Adhesive Bonding Technology and Testing 3527350519, 9783527350513

Table of contents :
Cover
Title Page
Copyright
Contents
Preface
Chapter 1 Simple Practical Demonstrations
1.1 Importance of Loading Mode on Bonded Joint Performance
1.1.1 Introduction
1.1.2 Equipment
1.1.3 Materials
1.1.4 Safety Precautions
1.1.5 Experimental Procedure
1.1.5.1 In Class
1.1.5.2 In the Laboratory
1.2 Surface Treatments and Methods to Evaluate Surface Energy
1.2.1 Introduction
1.2.2 Equipment
1.2.3 Materials
1.2.4 Safety Precautions
1.2.5 Experimental Procedure
1.2.5.1 In Class
1.2.5.2 In Laboratory
1.3 Stress Distribution Along the Overlap Length
1.3.1 Introduction
1.3.2 Equipment
1.3.3 Materials
1.3.4 Safety Precautions
1.3.5 Test Procedure
1.4 Visual Identification of Defects in Adhesive Joints
1.4.1 Introduction
1.4.2 Equipment
1.4.3 Materials
1.4.4 Safety Precautions
1.4.5 Test Procedure
1.5 Failure Analysis of Adhesive Joints
1.5.1 Introduction
1.5.2 Equipment
1.5.3 Materials
1.5.4 Safety Precautions
1.5.5 Test Procedure
Chapter 2 Production and Testing
2.1 Bulk Specimens
2.1.1 Introduction
2.1.2 Adhesive Pouring Technique
2.1.3 Metallic Mold
2.1.4 Adhesive Application
2.1.5 Curing Procedure
2.1.6 Machining Procedure
2.1.7 Testing Procedure
2.2 Thick Adherend Shear Specimens
2.2.1 Introduction
2.2.2 Metallic Mold
2.2.3 Surface Treatment of Adherends
2.2.4 Geometrical Control Using Shims
2.2.5 Specimen Manufacture
2.2.6 Final Specimen Preparation
2.2.7 Testing Procedure
2.3 Fracture Mechanics Specimens
2.3.1 Introduction
2.3.2 Metallic Mold
2.3.3 Surface Treatment of Adherends
2.3.4 Adhesive Spacers
2.3.5 Specimen Manufacture
2.3.6 Final Preparation of Specimens
2.3.7 Testing Procedure
2.3.8 Data Reduction Schemes
2.4 Single‐Lap Joint Specimens
2.4.1 Introduction
2.4.2 Surface Treatment of Adherends
2.4.3 Joint Manufacture
2.4.4 Final Preparation of Specimens
2.4.5 Testing Procedure
Chapter 3 Laboratorial Activities and Report Examples
3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints
3.1.1 Introduction
3.1.1.1 Joint Strength Prediction
3.1.2 Work Description
3.1.3 Materials
3.1.3.1 Adherends (Tables and )
3.1.3.2 Adhesive (Table )
3.1.4 Experimental Work
3.1.5 Report
3.1.5.1 Introduction
3.1.5.2 Experimental Procedure
3.1.5.3 Materials
3.1.5.4 Failure Load Prediction
3.1.5.5 Experimental Results and Discussion
3.1.5.6 Conclusions
3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints
3.2.1 Introduction
3.2.2 Work Description
3.2.3 Materials
3.2.3.1 Adherends
3.2.3.2 Adhesives (Table )
3.2.4 Experimental Work
3.2.5 Report
3.2.5.1 Introduction
3.2.5.2 Materials
3.2.5.3 Prediction of the Failure Loads
3.2.5.4 Experimental Results
3.2.5.5 Discussion
3.2.5.6 Conclusions
3.3 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints
3.3.1 Introduction
3.3.2 Work Description
3.3.3 Materials
3.3.3.1 Adherends:
3.3.3.2 Adhesives:
3.3.4 Experimental Work
3.3.5 Report
3.3.5.1 Introduction
3.3.5.2 Experimental Details
3.3.5.3 Prediction
3.3.5.4 Experimental Results
3.3.5.5 Failure Surfaces
3.3.5.6 Conclusion
3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints
3.4.1 Introduction
3.4.2 Work Description
3.4.3 Materials
3.4.3.1 Adherends:
3.4.3.2 Adhesives (Table )
3.4.4 Experimental Work
3.4.5 Report
3.4.5.1 Introduction
3.4.5.2 Characterization of the Tested Joints
3.4.5.3 Theoretical Prediction of Failure Load
3.4.5.4 Comparison with Experimental Results
3.4.5.5 Conclusions
3.5 Modeling a Single‐Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling
3.5.1 Introduction
3.5.2 Work Description
3.5.3 Materials
3.5.3.1 Adherends:
3.5.3.2 Adhesives (Table )
3.5.4 Modeling Procedure
3.5.5 Report
3.5.5.1 Introduction
3.5.5.2 Module/Part
3.5.5.3 Module/Property
3.5.5.4 Module/Section
3.5.5.5 Module/Step (First Phase)
3.5.5.6 Module/Load
3.5.5.7 Module/Mesh
3.5.5.8 Module/Step (Second Phase)
3.5.5.9 Module/Job
3.5.5.10 Module/Visualization
3.6 Case Study in Joint Design for a Structural Automotive Application
3.6.1 Introduction
3.6.2 Report
3.6.2.1 Introduction
3.6.2.2 Design Brief
3.6.2.3 Adhesive Selection
3.6.2.4 Surface Treatment Selection
3.6.2.5 Material Properties
3.6.2.6 Joint Design
3.6.2.7 Numerical Models
3.6.2.8 Design Validation
3.6.2.9 Design for Manufacturing
3.6.2.10 Quality Control Techniques
3.6.2.11 Health and Safety Concerns
References
Chapter 4 Essay and Multi‐choice Questions
4.1 Essay Questions
4.2 Multi‐choice Questions
Solutions
Essay Questions – Example Answers
Multi-choice Questions – Solutions
Index
EULA

Citation preview

Adhesive Bonding Technology and Testing

Adhesive Bonding Technology and Testing Ricardo João Camilo Carbas Eduardo André Sousa Marques Alireza Akhavan-Safar Ana Sofia Queirós Ferreira Barbosa Lucas Filipe Martins da Silva

Authors Dr. Ricardo João Camilo Carbas

INEGI - Inst. of Science and Innovation in Mechanical Engineering Dept. of Mechanical Engineering Rua Dr. Roberto Frias s/n 4200-465 Porto Portugal

All books published by WILEY-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for

Dr. Eduardo André Sousa Marques

INEGI - Inst. of Science and Innovation in Mechanical Engineering Rua Dr. Roberto Frias 400 4200-465 Porto Portugal

British Library Cataloguing-in-Publication Data

Dr. Alireza Akhavan-Safar

The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at .

