature than its Tg. To overcome this disadvantage, there are two alternative ways: changing the expansion agents or modifying the matrix resin. As alternative expansion agents, expandable graphite and aluminum hydroxide are promising, as mentioned above. They are not only chemical attack agents, but also expansion agents having larger expansion capabilities than TEMs. Modification of matrix resins is also important, but this is the next step and it has not been investigated often so far. A strategy is to modify the resin to be softer than the previous ones at high temperature so that the resin can deform due to the TEMs’ expansion. A problem is that the heat resistance of the resin might be damaged if the resin’s Tg becomes too low. Thus, a good compromise of heat resistance and dismantlability must be pursued. Kishi et al. has tried to develop a heat resistant dismantlable adhesive for joining composite materials and metals, and modified the composition of an epoxy adhesive accordingly (Kishi et al. 2006). This research was conducted as part of the national project “R&D of Carbon Fiber-Reinforced Composite Materials to Reduce Automobile Weight” supported by New Energy and Industrial Technology Development Organization (NEDO) in Japan. As mentioned above, the resin should be soft at the temperature of the TEMs’ expansion. A problem is that typical heat resistant resins have moduli and strengths enough to suppress the TEMs’ expansion even at high temperatures, over their Tg. In other words, such heat resistant resins are not soft enough for TEMs to expand even at high temperature. Therefore, a kind of resin which has a high Tg and enough low modulus and strength above the Tg is necessary for heat resistant dismantlable adhesives. In addition, steep change of the modulus and strength of the resin around the Tg is also important to increase the maximum allowable temperature of the adhesive, as shown in Fig. 15.

Heat resistant resin Non heat resistant resin

Temperature

Expansion force

Ideal variation of modulus

Young’s modulus

Fig. 15 An idealistic variation of modulus for dismantlable adhesive’s matrix resin

1770

C. Sato

Kishi et al. aimed a target for the adhesive as: 1. The maximum temperature for use is 80
C. Under this temperature, the resin has a modulus and a strength high enough to bear the applied loads. 2. Over 100
C, the resin has a steep decrease of the modulus and strength. 3. In the temperature range over 150
C, in which the resin is in rubbery condition, the resin should have a modulus and strength low enough for TEMs expansion. For the adhesive, they proposed a formulation, shown in Fig. 16, including diglycidylether of bisphenol A (DGEBA) 828, DGEBA 1001, and glycidylphthalimide (GPI) as the base resins cured with hardeners dicyandiamide (DICY) and 3-(3,4-dichlorophenyl)-1,1-dimethylurea (DCMU) that can increase the Tg of the resin. An important point of the formulation is the inclusion of GPI that has high polarity and can form strong bonds between the main molecular chains. The bonds work as if the cross-linking density of the resin increases at low temperatures. However, the bonds are dissociated at high temperatures, decreasing the substantial cross-linking density. The low modulus at high temperatures and the steep decrease of the modulus around the Tg are due to the destruction of the bonds. Thus, the resin has high modulus and strength in low temperature ranges, high Tg, steep decrease of the modulus, and low modulus over Tg, which are suitable characteristics for heat resistant dismantlable adhesives. For instance, the resin has a modulus higher than 3 GPa in the glass condition range around the ambient temperature. However, the modulus decreases to 2 MPa over 150
C. The resin was combined with an expandable graphite and was used to bond metal parts to carbon-fiber-reinforced plastics (CFRP) plate. The joint could be separated within 5 min at 250
C, as shown in Fig. 17. The adhesive includes 10 wt% of the expandable graphite and it begins to expand around 200
C, which is much higher than the Tg of the resin. The margin between the expansion temperature of the expandable graphite and the Tg of the resin was decided considering the chemical stability of the expandable graphite at high temperatures. The temperature of 250
C

H H2C C C O H2 O

CH3 C CH3

CH3

H O C C Cn O H2 H2 OH

C CH3

H O C C CH2 H2 O

DGEBA(828):n = 0.1,DGEBA(1001):n = 2 O C H N C C CH2 H2 O C

HN

C

NH

CN

NH2

O GPI

DICY

Fig. 16 Components of the heat resistant dismantlable adhesive (Kishi et al. 2006)

59

Recycling and Environmental Aspects

(Before heating)

1771

(After heating at 250°C for 5 min)

Fig. 17 Specimen of metal parts bonded on a carbon reinforced plastic (CFRP) plate with a heat resistant dismantlable adhesive

in the dismantlable test was decided as a compromise between time and temperature of separation.

59.3.7 Future Seeds One of the promising seeds to realize novel types of dismantlable adhesives is the use of degradable polymers. Polyperoxide-polymers, for instance, can be applied for that purpose because the resin can be cured and re-liquefied reversibly by heating (Matsumoto and Taketani 2006). Adhesives using the resin can be separated and bonded many times. Another kind of polymer having weak points such as Azo groups in the molecular chains is also promising for heat-degradable dismantlable adhesives. A type of rubber such as thermoreversible hydrogen bond cross-linked (THC) rubber, shown in Fig. 18, can be re-liquefied by heating, has been proposed recently (Chino et al. 2002; Chino and Ashiura 2001; Chino 2006). Such rubbers are also applicable for dismantlable adhesives. Since almost all recent dismantlable adhesives need heating to be separated, heating methods are also important to improve the efficiency of the dismantling process. The necessity of heating will remain even in the future. Heating the whole structure including the adhesively bonded joints is not efficient. Local heating of the joints is more convenient. Induction heating and microwave heating methods are appropriate in terms of efficiency. A new method of microwave heating using nanoferrite particles, shown in Fig. 19, has been proposed and it can be applied for the heating process of joints bonded with dismantlable adhesives (Sauer et al. 2004). Other stimulations except heating should be investigated in the future. Magnetism, radiation (neutron beams, gamma ray), hydrostatic pressure, compressive stress, UV irradiation, and high voltage discharge are candidates for the stimulation. Dry adhesion, which has been attracting attention recently, is another possibility to provide new ideas for joint dismantlement (Lee et al. 2007). Further investigation and research on interfacial phenomena must be conducted to obtain further

1772

C. Sato Crosslinked rubber

Low recyclability

Crosslink by covalent bonds

THC rubber High recyclability Heating Crosslink by hydrogen bonds Cooling

Fig. 18 Schematic illustration of thermoreversible hydrogen bond cross-linked (THC) rubber principle

Fig. 19 Microwave heating using nanoferrite as dipole conductors (Sauer et al. 2004)

Microwaves

Adhesive with nanoferrite

Temperature increase without energy loss

59

Recycling and Environmental Aspects

1773

knowledge because the ultimate dismantlable adhesive must be an adhesive with which its interface can be separated easily and completely with a small amount of energy used for the separation.

59.4

Conclusion

To promote the use of adhesives, eco-friendly attitudes of both adhesive makers and users are indispensable. Reducing carbon dioxide and energy consumption is the main topics of environmental issues. Recently, alternative options, such as the use of carbon-neutral or recycled materials for making adhesives, can be taken to meet the demands derived from the environmental issues. The optimization of bonding process and joint design is also very important and effective to reduce energy consumption and wastes. Recently, dismantlable adhesives have been available for joining the materials that have to be separated on demand. These adhesives can be applied to adherend recycling or product reworking. There are many technical seeds for dismantlable adhesives already. Since dismantlable adhesive technology has still some minor problems such as low heat resistance or low durability, they should be solved soon by researchers or industries, and then the use of the adhesives will expand to many applications. Dismantlable adhesive technology has contradictory aspects: good bonding and easy separation. High bonding strength tends to induce difficult separation. Therefore, it is also important to clarify the minimum bond strength necessary for each application instead of pursuing high performance in strength.

References Alcorta J, Papon E et al (2004) WO Patent 2004/087829 Bain PS, Manfre G (2000) WO Patent 2000/75254 Chang EP (1991) Viscoelastic windows of pressure-sensitive adhesives. J Adhes 34:189 Chino K (2006) Thermoreversible crosslinking rubber using supramolecular hydrogen bonding networks. Kautschuk Gummi Kunststotte 2006, April:158 Chino K, Ashiura M (2001) Themoreversible cross-linking rubber using supramolecular hydrogenbonding networks. Macromolecules 34:9201 Chino K, Ashiura M et al (2002) Thermo reversible crosslinking rubber using supramolecular hydrogen bonding networks. Rubber Chem Technol 75:713 Fay P (2005) History of adhesive bonding. In: Adams R (ed) Adhesive bonding – science, technology and applications. Woodhead, Cambridge, pp 3–22 Horner M, Boerio J (1990) Cathodic delamination of an epoxy/polyamide coating from steel. J Adhes 32:141 Ishikawa H, Seto K et al (2004) Dismantlable adhesive and evaluation of practical uses for building materials. J Soc Mater Sci Japan 53:1143. (in Japanese) Ishikawa H, Seto K et al (2005) Bond strength and disbonding behavior of elastomer and emulsiontype dismantlable adhesives used for building materials. Int J Adhes Adhes 25:193 Jos H, Leite L (2007) MSc thesis, Faculty of .Engineering of the University of Porto (in Portuguese) Kadota J, Hasegawa K et al (2004) Adhesive of lignophenols/polyacrylate for wood. J Adhes Soc Japan 40:101. (in Japanese)

1774

C. Sato

Kishi H, Inada Y et al (2006) Design of dismantlable structural adhesives with high temperature performance. J Adhes Soc Japan 42:356. (in Japanese) Krawinkel Th (2003) In: Proceedings of Swiss bonding 03, Rapperswill, pp 225–233 (in German) Lee H, Lee B et al (2007) A reversible wet/dry adhesive inspired by mussels and geckos. Nature 448:338 Matsumoto A, Taketani S (2006) Regiospecific radical polymerization of a tetra-substituted ethylene monomer with molecular oxygen for the synthesis of a new degradable polymer. J Am Chem Soc 128:4566 Mori S, Harano K (2009) Shuzu no seikei kakou. Seikei kakou 21:305. (in Japanese) Nishida H, Hirayama N (2006) PCT/JP2006/309543 Patent Nishiyama Y, Sato C (2005) Behavior of dismantlable adhesives including thermally expansive microcapsules. In: Possart W (ed) Adhesion – current research and applications. Wiley, New York, pp 555–568 Nishiyama Y, Uto N et al (2003) Dismantlement behavior and strength of dismantlable adhesive including thermally expansive particles. Int J Adhes Adhes 23:377 Papon E, Olive M et al (2008) In: Proceedings of the 31st annual meeting of the adhesion society, Austin, pp 296–297 Ratna D, Banthia A (2000) Epoxidized soybean oil toughened epoxy adhesive. J Adhes Sci Tech 14:15 Sauer H, Spiekermann S et al (2004) Accelerated curing with microwaves and nanoferrites. Adhesion. Adhes & Sealants 2003(2004):48–50 Sekine T, Tomita H et al (2009) An induction heating method with traveling magnetic field for long structure metal. Electr Eng Japan 168:32 Shiote H, Ohe M et al (2009) Effects of electrical treatment conditions on dismantlable properties of joints bonded with electrically disbanding adhesive. J Adhes Soc Japan 45:376. (in Japanese) Sprenger S, Eger C et al (2004) Nano-modified ambient temperature curing epoxy adhesives. Adhesion. Adhes & Sealants 2003(2004):20–23 Umemura K, Inoue A et al (2003) Development of new natural polymer-based wood adhesives I: dry bond strength and water resistance of konjac glucomannan, chitosan, and their composites. J Wood Sci 49:221 Viljanmaa M, Sodergard A et al (2002) Lactic acid based polymers as hot melt adhesives for packaging applications. Int J Adhes Adhes 22:219 Welsh J, Higgins J et al (2003) Evaluation of electrically disbonding adhesive properties for use as separation systems. In: AIAA 2003–1436 (44th AIAA/ASME/ASCE/AHS structures, structural dynamics, and materials conference), Norfolk, pp 1–3 Yamada K, Chen T et al (2000) Chitosan based water-resistant adhesive. Analogy to mussel glue. Biomacromolecules 1:252

Adhesion Technology Recap: Current and Emerging Areas

60

Lucas F. M. da Silva, Andreas Öchsner, and Robert D. Adams

Contents 60.1 60.2 60.3 60.4 60.5 60.6 60.7 60.8 60.9 60.10 60.11

Theory of Adhesion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Surface Treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Adhesive and Sealant Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Testing of Adhesive Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Joint Design . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Durability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Manufacture . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Quality Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Emerging Areas . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

1776 1776 1777 1777 1778 1779 1780 1780 1781 1782 1783

Abstract

This chapter gives a brief description of the subjects and chapters that are included in this second edition of the handbook and the most important changes in relation to the first edition. The handbook is organized in ten parts: theory of

L. F. M. da Silva (*) Department of Mechanical Engineering, Faculty of Engineering, University of Porto, Porto, Portugal e-mail: [email protected] A. Öchsner Faculty of Mechanical Engineering, Esslingen University of Applied Sciences, Esslingen, Germany e-mail: [email protected] R. D. Adams Department of Mechanical Engineering, University of Bristol, Bristol, UK Department of Engineering Science, University of Oxford, Oxford, UK e-mail: [email protected] # Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2_59

1775

1776

L. F. M. da Silva et al.

adhesion, surface treatment, adhesive and sealant materials, testing of adhesive properties, joint design, durability, manufacture, quality control, applications, and emerging areas. A total of 58 chapters are presented covering all aspects of adhesion and adhesives. In addition to the information contained in each chapter, an extensive list of references is given (approximately 4000 references).