INEGI - Inst. of Science and Innovation in Mechanical Engineering Rua Dr. Roberto Frias 400 4200-465 Porto Portugal Dr. Ana Sofia Queirós Ferreira Barbosa

INEGI - Inst. of Science and Innovation in Mechanical Engineering Rua Dr. Roberto Frias 400 4200-465 Porto Portugal Prof. Lucas Filipe Martins da Silva

University of Porto Dept. of Mechanical Engineering Rua Dr. Roberto Frias s/n 4200-465 Porto Portugal Cover Image: © Jenson/Shutterstock

A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek

© 2023 WILEY-VCH GmbH, Boschstraße 12, 69469 Weinheim, Germany All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Print ISBN: 978-3-527-35051-3 ePDF ISBN: 978-3-527-83800-4 ePub ISBN: 978-3-527-83801-1 oBook ISBN: 978-3-527-83802-8 Typesetting

Straive, Chennai, India

v

Contents Preface xi 1 1.1 1.1.1 1.1.2 1.1.3 1.1.4 1.1.5 1.1.5.1 1.1.5.2 1.2 1.2.1 1.2.2 1.2.3 1.2.4 1.2.5 1.2.5.1 1.2.5.2 1.3 1.3.1 1.3.2 1.3.3 1.3.4 1.3.5 1.4 1.4.1 1.4.2 1.4.3 1.4.4 1.4.5 1.5 1.5.1

Simple Practical Demonstrations 1 Importance of Loading Mode on Bonded Joint Performance 1 Introduction 1 Equipment 1 Materials 2 Safety Precautions 2 Experimental Procedure 2 In Class 2 In the Laboratory 3 Surface Treatments and Methods to Evaluate Surface Energy 6 Introduction 6 Equipment 7 Materials 8 Safety Precautions 8 Experimental Procedure 8 In Class 8 In Laboratory 9 Stress Distribution Along the Overlap Length 12 Introduction 12 Equipment 13 Materials 13 Safety Precautions 13 Test Procedure 13 Visual Identification of Defects in Adhesive Joints 15 Introduction 15 Equipment 17 Materials 17 Safety Precautions 17 Test Procedure 17 Failure Analysis of Adhesive Joints 19 Introduction 19

vi

Contents

1.5.2 1.5.3 1.5.4 1.5.5

Equipment 20 Materials 20 Safety Precautions 20 Test Procedure 21

2 2.1 2.1.1 2.1.2 2.1.3 2.1.4 2.1.5 2.1.6 2.1.7 2.2 2.2.1 2.2.2 2.2.3 2.2.4 2.2.5 2.2.6 2.2.7 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.3.5 2.3.6 2.3.7 2.3.8 2.4 2.4.1 2.4.2 2.4.3 2.4.4 2.4.5

Production and Testing 25 Bulk Specimens 25 Introduction 25 Adhesive Pouring Technique 25 Metallic Mold 26 Adhesive Application 30 Curing Procedure 31 Machining Procedure 33 Testing Procedure 35 Thick Adherend Shear Specimens 35 Introduction 35 Metallic Mold 36 Surface Treatment of Adherends 37 Geometrical Control Using Shims 40 Specimen Manufacture 41 Final Specimen Preparation 42 Testing Procedure 42 Fracture Mechanics Specimens 44 Introduction 44 Metallic Mold 48 Surface Treatment of Adherends 49 Adhesive Spacers 50 Specimen Manufacture 50 Final Preparation of Specimens 51 Testing Procedure 52 Data Reduction Schemes 53 Single-Lap Joint Specimens 56 Introduction 56 Surface Treatment of Adherends 56 Joint Manufacture 57 Final Preparation of Specimens 60 Testing Procedure 60

3 3.1

Laboratorial Activities and Report Examples 65 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints 65 Introduction 65 Joint Strength Prediction 66 Work Description 67 Materials 67

3.1.1 3.1.1.1 3.1.2 3.1.3

Contents

3.1.3.1 3.1.3.2 3.1.4 3.1.5 3.1.5.1 3.1.5.2 3.1.5.3 3.1.5.4 3.1.5.5 3.1.5.6 3.2 3.2.1 3.2.2 3.2.3 3.2.3.1 3.2.3.2 3.2.4 3.2.5 3.2.5.1 3.2.5.2 3.2.5.3 3.2.5.4 3.2.5.5 3.2.5.6 3.3 3.3.1 3.3.2 3.3.3 3.3.3.1 3.3.3.2 3.3.4 3.3.5 3.3.5.1 3.3.5.2 3.3.5.3 3.3.5.4 3.3.5.5 3.3.5.6 3.4 3.4.1 3.4.2 3.4.3

Adherends (Tables 3.1 and 3.2) 67 Adhesive (Table 3.3) 67 Experimental Work 67 Report 68 Introduction 68 Experimental Procedure 71 Materials 71 Failure Load Prediction 72 Experimental Results and Discussion 72 Conclusions 78 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints 79 Introduction 79 Work Description 83 Materials 83 Adherends 83 Adhesives (Table 3.7) 84 Experimental Work 84 Report 84 Introduction 84 Materials 86 Prediction of the Failure Loads 86 Experimental Results 90 Discussion 94 Conclusions 97 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints 97 Introduction 97 Work Description 98 Materials 99 Adherends: 99 Adhesives: 99 Experimental Work 99 Report 99 Introduction 100 Experimental Details 100 Prediction 101 Experimental Results 102 Failure Surfaces 103 Conclusion 103 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints 104 Introduction 104 Work Description 105 Materials 105

vii

viii

Contents

3.4.3.1 3.4.3.2 3.4.4 3.4.5 3.4.5.1 3.4.5.2 3.4.5.3 3.4.5.4 3.4.5.5 3.5 3.5.1 3.5.2 3.5.3 3.5.3.1 3.5.3.2 3.5.4 3.5.5 3.5.5.1 3.5.5.2 3.5.5.3 3.5.5.4 3.5.5.5 3.5.5.6 3.5.5.7 3.5.5.8 3.5.5.9 3.5.5.10 3.6 3.6.1 3.6.2 3.6.2.1 3.6.2.2 3.6.2.3 3.6.2.4 3.6.2.5 3.6.2.6 3.6.2.7 3.6.2.8 3.6.2.9 3.6.2.10 3.6.2.11

Adherends: 105 Adhesives (Table 3.17) 105 Experimental Work 105 Report 106 Introduction 106 Characterization of the Tested Joints 107 Theoretical Prediction of Failure Load 112 Comparison with Experimental Results 115 Conclusions 116 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling 117 Introduction 117 Work Description 117 Materials 117 Adherends: 117 Adhesives (Table 3.25) 118 Modeling Procedure 118 Report 118 Introduction 118 Module/Part 122 Module/Property 124 Module/Section 128 Module/Step (First Phase) 129 Module/Load 131 Module/Mesh 132 Module/Step (Second Phase) 134 Module/Job 136 Module/Visualization 137 Case Study in Joint Design for a Structural Automotive Application Introduction 139 Report 139 Introduction 140 Design Brief 140 Adhesive Selection 142 Surface Treatment Selection 145 Material Properties 146 Joint Design 148 Numerical Models 152 Design Validation 156 Design for Manufacturing 160 Quality Control Techniques 162 Health and Safety Concerns 163 References 163

139

Contents

4 4.1 4.2

Essay and Multi-choice Questions 165 Essay Questions 165 Multi-choice Questions 166 Solutions 187 Essay Questions – Example Answers 187 Multi-choice Questions – Solutions 190 Index 191

ix

xi

Preface This book is intended to serve as a didactic tool to support both those teaching and learning the subject of adhesive bonding. While the companion book “Introduction to Adhesive Bonding” is mostly dedicated to the theoretical aspects of this joining technology, this book is more concise and highly focused on hands-on learning, with exercises and their solutions and multiple experimental activities. The book is divided into four parts. The first is dedicated to simple practical demonstrations of adhesive bonding. These are all simple activities suitable to be carried out in a classroom setting, which quickly highlight the advantages and limitations of this technique. The second part is dedicated to production and testing of specimens that are used to characterize adhesives and the most commonly used types of joints. The third part describes in detail multiple laboratorial activities suitable for implementation in the laboratorial classes of engineering courses. These activities explore aspects such as the manufacture of defect-free bonded joints, the effects of geometry and materials properties in adhesive joint testing, surface preparation and joint design and strength prediction, among many others. Lastly, a set of exercises is provided in the form of developmental questions and multiple choice questions. This last part focuses on all of the knowledge areas discussed in the companion “Introduction to Adhesive Bonding” book. All problems are provided with solutions, and many are fully solved, helping bachelor or masters students in their study and providing evaluation reference materials for teachers. The authors would like to thank Paulo Nunes for the help in the preparation of figures. They also want to thank the team of WILEY, especially Felix Bloeck, for the excellent cooperation during the preparation of this book. Porto, Portugal, 2022

Ricardo João Camilo Carbas Eduardo André Sousa Marques Alireza Akhavan-Safar Ana Sofia Queirós Ferreira Barbosa Lucas Filipe Martins da Silva

1

1 Simple Practical Demonstrations 1.1 Importance of Loading Mode on Bonded Joint Performance 1.1.1

Introduction

Adhesive bonding shows many advantages over more traditional methods of joining such as bolting, brazing, and welding or even the use of mechanical fasteners. No other joining technique is so versatile, and its transversality lies in its capacity to join different materials, its ability to ensure permanent assembly, and its ease of use. In fact, a well-designed bonded joint allows for a reduction in production costs, while maintaining proper mechanical properties of the joint. Adhesives work by exploring the adhesion phenomena, and they are usually polymeric materials, typically thermosetting, that, compared to materials that are joined in structural applications (such as metals and composites), show a much lower strength. Nonetheless, adhesive joints can be applied to a wide diversity of structures, withstanding different types of loads. To understand the mechanics of a bonded joint, it is important to first establish that the behavior of the joint is highly dependent on the type of loads it is sustaining. In an attempt to obtain the highest joint strength, it is fundamental to load the adhesive under forces acting in the plane of the adhesive layer, minimizing peeling loads. Joints are generally more resistant when shear-stressed because the adhesive layer is relatively well aligned with the loading direction. In these conditions, the entirety of the adhesive layer can positively contribute to sustain the load (see Figure 1.1). Joints subjected to cleavage or peel stresses are much weaker than those subjected to shear because the stresses are concentrated in a very small area. All the stress is located at the edge of the joint (see Figure 1.1).

1.1.2 ● ●

Equipment

One set of scissors Tensile testing machine

Adhesive Bonding Technology and Testing, First Edition. Ricardo João Camilo Carbas, Eduardo André Sousa Marques, Alireza Akhavan-Safar, Ana Sofia Queirós Ferreira Barbosa, and Lucas Filipe Martins da Silva. © 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.

2

1 Simple Practical Demonstrations

Cleavage

Shear

F

F F

F

Figure 1.1 joints.

1.1.3 ● ●

Schematic representation of the shear and cleavage loads acting on adhesive

Materials

One roll of double-sided foam adhesive tape Small aluminum beams

1.1.4

Safety Precautions

Apply the necessary safety procedures for operating a test machine.

1.1.5

Experimental Procedure

1.1.5.1 In Class

Peel the adhesive tape off the roll by applying a pulling force or “peeling” action as shown in Figure 1.2. See how easily it peels away, even if the adhesive is quite strong. Figure 1.2 Adhesive joint under pull-out force.

F

Adhesive tape roll

1.1 Importance of Loading Mode on Bonded Joint Performance

Adhesive

Adhesive

F F

F

F

Figure 1.3 Adhesive joint subjected to shear stress, with the area being overlapped with and without the adhesive.