60.1

Theory of Adhesion

In ▶ Chap. 2, “Theories of Fundamental Adhesion” by Packham, the historical development and current status of the four classical theories of adhesion (mechanical theory, adsorption theory, electrostatic theory, and diffusion theory) are reviewed. The role of weak boundary layers is also discussed with emphasis on the importance of careful investigation of the locus of failure of an adhesive bond. Next, Brogly in ▶ Chap. 3, “Forces Involved in Adhesion” describes the main forces responsible for adhesion, from strong covalent bonds to weak van der Waals forces, also considering specific interactions such as acid-base or capillary forces. ▶ Chapter 4, “Wetting of Solids” by Shanahan and Possart considers thermodynamic aspects of wetting, which involves intimate contact between the two phases and the environment. A short overview of the relevant processes and parameters in the spreading of liquids on substrates is presented in ▶ Chap. 5, “Spreading of Liquids on Substrates” by Reiter. In a simplified view, the dynamics of these processes can be understood as being controlled by the balance of driving forces and resistance due to dissipative processes. Finally, in ▶ Chap. 6, “Thermodynamics of Adhesion,” Possart and Shanahan show that wetting data may be used to estimate the thermodynamic, or Dupré, energy of adhesion, provided certain assumptions are made and suitable models constructed for, in particular, interfacial tensions.

60.2

Surface Treatments

The use of surface treatments to optimize adhesion has been well-established. In ▶ Chap. 7, “General Introduction to Surface Treatments,” Critchlow considers the main treatment methods for metals and polymers in terms of how such processes are carried out and their influence on surface physical and chemical properties. Consideration has been given to a range of treatments from simple degrease options to the more highly complex multistage processes. Davies, in ▶ Chap. 8, “Surface Treatments of Selected Materials,” discusses high-performance surface treatments for several metals and other materials. Surface treatment of aluminum and other metals is used to illustrate how proper surface preparations meet these requirements. ▶ Chapter 9, “Surface Characterization and Its Role in Adhesion Science and Technology” by Watts reviews a variety of methods of surface characterization that have been found to be useful in the study of adhesion. The methods considered can conveniently be considered in three groups: surface topography, surface free energy of a material, and surface-specific chemical analysis. Next, Watts in

60

Adhesion Technology Recap: Current and Emerging Areas

1777

▶ Chap. 10, “Use of Surface Analysis Methods to Probe the Interfacial Chemistry of Adhesion” explores the manner in which the surface analysis methods of X-ray photoelectron spectroscopy (XPS) and time-of-flight secondary ion mass spectrometry (ToF-SIMS) can be used to extract information regarding the interfacial chemistry of adhesion from polymer/metal systems such as adhesive joints. The last chapter of this section, ▶ Chap. 11, “Organosilanes: Adhesion Promoters and Primers” by Abel, describes organosilanes from their genesis to their chemistry. The necessary interactions that such molecules have to develop on materials in order to fulfil their main role as adhesion promoters are explained. The use of silanes as primers particularly where the user aims to improve adhesion or protect from corrosion is also considered. Some other organic or nonorganic adhesion promoters are also discussed.

60.3

Adhesive and Sealant Materials

The first chapter ▶ “Classification of Adhesive and Sealant Materials” (Chap. 12) by Sancaktar classifies adhesive and sealant materials. For this purpose, various categories are considered depending on the polymer base, functionality in the polymer “backbone,” physical forms, chemical families, functional types, and methods of application. Next, in ▶ Chap. 13, “Composition of Adhesives,” Kim et al. give information about various ingredients that may appear in an adhesive or sealant: primary resins, hardeners, solvents, fillers, plasticizers, reinforcements, and various additives. In the third chapter of this section, Papon (▶ Chap. 14, “Adhesive Families”) classifies adhesives and sealants in three broad groups. First, there are adhesives where the polymer is pre-existing (and must be placed beforehand in fluid form: solution, emulsion, or “melt” state). Second there are adhesives where the polymer is formed during the course of a reactive process (polymerization). Third, Papon describes the particular category of pressure-sensitive adhesives where the polymer exhibits viscoelastic properties able to develop adhesion during the bonding step. The basic concepts, formulations, and test methods for pressure-sensitive adhesives are presented in ▶ Chap. 15, “Pressure-Sensitive Adhesives (PSAs)” by Paul, stressing the importance of interfacial interactions, viscous loss, and extensibility. Selection of the correct adhesive for an application can be a daunting task due to the many types commercially available ranging from different chemistries through different forms to an almost continuum of material properties. Kellar in ▶ Chap. 16, “Selection of Adhesives” gives a logical approach to this task and selection criteria.

60.4

Testing of Adhesive Properties

The first chapter by Dillard ▶ “Physical Properties” (Chap. 17) addresses several relevant physical properties of adhesives, including viscosity, density, and stressstrain behavior, quantities that are often thought to be intrinsic properties of a material. The thermal properties of adhesives are described next by Comyn.

1778

L. F. M. da Silva et al.

Consideration is given to the shelf life, the hardening process, the glass transition temperature, the thermal conductivity and thermal expansion, and the thermal breakdown of adhesives. In ▶ Chap. 19, “Failure Strength Tests,” da Silva et al. describe the major failure strength tests used to determine the intrinsic adhesive properties. Tensile, compressive, and shear tests are described with reference to the major international standards. Bulk and in situ (adhesive in a joint) tests are discussed and related. ▶ Chap. 20, “Fracture Tests” by Blackman considers first the fracture in a bulk adhesive specimen under mode I, i.e., tensile opening loading conditions. Mode II (in plane shear) and mixed-mode I/II testing of adhesive joints are also considered. There is often the need to measure the resistance to fracture in a joint with flexible substrates. One such test is the peel test, and some variants of the peel test are considered in that chapter. Goglio, in ▶ Chap. 21, “Impact Tests,” describes the main tests used to assess the impact strength of adhesives and joints. The main tests used in the experiments are pendulum, falling weight, and Hopkinson bar loading conditions or evolutions of these. The last chapter of this section by Dillard and Yamaguchi deals with ▶ Chap. 22, “Special Tests.” This chapter gives a brief description of special mechanical tests for various types of material and sample geometries, such as blister tests, tensile tests and shear tests for sealants/foam adhesives, indentation tests, scratch tests, tack tests, and tests for the evaluation of residual stresses.

60.5

Joint Design

This section starts with ▶ Chap. 23, “Constitutive Adhesive and Sealant Models” by Sancaktar. The chapter includes deformation theories and viscoelasticity with linearity and nonlinearity considerations, rubber elasticity, singularity methods, bulk adhesive as composite material, damage models, the effects of cure and processing conditions on the mechanical behavior, and the concept of the “interphase.” ▶ Chapter 24, “Analytical Approach” by Tong and Luo presents an analytical approach for determining stress and strength of adhesively bonded joints. Various closed-form solutions for adhesive stresses and edge bending moment for balanced single-lap joints are presented and compared. In ▶ Chap. 25, “Numerical Approach: Finite Element Analysis,” Ashcroft provides first a general background to the development and application of the finite element (FE) method. Then he discusses some of the practical aspects of FE modeling and gives applications of the FE method to adhesively bonded joints. FE analysis is currently the only technique that can comprehensively address the challenges of modeling bonded joints under realistic operating conditions. However, a reliable and robust method of using FE analysis to model failure in bonded joints is still to be developed. Öchsner, in ▶ Chap. 26, “Special Numerical Techniques to Joint Design,” introduces special numerical techniques to analyze adhesive joints. The first part covers special FE techniques which reduce the size of the computational models. The second part of this chapter introduces alternative approximation methods, the boundary element method and the finite difference method. In ▶ Chap. 27, “Design Rules and Methods to Improve Joint Strength,” da Silva et al. describe the main factors influencing joint

60

Adhesion Technology Recap: Current and Emerging Areas

1779

strength. Methods are then proposed to improve the joint strength by using fillets, adherend profiling and other geometric solutions, and hybrid joining. Repair designs are also discussed. Finally, configurations are recommended for several types of joint. ▶ Chapter 28, “Design with Sealants” by Anderson describes the many factors to go into the design of reliable sealant joints. The various joint types are discussed and illustrated, and their critical dimensions and materials are described. Sato, in ▶ Chap. 29, “Design for Impact Loads,” discusses design methods for adhesively bonded joints subjected to impact loading. Methodologies to treat the dynamic responses of structures are shown. Some examples of stress analysis are shown, where closed-form approaches and dynamic FE analyses are explained. The last chapter of this section ▶ “Vibration Damping of Adhesively Bonded Joints” (Chap. 30) by Adams et al. discusses the general concept of vibration damping in vibrating structures and how it can be used to limit vibration amplitudes. Simple equations for calculating the damping of a lap joint in tension and bending have been developed. As there is very little experimental data available, the damping of a variety of adhesively bonded single-lap joints has been measured by the authors.

60.6

Durability

In ▶ Chap. 31, “Effect of Water and Mechanical Stress on Durability,” Ashcroft and Comyn discuss the effect that absorbed water has on the strength of adhesively bonded joints. The influence of adherend surface treatment, applied stress, and adhesive type on the environmental degradation of bonded joints is demonstrated. Methods of modeling the environmental degradation of adhesively bonded joints using coupled hygro-mechanical finite element (FE) analysis are then described. Chapter “Radiation and Vacuum” of the first edition has been substituted by ▶ “Adhesives in Space Environment” (Chap. 32) by Dagras et al. In the first part, a description of the different uses and needs related to adhesives on spacecraft is given. Then, the characteristics of space environment are explained. Once in orbit, the adhesive has to withstand constraints as particles radiations, UV, and atomic oxygen that are not present on Earth. Before spacecraft launch, the adhesives have to be validated and tested on ground with conditions as representative as possible of the space environment. Finally, the impact of this space environment on adhesives and the associated degradations are described. ▶ Chapter 33, “Fatigue Load Conditions” by Ashcroft discusses some of the main factors when considering fatigue, with particular reference to issues applicable to bonded joint. The main methods of characterizing and predicting the response of bonded joints to fatigue loading are described. Finally, there are sections on the special cases of creep-fatigue and impact fatigue. In ▶ Chap. 34, “Creep Load Conditions,” Geiss considers viscoelastic models, superposition principles, experimental testing procedures, and predictive methods. In the last chapter of this section, ▶ “Durability of Nonstructural Adhesives” (Chap. 35) by Palmer, methods evaluating durability and test regimes, including exposure to elevated and reduced temperatures, UV radiation, moisture,

1780

L. F. M. da Silva et al.

saline solutions, stress, and fatigue, both individually and combined in cycles are considered.

60.7

Manufacture

The first subject in this section is ▶ “Storage of Adhesives” (Chap. 36) by Engeldinger and Lim. A major part of this chapter focuses on the shelf life and safety aspects. Adhesives are divided into four categories, solvent-based, waterbased, hot-melt, and reactive adhesives. In ▶ Chap. 37, “Preparation for Bonding,” Lutz discusses each item of the process chain: the storage of adhesives, transfer to the application area, different metering and dispensing technologies, mixing equipment used, substrate preparation, and quality control. Proposals are made regarding education of personnel and how to create a safe working environment. In ▶ Chap. 38, “Equipment for Adhesive Bonding,” Peschka describes the equipment required for manual adhesive bonding processes. Automation and robotics are outlined with a special emphasis on parameters affecting accuracy. Accelerated curing using such devices as UV radiation and inductive heating is also treated. The last chapter of this section, ▶ “Environment and Safety” (Chap. 39) by van Halteren, deals with the different aspects of consumer, work, and environmental protection, including health and safety information, related to the use of adhesives in industrial and domestic areas.