Now, cut two strips of adhesive tape, approximately 10 cm long. Bond the two strips parallel to each other with an overlap of approximately 3 cm. Bond the glued side of one strip to the unglued side of the other strip (see Figure 1.3). Pull on the joint in order to try to separate the strips by loading them parallelly to the adhesive layer, thereby subjecting the adhesive to shear, as schematically represented in Figure 1.3. It will be much harder to separate the joint as we are now loading it in shear; however, because of the low stiffness of the tape, it will bend and introduce some peeling loads, as shown in Figure 1.3, and this peeling can promote debonding. Repeat the same procedure, but this time, join the strips so that the sides that have adhesive are in direct contact, as represented in Figure 1.4. When the joint is made between the glued side of both strips, it is impossible to separate the strips under shear. Ultimately, the strips will break, while the bonded area remains intact. 1.1.5.2 In the Laboratory

In order to better understand the influence of load type when an adhesive joint is used, the same tape will be bonded to an aluminum plate, and the response for two different types of load (shear and peel) will be studied using a universal tensile machine.

3

4

1 Simple Practical Demonstrations

Adhesive Adhesive

F F

F

F

Figure 1.4 Adhesive joint subjected to shear stress, with the area being overlapped with the adhesive on both strips.

To this aim, the same tape will now be applied to metal (aluminum) adherends. Cut an adhesive tape strip, approximately 10 cm long, and join it to the surface of an aluminum adherend with a 3 cm overlap. This adhesive joint is subjected to shear stress, as shown in Figure 1.5. As the adherend is much stiffer, this adhesive joint is now subjected to an almost uniform shear stress. The procedure will now be replicated, but this time, the forces exerted will be in a peeling direction. Therefore, it is recommended for the tape strip to be slightly longer so that it can be easily pulled off. Cut an adhesive tape strip, approximately 15 cm long, and join it to the surface of an aluminum adherend with 3 cm of overlap. This adhesive joint can now be subjected to peeling stress, as shown in Figure 1.6. A comparison of the loads applied on the manufactured joints can be done manually or using a testing machine. Manually, it is possible to “feel” that the forces are different, but they cannot be quantified. Therefore, using a universal testing machine, the behavior of the joints loading under different types of stresses and different surface states can be easily quantified, leading to different results. Figure 1.7 shows a schematic representation of the peel and shear forces. As “felt” in a manual test, when an adhesive joint is tested in peel stress, at first, it is necessary to exert a greater force to peel off the adhesive, but over the course of the test, the force required decreases and the joint eventually fails. In turn, when the adhesive joint is being tested at shear, the force required gradually increases until failure occurs.

1.1 Importance of Loading Mode on Bonded Joint Performance

Aluminum Adhesive

F

Figure 1.5 Adhesive joint using adhesive tape and aluminum adherends, subjected to shear stresses.

Aluminum Adhesive

F

Figure 1.6 Experimental testing procedure of an adhesive joint using adhesive tape and aluminum adherends, subjected to shear stresses.

5

1 Simple Practical Demonstrations

Load

6

Figure 1.7 Schematic representation of the shear and peel behavior of an adhesive joint.

Shear load

Peel load

Displacement

1.2 Surface Treatments and Methods to Evaluate Surface Energy 1.2.1

Introduction

Surface preparation of an adherend is key to achieve a strong and durable adhesive joint, and it is a process step that should never be taken lightly. The type and the quality of surface preparation will unequivocally determine the behavior of the joint. Surface treatments can be divided into two major groups: passive and active treatments. Briefly, we can explain this categorization by saying that passive treatments do not change the chemical nature of the material surface and the active processes chemically change the adherend by cleaning and removing weak layers on the surface. How a liquid will wet a surface will mainly dictate the level of adhesion between the adhesive and the adherend. The formation of a drop of liquid on a solid surface is described by the contact angle, 𝜃, between the solid surface and the tangent to the surface of the liquid at the point of contact as schematically presented in Figure 1.8. The aim of surface treatments is to obtain a clean and wettable surface. Unfortunately, there is no standardized procedure or equipment to assess surface cleanliness. Furthermore, a clean surface is difficult to define and sometimes even quantify. One way of evaluating the level of cleanliness is to say that a surface is clean when no dirt is visible to the naked eye. However, this is a very subjective process, and the quality of the surface treatment should always be subject to a strict control. The value of 𝜃 can vary from zero – when there is complete liquid spreading, and we are experiencing perfect wetting – to 180∘ when the liquid assumes the shape of a spherical drop and does not wet the solid at all, as shown in Figure 1.9. Vapor θ Liquid Solid

Figure 1.8 Angle of contact (𝜃) formed between an adherend surface and a liquid.

1.2 Surface Treatments and Methods to Evaluate Surface Energy

θ

0°

90°

180°

cosθ

1

0

–1

Spread

Full wetting

Partial wetting

γSL = γSV

Negligible wetness

Nonwetting

Figure 1.9 Variation of the contact angle of a drop of liquid as a function of its spreading on a surface.

Water forms spherical drops

Water forms a uniform film

Figure 1.10 Wetting of a surface with a liquid before and after surface treatment: (a) untreated surface and (b) treated surface.

These differences can be easily observed in Figure 1.10, where the same liquid and the same surface (in this case, a composite material) behave in different ways. Before the surface treatment, the liquid does not wet the surface, forming very visible and spherical drops. On the other hand, after surface treatment, it is observed that the liquid wets the surface, forming a film. The contact angle decreases with the application of a surface treatment. In an ideal surface preparation, the contact angle should be as close as possible to zero to ensure good adhesion. There is a wide range of surface treatments available for use with adhesive bonding processes, and consequently, the quality of the surface may vary and lead to different morphologies and surface conditions. However, the final result of a surface treatment should always be the same: an increase in joint strength and durability, achieved by promoting adhesion between the materials to be bonded and the adhesive. Regardless of the procedure used, this should always be the final goal of the surface treatment.

1.2.2 ● ● ●

Equipment

Plasma generator device Pipette Dyne pens

7

8

1 Simple Practical Demonstrations

1.2.3 ● ● ● ● ● ● ●

Materials

Aluminum adherend (non-degreased surface) Aluminum adherend (surface degreased with acetone) Polymeric adherend (surface degreased with acetone) Polymeric adherend (treated with plasma) Water Acetone Cleaning paper (take care, as chosen paper must not contaminate the surface, not shred after the addition of the solvent, nor leave residues on the treated surface)

1.2.4

Safety Precautions

Avoid direct contact of acetone with the skin. The plasma and anodizing treatments should be performed under an effective air extraction system because harmful volatiles can be released during the process treatment. Surface blasting processes must be carried out in accordance with the manufacturer’s safety recommendations and PPE must be provided.