60.8

Quality Control

This section starts with ▶ Chap. 40, “Quality Control of Raw Materials” by Wakabayashi. The main components of adhesives are first described, and then quality control procedures are discussed such as chemistry, impurity content, molecular weight, viscosity, density, and quality control of raw materials in storage. Next, in ▶ Chap. 41, “Processing Quality Control” by Haraga, the processing quality control of adhesively bonded joints in actual production lines is discussed. Topics such as control of environmental conditions and materials, inspection of surface treatment of adherends, adhesive selection for easy process control, and education and training of operators are treated. In ▶ Chap. 42, “Nondestructive Testing” by Adams, the types of defect encountered in adhesive joints and the nondestructive testing techniques available to detect them are reviewed. Several techniques are available for void detection, the most commonly used being ultrasonics and different types of bond tester. The detection of poor cohesive properties is more difficult but can be achieved with ultrasonic or dielectric measurements. The last chapter of this section, ▶ “Techniques for Post-fracture Analysis” (Chap. 43), was rewritten by Créac’hcadec. This chapter presents three types of measurement systems: microscopy, compositional analysis, and mechanical observations. A table sums up an exhaustive list of the major observation devices. Then, the main methods are detailed

60

Adhesion Technology Recap: Current and Emerging Areas

1781

with macroscopic and mesoscopic observations, microscopic observation, physicochemico analysis, and mechanical observations.

60.9

Applications

The first application treated is the “Aeronautical Industry” by Hart-Smith. The following topics are treated: basic needs, adhesive characteristics, surface preparation, design of adhesive joints, durability, defects, thermal effects, quality control, composite structures, repairs, and examples of use. ▶ Chapter 45, “Aerospace Industry” was extensively revised by Pérès and Larnac. The chapter introduces some of the specific bonding-related issues that face those responsible for bonding operations in space applications such as extremely high mechanical loads over a short time, with the launcher and its payloads subjected to gravity magnified by about 20, followed by exposure to extremely high temperatures over a very long period, with the observation and telecom satellites under zero gravity. Dilger and Frauenhofer updated ▶ Chap. 46, “Automotive Industry.” Typical applications for automobiles are shown. Additionally, suitable adhesive properties and surface treatments are discussed. ▶ Chapter 47, “Adhesive Bonding for Railway Application” is next treated by Suzuki. It is shown that adhesives are an indispensable joining technology for railway industries. Adhesives are used for the fabrication of almost all rail cars. For example, in the steel main structure of conventional rolling stock, adhesives are applied for bonding decorated aluminum sheets of wall and ceiling to frames, bonding of floor covering to the floor plate, and fixing heat insulating material to the inside of the carriages. ▶ Chapter 48, “Marine Industry” by Davies describes the use of adhesive bonding to assemble structures in the marine industry. The marine environment is extremely aggressive, and this has resulted in widespread use of fiber-reinforced composite materials. Adhesive bonding is a lightweight and corrosion-resistant means of joining these materials. Three industrial applications are used to illustrate the use of adhesive bonding, small pleasure boats, high-performance racing yachts, and bonded structures in the offshore industry. In the well-revised chapter ▶ “Civil Construction” (Chap. 49), Böhm et al. describe adhesive bonding applications in the civil construction sector, specific requirements, and adherends. The execution of bonding and properties of adhesive bonded civil constructions are also treated. Jung and Kim, in ▶ Chap. 50, “Electrical Industry,” explain the mechanisms underlying the electrical conduction in adhesive joints, and the thermal and mechanical parameters that should be measured are introduced. In terms of the evaluation of the reliability of adhesives in electronics, the basic test procedures, including several specific test methods and analysis techniques, are explained. ▶ Chapter 51, “Shoe Industry” by Martín-Martínez constitutes one of the very few reviews in the existing literature on shoe bonding and gives an updated overview of the upper to sole bonding by means of adhesives. The surface preparation of rubber soles and both the formulations of polyurethane and polychloroprene adhesives are described. The preparation of adhesive joints and adhesion tests are also revised. Another chapter was added to this section on ▶ “Oil Industry” (Chap. 52) by de Barros et al. due to the increase adhesives application that field,

1782

L. F. M. da Silva et al.

especially for repairs. The main advantage by using adhesive bonding is to avoid the fire risk concerning the traditional weld process. Adhesives also feature protection against corrosion which is a main concern in offshore installations. This chapter shows that the use of composites has become a real alternative to structural repairs in situ as these materials become more and more reliable.

60.10 Emerging Areas ▶ Chapter 53, “Molecular Dynamics Simulation and Molecular Orbital Method” by Zhao et al., various simulation methods pertaining to adhesion technology are introduced, such as molecular dynamics, quantum mechanics, the molecular orbital method, and the density functional theory. Some examples are proposed investigating adhesion issues using various simulation methods. ▶ Chapter 54, “Bioadhesives” was rewritten by Richter. It gives an overview about natural adhesives and biological adhesives leading to bioinspired applications. Biological and microbiological adhesives have an enormous potential for industrial applications. The most prominent ones are constructional, packaging, and medical applications. Gorb, in ▶ Chap. 55, “Biological Fibrillar Adhesives: Functional Principles and Biomimetic Applications,” reviews functional principles of biological systems in various animal groups such as insects, spiders, and geckos with an emphasis on insects and discusses their biomimetic potential. Data on ultrastructure and mechanics of materials of adhesive pads, movements during contact formation and breakage, the role of the fluid in the contact between the pad and substrate are presented. ▶ Chapter 56, “Adhesives with Nanoparticles” are discussed by Taylor. This chapter outlines the principal types of nanoparticles and the methods that may be used to disperse the particles in a polymer matrix. It discusses how nanoparticles can alter the mechanical properties (stiffness and fracture), electrical properties (conductivity), and the functional properties (permeability, glass transition temperature) of thermoset polymers. In ▶ Chap. 57, “Adhesive Dentistry,” Nicholson describes the principal uses of adhesives in clinical dentistry, covering both the materials and the clinical techniques involved. The two types of tooth-colored material, namely, the composite resin and the glass-ionomer cement, are described, together with their current variations. Surface pretreatment is critical for the success of these systems, and the current state of the understanding of this aspect of the subject is described. ▶ Chapter 58, “Adhesion in Medicine” is covered by Chivers. Adhesives technology has two main applications in medicine as considered in this chapter, for internal fixation of tissues usually after surgery and for use on the skin, primarily to hold dressings in place. The last chapter by Sato concerns ▶ “Recycling and Environmental Aspects” (Chap. 59). Adhesively bonded adherends should be often separated before they can be recycled. For this purpose, dismantlable adhesives, which can be separated with stimulations (softening of adhesive, expansion force due to blowing agents or thermally expandable microcapsules, chemical degradation, and electrochemical reaction), have been developed recently. These adhesives can be applied to adherend recycling or product reworking.

60

Adhesion Technology Recap: Current and Emerging Areas

1783

60.11 Conclusions The preparation of this second edition has been an interesting experience for the editors. The review process gave a deeper insight into the various aspects of adhesion from basics to applications. Authors come from every background (chemistry, physics, mechanics) and work in either academia or industry, proving that adhesion is a truly multidisciplinary and widely applied subject. The editors would like to thank the authors for their patience with the preparation of this handbook. Finally, the editors especially thank Dr. Christoph Baumann, Ms. Tina Shelton, and Ms. Monika Garg, Springer editors, who helped enormously toward the success of this handbook.

Index

A Ab initio molecular dynamics (AIMD), 1590–1592 Abrasion, 145, 178, 231, 825 ACAs, see Anisotropic conductive adhesives (ACAs) Accelerated ageing, 1013, 1016, 1020, 1022, 1024, 1026 Accelerated curing, 1113 Acetone method, 1509, 1510 Acid anodizing, phosphoric and chromic, 153–155 Acid–base interactions, 58–60, 68, 125, 211 effect of amphotericity, 60–61 IRAS, 61–62 Lewis definition, 48 molecular orbital approach, 48–51 Acid-etching, 1708, 1709, 1725 Acidified sodium hypochlorite aqueous solutions, 1492 Acoustic emission, 1187–1188 Acrylate, 323, 1035 Acrylic, 386, 389, 392–405, 919 adhesives, 188, 287 Additives, in polymers, 37, 336–338 Adherend(s), 4, 1011, 1016, 1021, 1026, 1238, 1242, 1248, 1249, 1251, 1252, 1255, 1258, 1264, 1272, 1275–1265 metallic, 137, 145 polymeric, 148 Adherend fatigue failure, 943 Adhesion, 4, 347 acid–base interactions, 48–51 adhesion forces, 54–64 adsorption theory, 15–21 atomic force microscopy, 201–203 auger electron spectroscopy and scanning auger microscopy, 216–221 capillary forces, 66–67

chemical bonding, 51–54 contact angle, 207–209 diffusion theory, 27–35 electrostatic theory, 25–27 fracture stress, 14 fundamentals of, 15 inverse gas chromatography, 209–212 joint fails energy, 13 mechanical theory, 21–25 peel, 376 promoter, 268, 277–279, 1522 requirement for chemical analysis method, 212 scanning electron microscopy, 199–201 self assembly monolayers, 64–66 simple test methods, 205–206 spreading theory of, 16 thermodynamic models of, 119 time-of-flight secondary ion mass spectrometry, 221–225 van der Waals forces, 45–47 weak-boundary-layer theory, 35–38 wetting work of, 19 white light interferometry, 204 X-ray photoelectron spectroscopy, 213–218 Adhesion forces acid–base, 58–60 amphotericity effect, 60 atomic force microscopy, 62–63 capillary force complication, 57–58 dispersive work of adhesion, 56, 57 vs. distance measurements, 63–64 ionocity and covalency effect, 60 IR spectroscopic tool, 61–62 van der Waals, 55 wetting, 55, 65 work of, 54–55 Adhesion primers, organosilanes as, 272

# Springer International Publishing AG, part of Springer Nature 2018 L. F. M. da Silva et al. (eds.), Handbook of Adhesion Technology, https://doi.org/10.1007/978-3-319-55411-2

1785

1786 Adhesion technology adhesive property testing, 1777–1778 adhesives and sealant materials, 1777 applications, 1781–1782 joint design, 1778 quality control, 1780 surface treatment, 1776 Adhesive(s), 4, 290, 320, 346, 1120–1123, 1679, 1681, 1684, 1685, 1691, 1695, 1696, 1777 agitator design, 1062 applications, 320 blowing agents/expansion agents, 1758, 1759, 1761, 1763 bonds, 17, 19, 1291–1294 centrifugal mixing, 1064 chemically active materials, 1764 classification of, 295, 321, 347–350 continuous and batch process, 1060 cracks, 1175 crosslinked, 388 CTE, 607–608 dispensing, 1069–1073 elastomeric foam tapes, 600 failure, 599, 600 film, 296, 1039, 1044 formulation, 330, 340, 349 fracture energy, 546 health and safety, 1055–1058 implementation, 348 kneader, 1062 low to medium volume application, 1074–1077 medium to large volume application, 1077 product transfermation, 1077 PSA, 605 quality control, 1087–1088 residual stresses, 607–608 reworkability/dismantlability, 1757–1767 scratch tests, 605–606 sealant materials, 598 storage, 1065–1069 substrate preparation, 1081–1087 thickness, 781–783, 787, 792 training and education, 1055 types, 286, 287, 289, 293, 295, 297, 299–300, 311–312 vacuum mixing, 1060 vertical and horizontal mixing, 1061 waste, 1125 Adhesive-beam model, 670–672 Adhesive bonding, 174, 184, 185, 190, 192, 666, 1752–1757