1.2.5

Experimental Procedure

Surface treatment can be carried out using both passive and active methods. Depending on the class of material to be treated, a selection of the best suited methodology must be performed. 1.2.5.1 In Class

In this demonstration, a passive surface treatment method will be first used (cleaning with a solvent, acetone). This process aims to remove oily or greasy areas, which are the sources of very low wetting and adhesion. In many non-structural bonding applications, these processes are often sufficient, but they are frequently the first step of a more complex surface treatment process in structural applications. Degreasing is the simplest method suitable to obtain a clean surface, decrease the contact angle, and increase the adhesive spreading. This procedure can be applied to a wide range of materials, such as polymers, composites, and metals. There are several methods that can be followed for the application of a solvent. In this demonstration, manual cleaning was chosen, as shown in Figure 1.11. Cleaning should always be carried out in the same direction in order to remove the dirt without re-contaminating the surface. An active treatment procedure, as explained, will chemically alter the treated surface. It must therefore be carefully selected, taking into account the material to be treated. There are several approaches suitable for measuring the wettability, contact angle, and consequently the surface energy of a surface. Some are simple techniques, such as observing the shape of a drop of water, while others are much more complex, such as measuring the contact angle with specialized goniometers. The use of Dyne

1.2 Surface Treatments and Methods to Evaluate Surface Energy

Acetone

Cleaning direction

Cloth

Figure 1.11

Manual cleaning of a surface using acetone and a cloth.

pens is a widely used technique for assessing the quality of surface energy. This technique is a simple, cheap, and quick method, where pens are used to draw a line of a special ink along the surface of the adherent and thus visually observe the behavior of the liquid. To observe the shape of the drop on the treated surface, it is recommended to use a liquid whose properties are well known, and for this, distilled water is recommended. Using a pipette, a small drop is placed on the different prepared surfaces. The analysis of the drop shape will be mainly visual. If the drop has a spherical shape, this means that the liquid is not properly wetting the surface. If, on the other hand, the liquid easily spreads on the surface, this means that the treatment is facilitating the wetting of the liquid on the treated surface. It is easy to see that different surface preparations lead to different droplet shapes (Figure 1.12). Dyne pens can also be used to quantify the quality of the surface preparation. This analysis is more rigorous than the previously described one as the use of these pens allows us to determine an approximate value of the surface energy. Dyne pens use calibrated liquids, so if the applied liquid completely wets the surface, we can then have an approximate idea of a minimum surface energy value. This test is also very effective in predicting whether the surface shows differences before and after a surface treatment. Figure 1.13 shows the application of the same calibrated liquid before and after cleaning with acetone. Before cleaning, the liquid is unable to wet the surface. After cleaning with acetone, which is a simple and non-invasive surface preparation, the calibrated liquid can already wet the surface. 1.2.5.2 In Laboratory

A plasma generator will be used to treat the surfaces of polymeric adherends. Plasma treatment is the most effective technique to increase the surface energy of polymers because it is responsible for changing the chemistry of polymeric surfaces to be treated (see Figure 1.14).

9

10

1 Simple Practical Demonstrations

Without surface preparation

(a)

With surface preparation

(b)

Figure 1.12 Observation of the shape of a drop of water with a liquid of known properties on metallic surface without (a) and with (b) surface treatment.

Without surface preparation

(a)

Cleaned with acetone

Figure 1.13 Use of Dyne pens on a metallic surface before (a) and after (b) surface preparation.

(b)

Please note that these surfaces should not be touched with hands, avoiding the introduction of grease that can contaminate the already treated surfaces. It is advisable to join the treated surfaces immediately after treatment. If this is not possible, they should be conveniently stored in a manner that ensures that there is no contamination.

1.2 Surface Treatments and Methods to Evaluate Surface Energy

Figure 1.14 Surface of the polymeric material to be treated with plasma.

(a)

(b)

Figure 1.15 Observation of the shape of a drop of water with a liquid of known properties on a polymeric surface without (a) and with (b) surface treatment.

Figure 1.15 clearly shows that the wettability is higher when the polymer is exposed to plasma treatment, and this allows us to conclude that the adhesion is higher when the polymeric adherends are treated with plasma. The use of Dyne pens allows us to determine the surface energy of polymeric materials; this is achieved using different pens with different calibrated liquid energies, starting with 30 up to 38 mJ/m2 . Figure 1.16 shows that the surface energy of untreated polymeric adherends is between 30 and 32 mJ/m2 because it is clear that the liquids of this pens are able to uniformly wet the polymer. Figure 1.16 shows that the surface energy of the polymeric adherend after treatment with plasma increases from 30 mJ/m2 to more than 38 mJ/m2 . This means that the surfaces treated will show good adhesion with the adhesive, and after being bonded, the failure is probably cohesive.

11

12

1 Simple Practical Demonstrations

(a)

(b)

Figure 1.16 adherend.

Dyne pen application on a polymeric non-treated (a) and plasma-treated (b)

1.3 Stress Distribution Along the Overlap Length 1.3.1

Introduction

Tensile loads on a single overlap joint subject the adhesive to shear and peel stresses, as shown in Figure 1.17. The shear and peel distributions in the adhesive along the overlap length exhibit a large stress gradient at the end of the overlap, where a stress concentration is present. When considering the loads that an adhesive joint can be subjected to, shear is by far the more preferable for adhesive joints (see Figure 1.18). In this loading condition, the adhesive layer is relatively well aligned with the load direction, which means that the entire adhesive layer can contribute positively to support the load. When designing an adhesive joint, one should always try to ensure that the adherends carry the load in a manner that is as parallel to the adhesive layer as possible. Several analytical models allow the calculation of these stress distributions, as do many numerical methods such as the use of finite element modeling, which allows us to obtain precise stress and strain distributions along the adhesive layer, providing clear information about how loaded a portion of adhesive is under service. With this demonstration, it is intended to experimentally show how the strain F

F F Shear

Figure 1.17 stresses.

Peel

Schematic representation of adhesive joints subjected to shear and peel

1.3 Stress Distribution Along the Overlap Length

Figure 1.18 Typical bonded joint geometries for a joint under shear stress.