Index adherends, 668 analytical solutions, 674 edge moment factor, 673 joint configurations, 666 peel stress, 670 single lap joint (see Single lap joint (SLJ)) Adhesive-joint fracture tests mode I double cantilever beam test, 529 mode I tapered double cantilever beam test, 534 mode I/II test, 535–542 peel test (see Peel test) Adhesively bonded joints, 2, 5 cohesive zone model, 845 crash of car structures, 847 delamination, 694–696 energy release rate, 693 fracture mechanics, 845 impact strength evaluation, 844–847 infinite lap-strap joints, 834–839 single lap joints, 839–840 stress analysis, 834–844 stress singularity criterion, 845 stress–strain criteria, 844–845 Adhesive selection aesthetics, 421 costs, 426 fabrication issues, 427 form and processing conditions, 422 influencing factors for, 411 joint design, 414 joint perfomance, 412–413 manufacturers/suppliers, 433–437 manufacturing needs/constraints, 423 online assistance, 437 research and technology organisations, 429 substrate type, 415–422 surface pretreatments, 414–415 Adhesives implemented via a chemical process (AICPs), 357–369 Adhesives implemented via a physical process (AIPPs), 350–357 Adhesives, nonstructural, see Nonstructural adhesives, durability Adsorption isotherms, 246 Adsorption theory adhesive bond strength, 19 environmental stability, 19 oxidized aluminum surface, 17 rough surfaces, 20–21 thermodynamics of adhesion, 20 wetting and spreading of, 16 Advanced analysis methods, 732

Index Aerodynamic noise, 1388 AES, see Auger electron spectroscopy Aesthetics, 419–422 AFM, see Atomic force microscopy Ageing, 1013, 1016, 1017, 1020, 1021, 1023, 1024, 1026, 1526 Agitators, 1062–1065 AICPs, see Adhesives implemented via a chemical process (AICPs) AIMD, see Ab initio molecular dynamics AIMEN, 429 AIPPS, see Adhesives implemented via a physical process (AIPPs) Aircraft structures, adhesively bonded joints adhesive characteristics, 1238–1243 bonded repairs, 1276 design, 1248–1253 durability, 1256–1260 fiber-polymer composite structures, 1253 industry-specific factors, 1277–1278 inspection, testing and quality control, 1265–1271 load redistribution, 1260–1264 surface preparation, 1243–1248 thermal mismatch, effects of, 1264 use of adhesive bonding, 1278–1281 Alcoholysis, 264 Alfa-olefin resin, 1133 Alignment, 1686 Alkyl cyanoacrylate type adhesives, 368 Allergies, 1119, 1121 Alternating current (AC), anodizing, 155–156 Alternating current-direct current anodizing (ACDC anodizing), 156–159 Alternating failure, 1202 Aluminum, 165, 170, 173, 178, 181, 184, 191 ceiling panel, 1384, 1385 extrusion, 1374 substrates, 1083 treatment, 1085 Aluminum surface treatments anodization, 174–178 etching, 178 grit-blast, 179–181 sol–gel processing, 178 American Society For Testing And Materials (ASTM) Amino-resins, 287 Anaerobics, 300, 315, 369 Analysis, 1202, 1238–1243, 1253–1256, 1568–1570 Analytical solution, 685–691, 694–697 Angle joint, 818

1787 Anisotropic conductive adhesives (ACAs), 1453, 1456–1458 Anodic films, 151–153 Anodic oxides physical factors affecting, 153 Anodizing, 149–151 alternating current, 155 aluminum, 22 Antenna, 1293, 1309, 1314, 1318, 1327 Anti-flutter bonding, 1341–1344 Antifouling, 1631 Antioxidants, 1519, 1522 Applications, 1535, 1537, 1538, 1541, 1546, 1553 and emerging areas, 1782 equipment, 1078 technologies, 1081 Aqueous dispersions, 297 Aqueous solutions of sodium dichloroisocyanurate, 1492 Arcan, 506 joint method, 508 test, 505, 508, 509 Ariane, 1287, 1289, 1291, 1293, 1297, 1316 Arrhenius equation, 1046 Association for Innovation, Research and Technology Organisations (AIRTO), 429 Asymmetric double cantilever beam (ADCB) test, 540 Atmospheric agents bonded joints obtained by action of, 367–369 polymerization through action of, 365 Atmospheric plasma torch treatment, 1500 Atmospheric re-entry, 1298 Atomic force microscopy (AFM), 201, 1212–1213 Atomic oxygen (ATOX), 923, 928, 932 ATOX, see Atomic oxygen (ATOX) Atraumatic Restorative Treatment technique (ART), 1706, 1723 Attachment, 1629–1630, 1642, 1644, 1645, 1647, 1650, 1652, 1654, 1660, 1662 Attenuated total reflectance-Fourier transform infrared (ATR-FTIR), spectroscopy, 1223–1224 Auger electron spectroscopy (AES), 138, 216 Autohesion, 27, 29 Automatic moving system, 1106 Automating adhesive application, 1106 Automobile production requirements adhesives (see Automotive industry) anti-flutter bonding, 1341

1788 Automobile production requirements (cont.) hem flange bonding, 1341 hybrid joining, 1339 Automotive industry assembly line, 1344–1346 body shop, 1336, 1338–1344 durability, 1354–1358 failure types, 1363–1364 lap shear strength, 1351–1352 paint shop, 1336, 1344 power train, 1346–1347 press shop, 1336 quality control, 1364 repair and recycling, 1365 stress–strain diagram, 1353 trim assembly, 1336 Axially loaded butt joints, 498–500 B Back-face strain, 952, 953 Backing, 377, 379, 390 Batch process, 1060 BATTELLE, 429 Beakers, 1093 Benefits, 1238, 1246, 1259, 1278 Bioadhesion, 1642 Bioadhesive, 1597–1636 Bio-based adhesive, 1600, 1605, 1634 Biocompatible adhesive, 1608, 1630, 1636 Bioinspired adhesives, 1630–1631, 1642 antifouling, 1631–1632 biomimetic adhesives, 1631–1633 Bioinspired surfaces, 1643, 1666, 1669 Biological adhesives, 1625 attachment, 1629 construction, 1628–1629 defense, 1627–1628 locomotion, 1630 predation, 1626–1627 principles, 1611 Biomimetic adhesives, 1631 Biomimetics, 1642, 1643 Blister test, 594–598 Block copolymers, 35, 384 styrenic, 380, 400, 404–402 Block impact tests, 558–561 Blowing agents, 1751, 1758–1764 Boat, 1395, 1404, 1413 Body shop, 1336 Body structure, 1368, 1370 Bonded anchors, 1424–1426 Bonded blank technology, 1337, 1338

Index Bonded repairs, 1276–1277 Bonded structures, 859–860, 873 flexure, single lap joint in, 861–864 tension, single lap joint in, 860–861 Bonding agents, 1704, 1710 biocompatibility of, 1716–1717 definition, 1710–1713 evaluation of, 1713 leakage and, 1713 Bone, 1731, 1733, 1739, 1742 Boric-sulfuric acid anodizing (BSAA), 154 Boundary element method (BEM) advantages, 762 applications, 762, 770 vs. finite element method, 761 fundamental solution, 758, 760 representation formula, 758, 759 strong formulation, 755, 756 weak formulation, 757 weighted residual method, 756, 757 Boundary layer, 4 Brookfield type viscometer, 1142 Bulk tests, 506, 513, 566 Bulk torsion, 503–504 Burgers model, 979, 982, 984, 999 Butt joints, 797–798, 814–816 axially loaded, 498 in torsion, 509 C Calcium carbonate, 1506 Cambridge Consultants, 426 Capillary forces, 66 Capillary rise method, 88–89, 94 Carbohydrates, 1603–1604 blends, gum dispersions and exopolysaccharides, 1605–1606 cellulose derivatives, 1604–1605 starch and dextrins, 1605 Carbon black (CB), 1507 Carbon fibre, 1048, 1277, 1287, 1294, 1298 composite, 1287, 1294, 1304, 1311, 1326 Carbon fiber reinforced plasitics (CFRP), 1375, 1376, 1381–1384 Carbon nanofibers (CNFs), 1681, 1682, 1688, 1695 Carbon nanotube (CNT), 1679, 1682, 1686, 1687, 1691 Carboxylic acids, 1507 Car-Parrinello molecular dynamics (CPMD), 1591–1592 Carriers, 617, 618, 635–637, 650, 653, 654, 658

Index Cartridge guns, 1093 Casein adhesives, 288, 289 Castro-Macosko model, 30, 658 Catalysis, 264, 341 Cathodic disbondment, 237 Ceiling temperature, 464–465 Centrifugal mixing, 1064–1065 Certification, 1392 Chain disentanglement, 30, 32 Chain extender, 1512 Chain-growth polymerizations, 359 Chain pullout, 30, 32, 34 Charged particles, 920, 931 Chase-Goldsmith model, 651–653 Chemical adhesion, 1198 Chemical bonding, 51–53, 62 Chemical compatibility, 821 Chemical equivalent, 1140, 1146 Chemical properties, of sealants, 820–822 Chemical reaction, 463–464 Chemorheology, 658 ChemQuest Inc., 426 Chromic acid anodizing (CAA), 134, 153 Chromic acid etched (CAE), 949 Civil construction concrete construction, 1420–1432 glass construction, 1432–1437 steel construction, 1440–1443 timber construction, 1437–1440 Coatings blister test, 596 indentation tests, 602–603, 609 residual stress, 607 scratch tests, 603–604, 609 Cobalt, 60 (Co60), 925 Coefficient of thermal expansion (CTE), 607–609 Cohesion, 4 Cohesive failure, 596, 600, 694 Cohesive fracture, 1155, 1156, 1168 Cohesive zone model (CZM), 706, 845 Collagen, 1740, 1741, 1744 Commercial adhesives, 945 Compact tension test, 526 Compatibilisers, 33 Composites, 188–189, 1398, 1399, 1402, 1405, 1409, 1414, 1535, 1538, 1553–1406 material, 637–639 resins, 1704, 1709, 1714–1715 restorations, 1715 Comprehensive isothermal chemoviscosity model, 659 Composition, 151–153, 319–342

1789 Compression set, 825 Compressive tests, 491, 500–502 Concrete construction, 1420–1421 adhesive bonding, 1421–1422 flooring, 1423 pavement marking tapes, 1423–1427 structural reinforcement, 1427–1432 tiles, 1422–1423 Condensation, 263–264 Conditioned environment, 1034 Conductive adhesives, 310–311, 314–315, 585–588 Conductive filler materials, ECAs carbon-based nanomaterials, 1461–1463 conductive polymers and PEDOT hybrid metallic fillers, 1460–1461 PSS, 1463 Conductive nanocomposite adhesives, 299 Confocal microscopy, 1213–1214 Constant amplitude fatigue (CAF), 944, 947, 953, 965 Constant strain rate, 652 Constitutive behavior, 622–623 Consumer protection, 1119 Contact angle capillary rise method, 94 dew drops, 94 fiber, 95 goniometry, 92 hysteresis, 91, 107 puddle, 94 Wilhelmy plate technique, 94 Contact line dewetting, 110 dynamic contact angle, 106 energy dissipation, 104 equilibrium condition, 102 spreading rate, 108 Contact mechanics, 56 Contact resistance, 1465, 1468, 1469 Contact splitting, 1649, 1650 Continuous process, 1060 Continuum behavior, 616 Continuum damage mechanics (CDM), 964, 965 Core-shell rubber, 1138 Cork soles, 1490 Corner joint, see Angle joint Corona discharge, 1493 Corrosion, 258, 271, 272, 274–276, 278, 279 Corrosion protection films organosilanes as, 274 Cosmic rays, 920

1790 Cost, 411–413, 418, 419, 422 Coupling agent, 271 metal-based, 278 non-silane, 277 silicon-based, 278 CPMD, see Car-Parrinello molecular dynamics (CPMD) Crack growth partitioning approach, 969 Crash suitable adhesive, 1353 Crazing, 31, 32, 34 Creep behavior, 632–633 fatigue, 948, 966–970 modulus, 993, 994 predictive methods, 1005 relaxation test methods, adhesively bonded joints, 1000 retardation test methods, adhesively bonded joints, 991 rheometric test procedures, 989 stress relaxation and static load, 1003 superposition principles, 986 viscoelastic models, 977 Critical relative humidity, 883 Crosslinked adhesives, 388 Crosslinkers, 1508, 1515 Cryogenic temperature, 1289, 1295, 1329 Crystallinity, 1146, 1504, 1517 CTE, see Coefficient of thermal expansion (CTE) Cumulative fatigue damage model, 656 Cure, see Hardening process Curing, 1114, 1523 Cyanoacrylates, 310, 1734–1736 Cycle mix effect, 947 Cyclic loading, 942 Cylindrical joints, 775, 801–803 D Dahlquist’s criterion, 375 Damage mechanics, 964–967 models, 647–657 Damping bonded structure, level in (see Bonded structure) bondline thickness, effect of, 870 damping ratio, 858 joint manufacture, 864–865 logarithmic decrement, 857–858 loss factor, 858–859 overlap ratio, effect of, 867–869