F b

F

t l

τav

τ (shear stress on the adhesive)

(and hence stress) distributions vary along the adhesive layer using a joint where the adherends are composed of hard rubber and the adhesive is simulated by a relatively soft foam rubber.

1.3.2 ● ●

One tensile testing machine Black marker pen

1.3.3 ● ● ● ●

Equipment

Materials

Two sheets of a hard natural rubber 25 cm × 2.5 cm × 1 cm One piece of foam rubber 2.5 cm × 2.5 cm Contact adhesive Acetone

1.3.4

Safety Precautions

Avoid contact with the contact adhesive and acetone.

1.3.5

Test Procedure

Clean the surfaces of the hard rubber and the foam rubber with acetone. Join the foam rubber to the hard rubber sheets with the contact adhesive to form a single overlap joint with an overlap length of 2.5 cm. Using the marker pen, trace vertical lines along the overlap length on the joint as indicated in Figure 1.19. Load the joint by pulling both adherends on opposite directions and observe the deformation of the joint in the overlap length region. This demonstration can be performed in a classroom using a manually applied load. The same specimen can also be used in a tensile machine, which will allow for the stress concentrations at the ends of the overlap length to become more visible.

13

14

1 Simple Practical Demonstrations

Figure 1.19

Schematic representation of the vertical trace along the length of the joint.

Figure 1.20

Single lap joint in an unloaded state.

Figure 1.21

Single lap joint in a loaded state.

Observing the vertical lines made on the adhesive layer (see Figure 1.20) is a simple method to determine the level of stress present along the bondline in an adhesive joint. Figure 1.21 shows a loaded joint, and it is clear that the level of shear stress is higher at the ends of the overlap length. Figure 1.22 schematically shows the areas of the adhesive joint where the concentration of stresses is greater, i.e. the edges of the joints.

Stress concentration

Figure 1.22 Schematic representation of stress concentration in an adhesive joint when subjected to shear stresses.

1.4 Visual Identification of Defects in Adhesive Joints

1.4 Visual Identification of Defects in Adhesive Joints 1.4.1

Introduction

Because of poor storage conditions, manufacturing problems, large internal stresses, or unexpected service loads, defects may appear in an adhesive layer or within interfacial areas. Defects should be detected whenever possible as they can significantly impact the joint strength, leading to premature failure and compromising structural integrity. Defect type, size, and location are three important factors affecting the joint strength. Porosity, cracks, voids, detachments, presence of a foreign object, poor curing, and poor adhesion are some of the defects that can be observed in a poorly manufactured joint (see Figure 1.23). These defects can be grouped into three major groups: ●

●

●

Poor adhesion (poor bonding between the adhesive and the adherend), which results from poor surface preparation or by the presence of contaminating substances in the adherend; Poor cohesive strength, resulting from incorrect adhesive formulation, poor mixing, or insufficient adhesive curing; Voids and porosities that result from the presence of air bubbles, volatile release, inadequate curing, thermal contraction, or application of the adhesive. This type of defect is most easily detectable by non-destructive techniques. Voids are usually created because of the presence of trapped gas/air bubbles in the adhesive mixture, even before the adhesive is applied to the surface. Voids in the adhesive layer also come from the incorrect pattern of adhesive application on the bonding surface, which can cause air to trap inside the adhesive layer. Figure 1.24 and Figure 1.25 show some of the good and bad practices associated with the application of adhesives and the manufacture of bonded joints. Following these recommendations will minimize the probability of having defects in an adhesive joint.

Weak adhesion Void Crack

Porosity

Poor cure Unbound

Adherend

Adhesive

Adherend

The presence of voids leads to a decrease in joint strength, a decrease in the adhered area, and an increase in the stress level within the adhesive layer. The presence of voids leads to a decrease in joint resistance, a decrease in the adhered area, and an increase in the stress level within the adhesive layer. Hence, a proper

Figure 1.23

Types of defects that can be found in an adhesive joint.

15

16

1 Simple Practical Demonstrations

Good practice

Figure 1.24

Good and bad practice in adhesive application. Good practice

Figure 1.25

Bad practice

Bad practice

Good and bad practice in top adherend application.

quality control procedure will examine the presence of voids in adhesive layers. If air is trapped at the interface between the adhesive and the adherend, a disbonded region will be created. Discontinuities, voids, relative sliding between adherends, insufficient amounts of adhesive, fracture, and indentations or dents are examples of defects that can be identified by macroscopic observation of the bonding area. Visual inspection as a non-destructive control method is a simple task that only allows a first identification of the bond quality. One of the main methods used to perform the quality control of adhesive joints is visual control, not only at the time of manufacture but also after the execution of the joint. While the visual inspection method is a simple approach used to perform the quality control of adhesive joints, the inspection operator must be highly skilled and experienced. Moreover, it is crucial to provide adequate light intensity, ensure the correct viewing angle, and use the most suitable tools. The accuracy of this technique is highly dependent on the quality of the supporting installation. Visible defects and faults should be compared with reference images in inspection manuals to ensure that they are within an acceptable range. This type of control is appropriate for identifying defects or flaws that are noticeable on the surface of the joint. In addition, geometric faults such as misalignment, non-uniform adhesive thickness, incorrectly shaped fillets, etc., can also be visually observed. Lack of excess adhesive or filleting at the edges of the bondline after manufacture may be a sign of insufficient adhesion between the adherends or a poorly secured bondline (thicker bondline).

1.4 Visual Identification of Defects in Adhesive Joints

1.4.2 ● ●

Magnifying glass Metallic coin

1.4.3 ●

Equipment

Materials

Examples of adhesive joints that are misaligned, porous, with “burnt” adhesive, with non-uniform thickness, and with missing adhesive.

1.4.4

Safety Precautions

No hazards to report.

1.4.5

Test Procedure

In Figure 1.26, it is possible to observe the geometric misalignment of an adhesive joint. It is clearly visible that the bottom adherend is not in the correct position, while the correct position of the adherend is marked in red. These joint misalignments cause stresses to be different from those expected for a given adhesive joint, which, in addition to all the dimensional errors, can lead to serious constraints in joint performance. The correct curing process for an adhesive is always listed in the technical data sheet provided by the supplier. It is very important to follow these indications because the use of incorrect procedure (for example, too low temperature) might lead to insufficient cure, and the strength of the joint will be lower than as designed. On the other hand, if the curing temperature is higher, we can “burn” the adhesive (see Figure 1.27). In fact, both excessive temperatures and long curing times may

Misaligned substrate

Correct substrate alignment

Figure 1.26

Adhesive joint with geometric misalignment of one of the adherends.