Index specific damping capacity, 857 temperature, effect of, 869–870 temperature testing of joints, 866 vibration test apparatus, 865–866 Decorated sheet, 1371 Defects, 1175–1178, 1180, 1184, 1188, 1191 Deformation theory, 616–620 Degradation processes, 167–173 Delamination, 596–597, 602–604 Delivery form, 1034, 1038 Density, 434–443 Density-functional based tight-binding (DFTB), 1561, 1592 Density functional theory (DFT) gas sensors, 1582–1586 Hohenberg–Kohn theorem, 1578–1580 Kohn–Sham equations, 1580 Design, 414, 811–850, 1011, 1062–1064, 1093–1094, 1165, 1238–1243, 1248–1260, 1295–1301, 1538–1540 Dewetting, 102, 106, 110–112 Dextrin, 289 DFTB, see Density-functional-based tightbinding (DFTB) Differential movement, 825–827 Differential scanning calorimetry (DSC), 135, 484, 1219–1222 Differential thermomechanical analysis (DTMA), 949 Diffusion theory copolymer compatibilisers, 33–35 molecular dynamics, 29–30 polymer-polymer compatibility, 28 Digital volume correlation, 1226 Diglycidyl ether of bisphenol A (DGEBA), 135, 362 Diisocyanate, 1511 Diluents, 324 Dipalmitoylphosphatidylcholine (DPPC), 1590 Directional stability of cracks, 1477 Directives, 1025 Dismantlable adhesives, 1752 applications, 1757 types and characteristics, 1757 Dispersion adhesive, 355–356, 1678, 1684–1685 hazard potential, 1038 storage and shelf life, 1037 water-based adhesive group, 1037 Dominant damage approach, 969 Dosing process, 1108

Index Double cantilever beam (DCB), 946, 959, 960, 962 Dozer, 1057, 1077, 1080 2D Periodic boundary conditions (PBC), 1565 DPPC, see Dipalmitoylphosphatidylcholine (DPPC) Drop weight method, 88–89 Dry adhesion, 1650, 1669 Dry adhesives, 1668 Dry and wet epoxides, 471–472 Du Noüy ring technique, 90–91 Durability, 167, 169, 174, 177, 179, 181, 184, 192, 1779–1780 degradation mechanisms, 1410–1411 marine aging effects, 1411–1412 of nonstructural adhesives (see Nonstructural adhesives, durability) testing, 1412 Dynamic adhesion, 1644 Dynamic analysis, 718 Dynamic contact angle, 106 Dynamic finite element analysis, 843 Dynamic fracture energy, 562, 575 Dynamic mechanical analyzers (DMA), 484–485, 608 Dynamic mechanical thermal analysis, 484–485 Dynamic mixer, 1103 Dynamic response analysis, 720 Dyne inks, 206 E Earth, 1289, 1300, 1309–1311, 1325, 1326, 1329 Earth geostationary orbit, 1328 Eco-conscious materials, 1754 Ecological and economical development of innovative strategy and process for clean maintenance and dismantling further recycling (ECODISM) project, 1760, 1761 Edge joint, see Seam joint Edge moment factor, 673, 674, 677–681 Education, 1055, 1152, 1167–1168 Elasticity, linear, 382 Elastic limit, 616–619, 629, 634–636, 650, 651, 659 Elastic–plastic, 601 fracture mechanics, 525 Elastomer-epoxies, 304 Elastomeric adhesives, 598, 609 Elastomers (rubbers), 288, 298, 306

1791 Elasto-plastic fracture mechanics (EPFM), 959, 967 Electrically conductive adhesives (ECAs), 1450 in advanced electronic packaging technologies, 1451 anisotropic conductive adhesives, 1456 conductive filler materials (see Conductive filler materials, ECAs) isotropic conductive adhesives, 1454 non-conductive adhesives, 1459 Electrical properties, 1682, 1690–1691 Electrochemical treatments, 149–159, 1493 Electrolytic deoxidizing process (EPAD), 154 Electron microscopy, see Scanning electron microscopy (SEM) Electrostatic theory, 25 Element, 426 Emission of organic solvents, 1124 Emulsifiers, 1514 End-loaded split (ELS), test, 536 End-notch flexure (ENF), test, 536 Energy consumption, 1368 Energy dispersive X-ray (EDX) spectroscopy, 1215–1216 Energy dissipation, 859, 860, 867 Energy minimization, 1566 Enforced displacement, 753, 754 Engineering polymers, 285–290 Engineering thermosets, 287 Entropy contribution, 620 Environment ageing, 893, 908 assessment, 1123 compatibility, 1118 factors, 1016 hazards, 1012 stability, 19–20 testing, 1472, 1474, 1476–1478 Environmental protection, 1033, 1123 adhesives, 1125 air, 1124 soil, 1124 water, 1124 Epoxy adhesives, 302–305, 361, 933, 936, 1334, 1339, 1351, 1356, 1370, 1382 Epoxy resins, 340 Equilibrium contact angle, 102, 107, 111 Equivalent sun hours (ESH), 926, 928 Etching, 133, 138, 143, 146, 148 Ethylene-vinyl acetate (EVA), 1501–1502 copolymers, 357 hot melts, 472–473 Ethylene vinyl acetate hot melts, 472–473

1792 The Euler and Timoshenko beam theories, 670, 674, 678, 681, 682, 685–688, 690, 695 European Association of Research and Technology Organisations (EARTO (EU)), 426 EVA, see Ethylene-vinyl acetate (EVA) Evaporation, solvent/water, 461–462 Exchange-correlation energy, 1581 Excitation, 929 Expandable graphite, 1764, 1765, 1769, 1770 Extracellular polmeric substances (EPS), 1631 Eyring theorem, 634–635 F Failed adhesive joint, forensic investigation of, 233–241 Failure analysis, 730 Failure criteria cohesive zone model, 845 continuum mechanics, 691–692 fracture mechanics, 692–698, 845 stress singularity criterion, 845 stress–strain criteria, 844 Failure path, 525, 541, 551 Falling weight test, 587 Fatigue, 1694–1695 crack growth, 946, 947, 950, 952, 958–964, 966–969 creep fatigue, 966 cycle, 957 damage mechanics methods, 964 environmental factors, 948 fracture mechanics methods, 959 impact fatigue, 970 initiation and propagation, 945–946 load factors, 947 life, 655, 657 phenomenological methods, 955 prediction methods, 950 prevention, 942 strength, 1357, 1359–1360 testing, 946 total-life methods, 950 Fatigue crack growth (FCG), 961, 964, 966, 967 Fiber(s), 285, 287, 288, 294, 299, 304, 311 Fiber-polymer composite structures, 1253–1256 Fibrillar adhesives, 1643, 1655, 1664, 1666 Fibrin, 1731, 1738–1742 Fibrous proteins, 1607 Fickian diffusion, 895–897 Field problems, 742, 763

Index Fillers, 325–331, 1506 Fillet, 775, 778, 787–788, 791, 794, 804, 805 Film adhesives, 296 Finite difference method (FDM) advantages, 768 applications, 768, 770 vs. finite element method, 768 one-dimensional problem, 763, 767 Finite element analysis (FEA), 657, 896, 898, 908, 913 adhesive joints, 726 application of, 707, 720 cohesive zone modelling, 706 damage, 705, 707, 730, 732, 734 dynamic analysis, 718 field problems, 737 high performance compute clusters, 705 isoparametric elements, 712 method, 703, 708 multi-physics, 703, 704, 735, 736 non-linearity, 714 smoothed particle hydrodynamics, 706 thermal analysis, 716 virtual crack closure technique, 706 Finite element method (FEM), 1778 application of, 720, 770 commercial packages and usages, 742, 743 fundamentals of, 709 history, 703 isoparametric elements, 712 natural coordinates, 711 submodel (see Submodel) substructures (see Substructures) First-in first-out, 1149 Fixed-arm peel test, 544 Fixed-ratio mixed-mode (FRMM) test, 540 Flame retardant, 1033, 1043 Flammability, 1309 Floor adhesives, 1423 Flory-Huggins interaction parameter, 28 Foam adhesives, see Elastomeric adhesives Foolproof, 1166, 1169 Forces involved in adhesion, 38 Form, 410–412, 414, 415, 418–420, 422, 429 Formaldehyde, 1735, 1737 Formo-phenolic network, 361 Formo-phenolic resin precursors, 360 Formulation, 346, 349–352, 355, 359, 363–370 Fowkes approach, 123–126 Fracture, 1679 adhesive-joint specimens (see Adhesivejoint fracture tests) analysis, 959

Index bulk adhesive samples, 526–528 energy, 13–14, 23 mechanics, 706, 730, 732, 734, 735, 845, 959–964 stress, 14 toughness, 528, 546, 550, 551, 1691–1694 Fraunhofer Institute for Manufacturing Technology and Advanced Materials, 426 Free energy of mixing, 28–29 Free vibration, 719 Free volume, 440, 441, 469–471 Free volume theory, 469–471 Friction, 1599, 1644 Friction stir welding (FSW), 1369 Full-condensation adhesives, 361–365 Function, 284, 290–295, 301, 308, 316 Functional components amorphous polyalphaolefin, 1137–1138 core-shell rubbers, 1138 functional acrylic materials, 1136 oxetane polymers, 1136 silane coupling agents, 1135 thermally expandable microcapsules, 1138 Functional fillers carbon nanotubes, 1138–1139 fine calcium carbonates, 1139 micro balloons, 1139 Fundamental adhesion, 15 Fundamental solution, 758, 760 G Gecko, 1643, 1645, 1647, 1648, 1652, 1654, 1655, 1664, 1666, 1668, 1743 Gelatin, modified, 1737 Generalized gradient approximations (GGA), 1581 Geometrically-linear and nonlinear analyses, 679, 680, 690–691 Geometry optimization, 1566 Geostationary missions (GEO), 922, 925, 928 Geostationary orbit, 1328 GGA, see Generalized gradient approximations (GGA) Gibbs’ model, 75–82 Girifalcsood concept, 121–123 Glass construction, 1432 adhesive bonding, 1432–1433 elements, 1436–1437 glass facades, 1434–1436 laminated glasses, 1433–1434 Glass-ionomers, 1719, 1721, 1722, 1724

1793 Glass transition, 856, 866 point, 949 Glass transition temperature (Tg), 467–469, 933, 934, 948, 949 Glue, 346, 351, 353, 355, 360, 361, 363, 368 Gluing, 1441, 1442 G-modulus, 1350–1352, 1361, 1362 Graphene, 1680, 1682 Al-doped graphene, 1585 intrinsic, 1585 Graphite, 329 Greenhouse gas emission, 1754 Grit-blast treatment, 179 Gun, 1069, 1070, 1072, 1074, 1076, 1078, 1079 H Hardening process ceiling temperature, 464–465 cooling, 463 chemical reaction, 463–464 evaporation of solvent/water, 461–462 kinetics, 465–467 kinetic approach to equilibrium, 465–467 latex adhesives, 462–463 pot-life, 464 water vapor reaction, 467 Hardness, 601, 602 Hazardous material, 1033 Hazard potential, 1120 Hazard symbol, 1036 H2CO-Al-doped adsorbed systems, 1586 H2CO-graphene system, 1586 Health protection adhesive selection, 1123 chemically curing adhesives, 1122 consumer, 1119 physically hardening adhesives, 1120–1122 risks, 1119–1120 work, 1119 Heat-activated films (HAF), 1041, 1042 Heat-activated non-tacky solids, 295 Heat conduction, 716 Heat-curable rubber (HeR), 310 Heat flow, 716 Heat of polymerization, 465–466 Hem flange bonding, 1341–1342 Hemostasis, 1737, 1741 Hierarchically structured adhesives, 1645 High cycle fatigue (HCF), 950, 951 High energy radiation, 936 Higher volume dispensing, 1072

1794 High performance compute clusters (HPCs), 705 High-speed trains, 1368, 1377, 1381, 1388 High temperature, 1289, 1298, 1299, 1301, 1325 adhesives, 292–293, 297–298, 314 Hohenberg–Kohn (H–K) theorem, 1578–1580 Holography, 1190 Honeycomb panel, 1375 Hopkinson bar, 563 HORIBA MIRA, 426 Horizontal mixing, 1061 Hot-melt adhesives, 291, 311, 313, 1034, 1757 hazard potential, 1040 hot-melt adhesive group, 1038–1039 storage and shelf life, 1039–1040 Hot melt dispensing, 356, 1072 Humidity, 1034, 1035, 1039, 1049 Hybrid adhesives, 287 bonding, 1357, 1360 joining, 1339–1341 joints, 775, 778, 792–796 structures, 1339 Hydrogel, 1738 synthetic, 1737–1738 Hydrolysis, 261–263 I Impacts energy, 577, 578 fatigue, 970–971 peel test, 1354, 1355 properties, 1464 Impact loads, adhesively bonded joints, 834 Impact tests block impact tests, 558 conductive adhesives, 585 drop-weight based tests, 577–584 environmental conditions, 588 fracture energy, 574–576 Hopkinson bar apparatus, 563 joints, 580 wedge peel, 561 Implicit solution, 1580 Incompatible interfaces, 32–33 Incompressibility, 620, 623, 624 Indentation tests, 601, 609 Induction, 1114 Induction heating (IH), 1115, 1758 Infinite lap-strap joints, 834 Infrared spectrometer, 1507 Initial and equilibrium spreading coefficient, 104, 108

Index Initial spreading coefficient, 108 Inorganic adhesives, 300 Insect, 1642, 1644, 1645, 1652, 1660, 1663, 1666, 1668 Inspection, 1155, 1164, 1168, 1265–1276 Interaction, 265–269, 272, 276 Interfacial failure, 1156–1157 Interfacial tension, 4, 76–77 Interphase concept, 639–647 Interlocking, 1490–1491 Intermolecular interactions, 117–119 Internal forces, 751 International organization for standardization (ISO), 1140–1142 Interphase, chemistry, 225, 228 In vacuo testing, for provision of interfacial failure surfaces, 249–253 Invariants, 617, 622, 623, 631, 660 Inverse formulation, 757, 758 Inverse gas chromatography, 209 Ion-exchange layer, 1722, 1725 Ionic groups, 1512 Ionisation, 929 Iosipescu test, 506, 508 Irradiation effects electrical properties, 936 mechanical properties, 932 optical and thermo-optical properties, 934 outgassing, 937 Isochronous stress–strain, 993, 997 Isocyanate groups, 365, 1520 Isoparametric elements, 712 Iso-strain, 993, 996, 998 Isostress condition, 642 Isotherms, adsorption, 246–247 Isotropic conductive adhesives (ICAs), 1451, 1453–1456 ISQ, 426 J Japan industrial standards (JIS), 1140–1142, 1144, 1146 Joint design, 2, 4, 1778–1779 angle joint, 818 butt joints, 814 lap joints, 816 sealant properties, 820–825 seam joint, 819 Joint tests, 502, 508, 513, 515, 517, 557, 565, 569, 577 Jovian missions, 922

Index K Kelvin model, 626–627 Kelvin-Voigt model, 981–985 Kenics mixer, 1102 Kissing bonds, 1176, 1190 Kneader, 1062 Kohn–Sham equations, 1580 L Lap joints, 798–800, 816–818 Lap shear strength, 1351, 1352, 1356, 1357, 1361, 1386, 1393 Lap shear tensile tests, 847 Latex adhesives, 462–463 Launchers, 1288–1289, 1291–1293, 1295–1297, 1329–1330 Layered silicate, 1683, 1687, 1691, 1692 LCAO, see Linear combination of atomic orbitals (LCAO) Leather soles, 1490 Leather upper, 1488–1489, 1523 Lifetime prediction, 894 Light stabilizers, 338 Linear Aromatic Condensation-Thermoplastic Imide (LaRC-TPI), 364 Linear combination of atomic orbitals (LCAO), 1576 Linear elastic fracture mechanics (LEFM), 526, 959, 967 Linear elasticity, 382–385, 631 Linear overlap based on the Euler beam theory, 685 Linear viscoelasticity, 625–626 Link-atom method, 1588 Liquid-phase adhesives, 350–351 Liquid spreading dewetting, 110–111 dynamic contact angle, 106–108 energy dissipation, 104–106 equilibrium conditions, 102–103 retraction/dewetting, 102 spreading coefficient, 103–104 spreading rate, 108–110 Loading rate effect, 566, 569, 574, 578, 584 Load point, 759 Load update, 673–674, 677–679 Local spin-density approximation (LSDA), 1581 Locus of failure, 38 Long-term reliability, 1453, 1464, 1471 Loop tack test, 605, 606 Low cycle fatigue (LCF), 947, 951, 955

1795 Low earth orbit (LEO), 922, 925, 928, 1289, 1309, 1311, 1329 Low energy substrates, 1086–1087 Low energy surfaces, 1082, 1086 Low pressure plasma, 1086 Low-pressure RF gas plasma, 1493 Low temperature storage Arrhenius equation, 1046 basics of adhesive storage, 1043–1046 below freezing point, 1048 freezing point and room temperature, 1048 LSDA, see Local spin-density approximation (LSDA) Lubrication approximation, 104 Ludwik’s equation, 636 M Machine-processing of adhesives, 1093 double acting piston pumps, 1096 drum presses, 1095 eccentric screw pumps, 1098 peristaltic pumps, 1099 pressure containers, 1095 pumps, 1095 single acting piston pump, 1097–1098 valves, 1099 Macrofouling, 1631, 1632 Maglev trains, 1382 Magnesium substrates, 1083–1086 Main component, 1131–1133, 1136, 1141, 1146 Manual processing of adhesive systems, 1092–1093 Manufacture, 419, 424–425 Marine industry boats, 1395–1399 certification, 1392–1394 durability (see Durability) marine environment, 1404 mechanical behaviour, 1407–1410 offshore structures, 1401–1404 quality control, 1413 racing yachts, 1399–1401 repair and recycling, 1413–1414 surface preparation, 1404–1407 Material properties, stress–strain relations, 832–834 Material safety data sheet (MSDS), 1032 Material standards crystallinity, 1146 fillers, 1147 pigments, 1148 plasticizers, 1147

1796 Material standards (cont.) polymers, 1146 solubility parameters, 1146 solvents, 1146 tackifiers, 1146 thickening agents, 1148 Material transfer, 1078 Maximum bubble pressure method, 87–88 Maxwell model, 627–629 McBain and Hopkins, 25 Mechanical behavior, 616, 619, 650, 654, 657–660 mismatch, 1168 properties, 823–825, 1687 theory, 15, 20–25 Mechanical stress environmental ageing, 893, 908 failure modelling, 907–912 Fickian diffusion, 895 modelling diffusion, 895 moisture diffusion, 895 non-Fickian diffusion, 897 strain modelling, 902–907 Mechanical theory of adhesion, 1487, 1498 macro and microroughness, 21–23 mechanism, to rough surfaces, 23–25 Membrane, 595–598 Metal(s), 132, 133, 135, 137–139, 143–145, 149–151, 154 Metal-based coupling agents, 278 Metal inert gas (MIG), 1369 Metallic adherends, 137–141 chemical treatments for, 145–148 Metallic substrate, 229–231 interactions with, 265–266 Metal oxides, 1518, 1522 Metering accuracy, 1108–1112 Method of application, 312–316 Method of reaction, 284, 295–301 Micro balloon, 139 Microfabrication, 1642 Microfibrous surface, 22–23 Microporous surface, 22 Microscopy, 1204, 1211, 1228 Microstructure, 1666, 1667, 1669, 1670 Microtomy, 234, 241, 243–245 Minimum film-formation temperature (MFFT), 462–463 Mixed-mode I/II adhesive-joint tests, 540 Mixing devices, 1102–1104

Index Mode I double cantilever beam test, 529–534 tapered double cantilever beam (TDCB), test, 534–535 Mode II adhesive-joint tests, 535 Modified Bingham model, 651 Modified thick-adherend shear test, 1202 Modulus, 1682, 1687 Moisture absorption, 885, 900, 910 Moisture curing adhesives, 300, 315 Moisture-reactive hot-melt polyurethane adhesives, 1516 Molecular dynamics (MD) energy minimization, 1566 force field, 1562–1564 implementation, 1571–1576 isolated boundary condition, 1564 Monte Carlo method, 1562 periodic boundary condition, 1564 thermodynamic properties, 1568 time-dependent properties, 1569–1570 velocity Verlet algorithm, 1567–1568 water models, 1570–1571 Molecular mechanics (MM), 1561, 1566, 1585–1590, 1592 Molecular orbital (MO) theory, 48 atomic orbitals, 1576–1578 density functional theory, 1578–1581 Molecular weight, 1136, 1141, 1146, 1147 Monomer, 392, 401 Monte Carlo method, 1562 Mooney–Rivlin equations, 623–625 Mooney viscosity, 1133 Multivariate analysis (MVA), 225 Mussel, 1630, 1633, 1742, 1743 N Nanoadhesion, 62 Nanocomposite adhesives, 294–295 Nano-indentation, 1212 Nanoparticles, 1678 definition, 1680 equi-axed, 1680 manufacturing using, 1684–1685 plate-like, 1683 reinforcement, 1679–1680 Natural adhesives, 289–290, 1599–1600 carbohydrates, 1603–1606 natural polyesters, 1602–1603 natural resins, 1600–1601 natural rubber, 1601

Index polyphenol based adhesives, 1608–1611 protein based adhesives, 1606–1608 Natural polyesters, 1602 Natural resins rosin, 1601 terpene resins, 1601 Natural rubber, 289, 400, 1601–1602 N-chloro-p-toluensulphonamide solutions, 1492 Neumann approach, 126–127 Newtonian behavior, 443–446 Non-conductive adhesives (NCAs), 1453, 1458–1459 Nondestructive testing (NDT), 1178–1180 acoustic emission, 1188 acoustic methods, 1187 defects, 1175 definition, 1172 optical holography, 1190 sonic and ultrasonic vibrations, 1184–1186 spectroscopic methods, 1186 thermal methods, 1188–1190 ultrasonics, 1180 visual, 1190 Non-Fickian diffusion, 897–902 Non-linear finite element analysis, 714 Nonlinear analysis, 632, 633, 642 Nonlinear mechanical models, 631 Nonlinear overlap based on the Timoshenko beam theory, 685 Nonlinear viscoelasticity, 631–632 Non-Newtonian behavior, 446–449 Non-silane coupling agents, 277 Non-structural, 1009–1027 Nonstructural adhesives, durability adhesive selection, 1012 antioxidants and stabilizers, 1018 applications, 1014 electronics, 1024 elevated temperature and UV radiation, 1016 insulating glass units, 1023 joint assembly, 1015 joint design, 1011 mixing procedures, 1014 plasticizers, 1012 pressure sensitive adhesives, 1019 storage conditions, 1014 stress and fatigue, 1022 surface preparation and primer selection, 1013 sustainability, 1026

1797 water, humidity and saline solutions, 1021 workmanship, 1015 Normal conditions, 1044 Notched plate shear method (Arcan), 506 NPL, 426 Numerical crack growth integration (NCGI), 960, 961 Numerical integration, 712 Numerical methods boundary element method (see Boundary element method (BEM)) finite difference method (see Finite difference method (FDM)) finite element method, 742, 743 submodel, 751, 754 substructures (see Substructures) Nylon-epoxy, 302 Nylon fabric, 1490 O Offshore, 1392, 1401, 1410, 1534, 1535, 1537, 1538, 1543, 1548 Oils, 402 One-part adhesives, 309 Optimizing performance, 386–388 Organosilanes, 258–263, 266, 268–275, 278, 279 as adhesion primers, 272–274 as corrosion protection films, 274–275 definition, 259 origin of, 259–260 as primers, 271–272 silanes chemistry, 260 uses of, 268–271 Oscillatory failure, 1202 Overlap, 778, 780, 782–787, 790–796, 798, 799, 801, 803, 805 Oxetane polymer, 1136–1137 Oxidative degradation, 336, 479–482 Oxidized aluminum surface, 17 P Packaging, 1032, 1033, 1035, 1040, 1042, 1044, 1049, 1065, 1068, 1087 PA Consulting Group, 426 Paint shop, 1336–1337 Palmgren–Miner (P–M) rule, 953 Parafin wax, 483 Partial differential equation, 755 Particle adhesion, 27 Patterned adhesives, 1666, 1667 PBC, see Periodic boundary condition (PBC)