17

18

1 Simple Practical Demonstrations

Burned adhesive

Figure 1.27 Adhesive joint showing areas of burnt adhesive due to excessive curing temperature. Lack of adhesive

Figure 1.28

Adhesive joint missing adhesive in parts of the overlap area.

lead to a degradation of the polymeric chains, leading to a loss of chemical and mechanical properties. One of the most common defects observed in joint production is the lack of adhesive at the edges (Figure 1.28). This defect can lead to premature failure of the adhesive joint, as only a small part of the joint is being used to attain the strength for which the joint was designed. This defect is mainly due to poor quality execution of the adhesive joint, usually by an insufficient application of adhesive or poor application of the adhesive along the length of the overlap. This defect may also be associated with an inhomogeneous variation of the adhesive layer thickness if there is no tight control of this dimension. The tap test, like the visual test, is among the simplest NDT approaches used in practice. In this test, the joint surfaces are tapped with a tool, as presented in Figure 1.29. In our example, we will use a coin, but a hammer or even your knuckles could be used instead. An operator will listen to the reflected sound wave to determine whether the joint is qualified or not. Large unbound areas between the joint and the adhesive or the presence of significant voids will visibly alter the reflected sound. Voids or defects generate a resonant sound, whereas when there is no defect, the sound is usually hollow. While effective in detecting the existence of many defects, this approach is unable to provide information about the size or

1.5 Failure Analysis of Adhesive Joints

Figure 1.29

Detection of voids with the aid of a coin by the analysis of the sound emitted.

type of defects. Furthermore, if the defect is far from the surface, it will be difficult to detect using the tap test. Once again, this test is closely related to the experience of the operator.

1.5 Failure Analysis of Adhesive Joints 1.5.1

Introduction

The ultimate purpose of an adhesive joint is load transfer between the two bonded components, maintaining its structural integrity under static and/or dynamic stresses and adverse environmental conditions (temperature and humidity). It therefore becomes fundamental to correctly assess the distribution of the stress profile and, consequently, the failure modes induced in the bonded joints. In the majority of failures in adhesive joints, it is possible to distinguish three different failure modes (Figure 1.30): ● ● ●

cohesive breakage inside the adhesive, adhesive breakage at the interface between the adherend and the adhesive, and breakage of one of the adherends.

These three failure modes are schematically presented in Figure 1.30. Adhesive failure occurs when there is poor surface preparation. This happens when there is a loss of adhesion between the adhesive and the adherend. The bond (chemical and mechanical) that should solidly connect the adhesive to the adherent is somehow lost and there is a clear separation between these two materials at the interface. When we have adhesive failure, it is possible to accurately determine the strength of the adhesive layer using a variety of models that are based on the mechanical properties of adhesive materials. The same is not true when we have adhesive failure, as the interface properties are extremely difficult to determine and may depend on several complex factors.

19

20

1 Simple Practical Demonstrations Types of failure

Cohesive failure (undesirable)

Joint before failure Adhesive failure Adhesive

(undesirable)

Substrate failure (desirable)

Figure 1.30 Schematic representation of the different failure modes in single overlap bonded joints.

On the other hand, a joint must be designed so that the weakest element is the adherend, i.e. failure never occurs through the adhesive. When this type of failure occurs, it is known that the adhesive joint is designed to be stronger than the material being joined. When failure occurs in the adherend, this does not necessarily mean that the adherent will break cleanly, and the adhesive layer will remain intact. Often, the adherent will instead yield and become permanently deformed (in the case of metals) or delaminate (in the case of composites), and this can then lead to the failure of the adhesive layer. Ultimately, a correctly designed and produced joint will be one in which adherend failure occurs first, even if the adhesive itself becomes damaged and fails as a result of this. When an adhesive joint is manufactured using composite laminates, the situation may be more complex, so it is advisable to use adherends that have surface layers with fibers oriented parallel to the direction of stress, seeking to avoid interlaminar failure of these layers.

1.5.2 ● ●

Magnifying glass Protective gloves

1.5.3 ● ●

●

Equipment

Materials

Fractured adhesive joints, showing adhesive, cohesive, and interfacial failures Adhesive joints with metallic and polymeric adherends (with and without deformation) Composite adhesive joints

1.5.4

Safety Precautions

Gloves should be worn to avoid skin injuries caused by the sharp fibers that are presented in the fracture composite materials.

1.5 Failure Analysis of Adhesive Joints

Figure 1.31

1.5.5

Adhesive joint with adhesive failure.

Test Procedure

This work consists in the careful observation of the fracture surfaces of different adhesive joints. Ideally, the observation can be done with the naked eye, but in more complex cases, more sophisticated equipment should be used, such as magnifying glass or a microscope (optical or electron). In Figure 1.31, an adhesive failure is observed, whereupon the adhesive does not bond to one of the adherends, and the failure occurs at the interface between the adhesive and the adherend. As already described, this type of failure is highly undesirable. This sort of failure means that the surface preparation is not ideal, leading to a premature failure of the adhesive joint. A cohesive failure in the adhesive is demonstrated in Figure 1.32. Failure occurs in the middle of the adhesive layer, demonstrating good adhesion of the adhesive to the adherends. Although better than adhesive failure, this type of failure is also undesirable in practical applications as the adhesive is still the weakest link within the joint. The adhesive joint shown in Figure 1.33 shows adherend failure. In other words, it can be concluded that the joint is well designed, and the surface preparation is well executed as the weakest element of the adhesive joint is now the adherend material and not the adhesive layer or its interface. In this specific case, it is easily seen that there is a great concentration of stresses at the edge of the adhesive joint overlap coupled with massive plastic yielding of the polymeric adherend. The same adhesive can be used to manufacture adhesive joints with metallic adherends using hard steel. In this case, failure occurs within the adhesive, and, unlike the joints bonding the polymeric material, there is no yielding of the metallic adherends. However, some voids can still be noticed, suggesting that the adhesive mixture or applications was not perfect. In this case, the surface preparation is also

21

22

1 Simple Practical Demonstrations

Figure 1.32

Adhesive joint with cohesive failure in the adhesive.