1798 PEC, see Predicted environmental concentration (PEC) Peel, 544, 1778 adhesion, 376–378 geometry-dependent, 544–545 geometry-independent, 545–550 plies, 1246, 1247 stresses, 1239, 1240, 1242, 1243, 1250, 1251, 1254–1256, 1275, 1276 Pendulum test, 558, 561, 578, 584 Performance, 376–382, 386–388, 393, 394, 396, 398 Periodic boundary condition (PBC), 1564 Personal protection equipment, 1119 Phase boundary, 73–82, 85, 95, 99 energy balance, 82–83 Gibbs’ model, 75–82 interphase, 74–75 thermodynamic equilibrium, 84–86 Phenolic resins, 287 Phosphoric acid anodizing (PAA), 134, 153 Physical adhesion, 1198 Physical aging, 436, 437, 441, 442 Physical bonds, 168 Physical form, 284, 289, 296, 301–312, 316 Physical properties, of sealants, 822–823 Pin-and-collar test, 515 Plasma treatment, 1086 Plastic flow, 616–617 Plasticizers, 331–332 Plastics, 189–190 PMMA, see Polymethylmethacrylate (PMMA) Point, 860, 873 Poker chip test, 498 Polyester resins, 287, 304 Poly(3,4-ethylenedioxythiolphene) poly (styrenesulfonate) (PEDOT:PSS), 1463–1464 Polyacid-modified, 1715 Polyamide hot melts, 357, 473–475 Polychloroprene adhesive, 353, 1489, 1501, 1503, 1517. 1518 solvent-born, 1517 waterborne, 1520 Polyimide, 305–306 Polymeric adherends, chemical treatments for, 148–149 Polymeric methylene diphenyl diisocyante (PMDI), 249, 251 Polymeric substrate, interactions with, 266–268

Index Polymerization AICPs, 358–359 through action of atmospheric agents, 365–366 Polymerized adhesives, 360–361 Polymer-polymer compatibility, 28 Polymers, 285 substrates, 231–233 treatment, 141–142 Polymer-to-metal adhesion, 241 Polymethylmethacrylate (PMMA), 17 Polyol, 1512 Polyphenol based adhesives, 1608 Poly(β-pinene), 353 Poly(styrene-b-butadiene-b-styrene), 354 Polyurethane adhesive, 306, 364–308, 1485, 1489, 1494, 1496, 1502, 1503, 1737 solvent-born, 1503–1508 solvent-free, 1515–1517 waterborne, 1508–1515 Polyurethane-coated fabrics, 1490 Polyurethane soles, 1502 Polyvinyl acetate (PVA), 286 Polyvinyl alcohol, 352 Polyvinyl chloride (PVC) coated fabrics, 1489 soles, 1502 uppers, 1489 Polyvinylpyrrolidone, 352 Poromerics, 1490 Porosity, 1175, 1263 Postfracture analysis macroscopic and mesoscopic optical observations, 1204–1209 microscopic optical observations, 1209–1215 physico-chemico analysis, 1215–1224 Pot-life, 464 Power law, 632, 636 Power train, 1337 Practical adhesion, 12–16, 19, 21–23, 36 Precursor film, 108, 109 Predicted environmental concentration (PEC), 1123 Predicted no-effect concentration (PNEC), 1124 Preparation methods, 1014, 1022 Prepolymer method, 1510 Press shop, 1336 Pressure sensitive, 376, 379, 382, 383, 386, 388, 396, 401, 402, 405 Pressure sensitive adhesive (PSA), 1035, 1037, 1378, 1743, 1745–291, 295, 311, 605, 606, 1380–312, 369, 374

Index Pre-treatment, 411, 413–415 Preventive measures, 1056 Primary Adhesively Bonded Structures Technology (PABST), 1243, 1245, 1251, 1258, 1261, 1278 Primary adhesives, 320–324, 331, 332, 339, 342 Primary resins, 321–323 Primers, 4, 300, 315 Probabilistic design, 1152, 1168 Probe tack tests, 605–606 Process design, 1164–1165, 1168 Processing, 1683, 1685, 1686 Product description, 1034 Properties, adhesive material, 917 Protective measures, 1120 Protein, 1730, 1733, 1734, 1737, 1742, 1744 Protein based adhesives, 1606 fibrous proteins, 1606–1607 tissue engineering, 1607–1608 vegetable proteins, 1608 Protein-based glues, 289 Proton irradiation, 932, 934 PSA, see Pressure sensitive adhesives (PSA) Pumps, 1094–1099, 1106, 1107, 1112, 1115 Pyrogenic (fumed) silicas, 1506 Q Quality assurance, 1032 Quality control, 1152, 1270, 1780–1781 adhesives, use of, 1165–1167 acceptance inspection, 1140–1143 foolproof, 1165 operator’s tasks, reduction of, 1165 inspection for shipment, 1143–1145 raw material acceptance, 1139–1140 raw materials in storage, 1143 Quantum mechanics/molecular mechanics (QM/MM) method his-tags, 1588–1590 link-atom method, 1588 Quantum mechanics (QM), 1561 Quick stick test, 606–607 R Radiations, 924 Radiography, 1174, 1189 Raman spectroscopy, 1216 Ramberg–Osgood model, 566 Random copolymer, 35 Rate effect, 617, 658

1799 Raw materials fillers, 1134 functional components, 1135–1138 functional fillers, 1138–1139 main components, 1131–1133 material standards, 1145–1148 pigments, 1134 plasticizers, 1134 quality control procedures for, 1139 solvents, 1133 tackifier, 1133 thickeners, 1134 types and roles of, 1130 Reactive adhesive, 1033, 1034 hazard potential, 1043 reactive adhesive group, 1040–1041 storage and shelf life, 1041 Recyclable design, 1125 Recycling, 1026 Reducing of weight, 1368, 1377 Reflector, 1293, 1298, 1309, 1314, 1318, 1327, 1329 Regulation, 1025 Reinforcements, 332–335, 801–802 Relaxation, 979 adhesively bonded joints, 1000 Burgers model, 984 Kelvin-Voigt model, 983 Maxwell model, 983 stress relaxation and static load creep, 1003 time, 627–629, 632–634, 636 Release liner, 390–392 Reliability, 1471–1472 harsh environments, storage test in, 1473–1478 mechanical, 1478 test vehicle design, 1472–1473 Repairs, 1534, 1535, 1538, 1539, 1541, 1543, 1545, 1547, 1549 techniques, 775, 796–797 Representation formula, 758, 759 Reptation theory, 29 Research and Technology Organisations (RTOs), 429 Residual stresses, 595, 602–603, 607 Residuum, 756 Resin-modified glass-ionomers, 1717, 1719, 1720, 1722 Resins, 1519, 1522 primary, 321 Responsible care, 1118 Restriction of hazardous substances (RoHS), 1145

1800 Retardation, 979, 991 Reusing, 1755 Reversible adhesion, 1669 Rework, 1755, 1757 Rheological tests, 380–385 Rheology, 443–445 Rheometer, 1142 Robot, 1106 Rolling ball tack test, 605, 606 Rolling stock, 1368 Room temperature hot-melt adhesive, 1038–1040 reactive adhesives, 1040–1043 solvent-based adhesives, 1035 water-based dispersion adhesive, 1036–1038 Room temperature vulcanizing (RTV), 309 Rosin, 1600–1601 Roughness, 168, 169, 181, 182, 186–188 Rubber adhesives, 1372, 1379 Rubber elasticity constitutive elastic rubber behavior, 622–623 material properties, 620–622 Mooney–Rivlin equations, 623–625 Rubber sole, 1490–1501 Rules-of-thumb, 221, 271, 823, 1000, 1017 Rupture envelope, 996 S Safety, 1524 Safety data sheets, 1033, 1038, 1043, 1123 Sancaktar’s analytical model, 641 Sandwich structure, 1288, 1293, 1295, 1298, 1305, 1325–1326 Satellite, 1293–1295, 1297–1301, 1330 SATRA, 426 Scanning auger microscopy, 216 Scanning electron microscopy (SEM), 199, 1210–1212 EDX spectroscopy, 1216 Raman spectroscopy, 1216–1218 Scratch tests, 603, 609 Sealants, 4, 6, 290, 293, 298, 1010, 1012, 1015, 1017, 1019, 1022, 1024, 1027, 1237, 1277, 1731, 1737, 1739, 1741, 1777–299, 308, 314, 330, 598–600, 1376–1378 chemical properties, 820 differential movement tolerances, 826 joint design (see Joint design) mechanical properties, 823 physical properties, 822

Index Seam joint, 819 Secondary ion mass spectroscopy (SIMS), 16, 212, 221, 223 Secondary phases, 616, 637 Second generation acrylic adhesives (SGA), 1375 Selection, 409–430 Self-etching primers, 1711 Self-healing, 1085 Semicrystalline thermoplastic elastomer, 354 Semiempirical approach, 633 Service life, 1011, 1012, 1014, 1016, 1019, 1020, 1024, 1027 Shear, 379–380 stresses, 1257, 1261, 1263, 1265, 1272 tests, 491, 500, 502, 599, 609, 629 Shear adhesion failure temperature (SAFT), 379, 399 Shear test, 629 Shelf life, 460–461 hot-melt adhesives, 1039 reactive adhesives, 1041–1043 solvent-based adhesives, 1035–1036 water-based dispersion adhesives, 1037–1038 Shellac, 1602–1603 Shellfish, adhesives from, 1742 Shift factor, 630, 632–635, 642 Shoe bonding, 1486–1487 adhesive joints, testing of, 1523 methodologies for, 1484 sole materials, 1490–1502 upper materials, 1488–1490 Shoe industry adhesive joints, testing of, 1523–1526 polychloroprene adhesives, 1517–1523 polyurethane adhesives, 1503–1517 surface preparation (see Surface preparation) Shrink fit process, 1347 Sick house syndrome, 1133 Silanes adhesion promoter, 16 chemistry, 260 in formulations, 275–277 Silicate-based adhesives, 300 Silicon-based coupling agents, 278–279 Silicones, 308–310, 919, 932, 936, 937 adhesives, 365–366 glues, 935 rubber, 932, 934 tape, 935 Simple point charge (SPC) model, 1571

Index SIMS, see Secondary ion mass spectroscopy (SIMS) Simulations, 1559–1593 Singularity methods, 636–637 Single component dispensing, 1069 Single edge notched bend (SENB) test, 526 Single lap joint (SLJ), 513, 514, 839 adhesive-beam model, 670–672 adhesive joints with material nonlinearity, 682–684 adhesive stresses, 681 asymmetric and unbalanced joints, 686–691 average shear stress model, 668–670 composite, 680 coupled formulation for load update, 677–679 geometric nonlinearity and transverse shear, 674–677 load update and bending moment factor, 673–674 edge moment factor, 680 test, 513 two-dimensional elastic theory, 672–673 Single stranded DNA (ssDNA), 1574 Skin, 1730, 1732, 1735, 1736, 1741 adhesive, external application, 1744, 1746 Skin irritation, 1121 SLJ, see Single lap joints (SLJ) Smear layer, 1704, 1707, 1708, 1711 Smithers RAPRA, 426 Smoothed particle hydrodynamics (SPH), 706 Soft-block monomers, 401 Solar electromagnetic spectrum, 930 Solar spectrum, 920 Solder, 1744 Solid content, 1034, 1035, 1037, 1046 Solid-phase adhesives, 356–357 Solids content, 1512 Solubility parameter, 1133, 1146 Solutions, 258–264, 266, 268, 272–275, 278 Solvent, 1033, 1035, 1045, 1519 Solvent-based adhesives, 295, 1035–352, 1036–355 Solvents, 323–324 Sonics, 1185 Soybean adhesives, 289 Space environment components, 920 degradation mechanism, 929 irradiation effects, 932 mechanical fixation, 917 optical properties, 919

1801 properties, 917 radiations, 924 testing methodology, 923 Space structural bonding, 1302 SPC model, see Simple point charge (SPC) model Special mechanical tests blister test, 594 indentation tests, 601–603, 609 loop tack tests, 606 peel tests, 600 probe tack tests, 605 quick stick tests, 606 rolling ball tack tests, 606 scratch tests, 603–604, 609 shear tests, 599–600, 609 tensile tests, 598, 600 Specialty thermosets, 288 Specification sheet, 1143 Spectroscopic methods, 1186–1187 Speed mixer model, 1065 Spider, 1650, 1652, 1654, 1664, 1666, 1668, 1669 Spot welding, 1339 Spreading coefficient, 103, 109 Spreading of liquid, see Liquid spreading ssDNA, see Single stranded DNA (ssDNA) Standards, 1010, 1013, 1015, 1016, 1019, 1021, 1023, 1025, 1027, 1537 and recommended practices, 1537–1541 Starch, 289 based adhesives, 289 macromolecular structure, 351 Static condensation, 745 Statistics, 1153, 1155 Steel, 184–188 Steel construction, 1441 adhesive bonding, 1441–1442 bridge building, 1442 hybrid bonding, 1443 metal framework facades, 1442–1443 substrates, 1082–1083 Step-growth polymers, 296 Stepped-lap joints, 1255, 1260, 1275 Stiffness wearout, 955, 966 Storage conditions, 1032, 1033, 1042, 1044, 1048 Strain, 442, 443, 445, 446 energy density, 622, 623 hardening, 616 Strain-invariant polymer failure model (SIFT), 1239, 1256 Strain-life approach, 955

1802 Strap joints, 799 Strength wearout, 955–957 Stress, 442, 443, 445, 447 analysis, 727 concentration, 503, 505, 506, 509–510, 513, 518 relaxation, 475–478 singularity criterion, 845 strain behavior, 435, 453, 456–455 strain criteria, 844 strain diagram, 618, 619, 1353 waves, 831 whitening, 616, 618, 619, 658, 660 Stress–strain diagram, 618, 619 Stress free temperature (SFT), 607, 609 Stress relaxation, 475–478 Stress–strain relations of metals, 832 of plastics, 833 Stretch ratio, 622, 623 Strong formulation, 756 Structural acrylic adhesives, 4, 291, 311, 367, 461, 1200–1201 Structural bonding, 1196–1198 Structural timber engineering, 1439 Styrene butadiene rubber (SBR), 288 Styrene-isoprene-styrene (SIS), 401 Styrenic, 378–380, 382–384, 386, 390, 395, 400–402, 404–406 Styrenic block copolymers, 380, 400, 404 based adhesives, 379 Submodel coarse mesh, 751, 753 enforced displacement, 754 internal forces, 751, 754 Substrates, 1012, 1019, 1021, 1025 characterization, 229–234 metallic, 265 polymeric, 266 Substructures finite element model, 748, 751 one-dimensional bar elements, 743, 747 static condensation, 747 Suction, 1387 Sulfuric acid anodizing (SAA), 154 Superelement, 745, 746 Superposition, 1006 Boltzmann superposition principle, 986 correspondence principle, 987 principle, 986–989, 1005, 1006 time-temperature superposition principle, 988 Supplier test, 1139

Index Surface chemical analysis methods, 228 free energy, 133–135, 137 functionalization, 141 and interfacial tensions, 205 marking, 1343, 1351 topography, 199–204 unbonds, 1176 Surface contamination, 165, 188 problem, 133–137 removal of, 137–142 Surface energetics, 55, 65, 67 contact angle, 207 inverse gas chromatography, 209 simple test methods, 205 Surface preparation, 168, 181, 184, 187, 188, 191, 1348–1349 leather, upper materials, 1488 polymeric soles, 1501–1502 rubber sole materials, 1490 synthetic upper materials, 1489 Surface pre-treatment, 886 Surface tension, liquid capillary rise, 88–89 drop weight, 88–89 du Noüy ring, 90–91 maximum bubble pressure, 87–88 Wilhelmy plate, 89–90 Surface treatments, 1776 aluminum (see Aluminum surface treatments) ceramics, 190–191 composites, 188 degradation processes, 167 plastics, 189 pretreatment steps, 165–167 processibility, 173–174 steel, 184 titanium, 181 Surfactants, 356, 1522 Surgery, 1730, 1733, 1735 Sustainability, 1026 Sustainable development, 1118 Suture, 1730, 1731, 1736, 1739, 1744 Synthetic upper materials, 1489–1490 T Tack, 374–375, 377–378, 604–607 Tackification, 398 Tackified ethylene-vinyl acetate hot-melts, 286 Tackifiers, 398, 400, 402–338, 405, 1505 Tall oil rosin (TOR), 1601

Index Tan delta, 383, 385, 394, 407 Tap testing, 1187 Teeth, 1704, 1706, 1716, 1722, 1724, 1725 Temperature effects, 439–441 Temperature limits, 471, 482–483 Tensile tests, 491, 598–600, 629 Terpene resins, 1601 Testing, 1016, 1017, 1019, 1022, 1023, 1025, 1486, 1523 Testing of adhesive properties, 1777–1778 Test specimens, 136, 171, 273, 503, 510, 531, 1354, 1357 Test vehicle, 1472–1474, 1480 Theories of adhesion, 15–17, 20, 21, 25, 26, 28, 36, 38, 39, 1776 Thermal analysis, 716 conductivity, 479, 919 contraction, 823 cycling, 934, 935 expansions, 479, 823, 826, 918 insulation, 1291, 1295, 1297, 1299 mismatch, 1264–1265 stability, 1689 Thermal gravimetric analysis (TGA), 136 Thermally expandable microcapsules (TEMs), 1138 Thermally reversible adhesive, 483 Thermal mechanical analyzers (TMA), 607 Thermal properties of adhesives breakdown of, 479–483 glass transition temperature, 467–472 hardening process, 461–467 measurement of, 483–486 melting, 472–475 shelf-life, 460–461 thermal conductivity, 479 thermal expansion, 479 viscoelasticity, 475–478 Thermodynamics equilibrium, 84–86 models, 20, 119–127 potentials, 82–83 Thermogravimetric analysis (TGA), 485–486, 1223 Thermomechanical analysis (TMA), 1145 Thermoplastic(s), 285–286, 297, 305, 313 adhesives, 413, 1757 resins, 323 Thermoplastic polyimides (TPI), 305 Thermoplastic polyurethane adhesives, 1516–1517

1803 Thermoplastic rubber (TR) soles, 1490–1494 Thermoset(s), 287–288, 297, 301–302, 313 Thermoset plastics, 339 Thermosetting resin, 322 Thermo-structural adhesives, 363 Thick adherend shear test (TAST), 511 Thickeners, 1513–1514, 1522 Thin films, 596, 608 Thin sheet adherends, 513 3D digital image correlation, 1224–1226 Three-parameter model, 650 TID, see Total ionizing dose (TID) Timber construction, 1437 adhesive bonding, 1437–1439 engineered wood products, 1439 glued-in rods, 1439–1440 reinforcement of timber structures, 1440 Time-dependent fracture mechanics (TDFM), 967, 969 Time-of-flight, 212, 221–225 Time-of-flight secondary ion mass spectrometry (ToF-SIMS) dynamic SIMS, 221 galvanized steel sample, 224 MVA method, 225 poly(vinylidene fluoride) coating, 222 Time-temperature superposition principle (TTSP), 629–630, 986, 988–989, 1005, 1006 Tissue, 1730, 1732, 1734, 1737 Titanium, 168, 172, 181–184, 278, 1287, 1298, 1301, 1316, 1318, 1319, 1328 T joints, 775, 783, 797, 803–805 Torsion butt joints in, 509 modulus, 1363 stiffness, 1345, 1360, 1362 tests, 510, 517, 519 Total ionizing dose (TID), 925 Total-life methods, 953–954 Toughening, 1691, 1693 Toxicity, 1124, 1678 T-peel test, 544, 545 Training, 1168 Treatment, 131–160 Trim assembly, 1336–1337 TWI Ltd., 426 Two component dispensing, 1069–1072 Two-part adhesives, 390 Two part dispensing, 106 Typical bondings, 1335

1804 U Ultra-low angle microtomy (ULAM), 242, 243 Ultrasonics, 1174, 1180, 1185, 1187, 1191–1184 Ultrasonic testing, 1226–1229 Ultra violet (UV), 918, 926, 930 electron beam curing adhesives, 292, 312, 314 ozone, 1493, 1494 radiation, 930, 934 resistance, 1113–1114, 1347, 1357 simulation, 927 wavelength phenomena, 930 Unbalanced Young force, 107, 109 Urethane, 1515, 1517 V Vacuum mixing, 1060 Vacuum outgassing, 1289, 1308 Valves, 1094, 1096–1102, 1104, 1105, 1107, 1108 Vapor pressure osmometry (VPO) method, 1146 Van Allen belts, 921 Van de Graaff, 925 Van der Waals forces, 18, 27, 31, 45–47, 55 Variable amplitude fatigue (VAF), 944, 953 Vertical mixing, 1061–1062 Vibration, 853–875, 1779 damping, 1779 (see also Damping) Virtual crack closure technique (VCCT), 706 Viscoelasticity, 382, 977, 979, 986, 987, 989, 990, 996, 997, 999 linear, 625–626 nonlinear, 631–632 Viscosity, 434, 443–453 adhesive, 452 coefficient, 626, 629, 631 dilute solution, 449 filler content, 450 molecular weight dependence, 450 Newtonian behavior, 443 non-Newtonian behavior, 446 temperature dependent, 451 Viscous dissipation, 105, 108 Vision systems, 1087 Visual examination, 1190 V-notched beam shear method (Iosipescu), 505 Voids, 1175, 1176, 1180, 1181, 1183, 1185, 1187, 1189 Volatile organic compounds (VOCs), 138 Vulcanized natural rubber and styrenebutadiene (SBR) rubbers, 1494–1501

Index W Warwick Manufacturing Group, 431 Waste, 1124, 1125 Water, 1397, 1404, 1407, 1410, 1412, 1415 absorption, 1039 break test, 206, 225 Water-activated adhesives, 295 Water-based adhesives, 291–292, 312, 314, 351–352 Waterborne polychloroprene adhesives, 1520 adhesion promoters, 1522 antioxidants, 1522 composition of, 1520, 1521 curing agents, 1523 metal oxide, 1522 polychloroprene latex, 1521 resins, 1522 surfactants, 1522 thickeners, 1522 Waterborne polyurethane adhesives acetone method, 1509, 1510 adhesion properties of, 1511 chain extender, 1512 characteristics of, 1510 crosslinker, 1515 diisocyanate, 1511 emulsifiers, 1514 hard-to-soft segments/NCO/OH ratio, 1511 ionic groups, 1512 limitations, 1509 polyol, 1512 prepolymer method, 1510 solids content, 1512 thickeners, 1513 Water effect adhesive interfaces, 892–893 diffusion, 890 durability, 885, 888 humidity, 881–885 moisture absorption, 885 natural and accelerated aging, 887 stress, 888–890 surface treatment, 886–887 Water models, molecular dynamics, 1571 Weak-boundary-layer theory, 15, 36–38 formation, during bonding, 37 polymer surfaces, 37 Weak formulation, 756, 758 Wedge-peel impact test, 561–562 Weighted residual method, 743, 756

Index Weighting function, 756, 758 Welding, 1534, 1540, 1541, 1548, 1551–1553, 1734, 1744 Wet adhesion, 1505, 1515 Wetting, 72 contact angle, 72 free energy balance, 95–98 White light interferometry (WLI), 204 Wilhelmy plate technique, 89–90, 94 Williams–Landel–Ferry (WLF) equation, 630 Windshield adhesives, 1349 bonding, 1345, 1346, 1349, 1357 fatigue curve, 1358 strength, 1357 Wöhler diagram, 1358 Wood, 191 Work of adhesion, 14, 19, 54–68 Work of cohesion, 14 Workplace limit value, 1121 Work protection, 1118, 1119 Wound, 1735, 1736, 1746

1805 X XPS, see X-ray photoelectron spectroscopy (XPS) X-ray microtomography, 1214–1215 X-ray photoelectron spectroscopy (XPS), 17, 1218–1219 Beer–Lambert equation, 215 electron probe microanalysis, 216 X-ray photoelectron spectrum, 213 X-ray spectrometer, 1211, 1215 Y Yield, 616, 621–624, 627, 632, 650, 657, 659, 660 Young’s modulus, 528, 531, 543, 565, 575, 576, 581, 588 Z Zero-volume unbonds, 1176 Zinc oxide (ZnO) mechanism, 1582 Zinc polycarboxylates, 1718 Zinc stearate, 483 Zirconium, 278 Zisman approach, 120–121