Figure 1.33

Adhesive joint with adherend failure.

appropriate because there is adhesion between the adhesive layer and the adherend (Figure 1.34). In Figure 1.35, showing a mixed failure, it is visible that the failure mechanism is half cohesive and half adhesive. Occurrence of adhesive failure means that the surface treatment can be improved in order to improve the adhesion, obtaining cohesive failure. In Figure 1.36, it is visible that adherend delamination occurs, as composite fibers separate from the matrix material and become exposed. This type of failure is very disadvantageous because it occurs in the areas where the peel stresses generated

1.5 Failure Analysis of Adhesive Joints

Voids

Figure 1.34

Cohesive failure in the adhesive layer, in metallic adhesive joints. Adhesive failure

Figure 1.35

Mixed failure (cohesive and adhesive failure).

23

24

1 Simple Practical Demonstrations

Delamination

Figure 1.36

Example of delamination failure of composite joints.

by the adhesive layer are at the highest and the poor transverse strength of the composite is overcome. In these cases, it is necessary to redesign the joint in order to ensure that there is no delamination. This can be done, for example, by reinforcing the overlapping area with fillets or chamfers.

25

2 Production and Testing 2.1 Bulk Specimens 2.1.1

Introduction

Multiple test methods are available to determine the strength of an adhesive. These methods can be divided into two different techniques, which are tests of neat resin/bulk specimens or tests carried out in a joint. The bulk specimens most commonly used for adhesive testing are shaped as a “dog bone,” following diverse standards for testing of plastic materials. The use of joints to characterize the adhesive adds additional challenges to this process, such as the proper selection of surface treatment of adherends (in order to ensure that the adhesive is properly characterized) and to ensure that the stress field in the adhesive layer is as uniform as possible. Because of this fact, the use of bulk specimens is generally preferred. Bulk specimens can be manufactured by pouring or injecting adhesives. The injection technique consists in the injection of the adhesive directly in a mold with the final shape (Figure 2.1). It allows us to avoid problems associated with machining and cutting of the specimens, something that is especially hard when very ductile adhesives (e.g. silicone adhesives) are used. The metallic mold is composed of top and bottom parts and a middle frame with the final shape of the specimens. The three parts are fitted together with screws, and it is the torque applied to these screws that will control the pressure applied in the adhesive layer (pressure limited after the three parts come in contact with each other). The main problem associated with this technique is associated with the shrinkage of the adhesive during the curing process, leading to important defects in the surface (e.g. voids). The adhesive pouring technique consists of two main steps: the manufacture of a bulk plate and the subsequent machining of the “dog bone” shape (Figure 2.2). This technique is suitable for any type of adhesive and in any form (liquid, paste, or film), allowing the production of adhesive plates without defects.

2.1.2

Adhesive Pouring Technique

The production of adhesive plates follows the French standard NFT76-142 that describes the manufacture of plates in a mold with a silicon rubber frame under Adhesive Bonding Technology and Testing, First Edition. Ricardo João Camilo Carbas, Eduardo André Sousa Marques, Alireza Akhavan-Safar, Ana Sofia Queirós Ferreira Barbosa, and Lucas Filipe Martins da Silva. © 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.

26

2 Production and Testing

Figure 2.1

Tensile specimens produced with injecting techniques.

Figure 2.2

Tensile specimens produced with adhesive pouring techniques.

high pressure (2MPa). The different elements necessary to obtain adhesive plates with high quality are shown in Figure 2.3. A special metallic mold with an internal silicone frame is used, and the adhesive should fill this silicone frame with an excess of 5% in volume. This excess amount will flow out from the mold cavity and draw with it voids and bubbles. When pressure is applied to the mold, the internal silicone frame will expand sideways and generate a hydrostatic pressure in the plate, further ensuring that voids are removed and ensuring a consistently compressed material. Moreover, it is this silicone frame that controls the final thickness of the plate.

2.1.3

Metallic Mold

The French standard (NF T 76-142) suggests a generic dimension for the metallic mold used to obtain good-quality adhesive plates. Figure 2.4 schematically shows a metallic mold suitable to obtain these plates. The mold is composed of a base and a

2.1 Bulk Specimens

Figure 2.3

Schematic representation of pouring technique. Mould Base and lid 235 195

7

90 130

7

Metalic frame 1

14

A–A

20

A

ø7 H14

20

40

20

37,5

10

A 235

10

7 14

1

B–B

B

7

10 7 13

18 8

B 20

20 25 90

Figure 2.4 Technical drawing of a metallic mold used in the pouring technique (dimensions in mm).

27

28

2 Production and Testing

Figure 2.5

Removal of dust and particles from the mold surface.

lid with the same dimensions, showing a working area of 195 × 90 mm2 . It is within this working area that the silicone frame should be inserted. In order to ensure that the silicone remains fixed during the curing process, the working area must be restricted by a metallic frame composed of four metal beams fastened together. The surfaces that come in contact with the adhesive must be prepared in advance. Contaminants on the metal surface must be removed to ensure good adhesion of the adhesive through manual cleaning with a sandpaper, followed by degreasing of the surface using acetone (Figure 2.5). Before closing the mold, the different parts of the mold should be covered by a thin layer of release agent to facilitate the remotion of the plate after curing (Figure 2.6). The release agent should be properly selected as a function of the adhesive being used as different adhesives react differently to the mold release agents available in the market. The different parts should be mounted, and the screws are properly tightened to ensure that all parts remain in position during the pressure application. Before the application of the adhesive, the silicone frame should only be applied to the mold after the release agent is fully cured (Figure 2.7). Please note that the silicone frame does not require coating of the release agent because most adhesives do not readily

2.1 Bulk Specimens

Figure 2.6

Application of mold release agent.

Figure 2.7

Assembly of different parts of the mold and positioning of the silicone frame.

29

30

2 Production and Testing

bond well to silicone. Moreover, no contact should be allowed between the adhesive and the uncured mold release agent as this will damage the adhesive. The silicone frame has two main purposes. The first is to control the final thickness of the plate, whereas the second is the application of a hydrostatic pressure to the adhesive during the curing process, ensuring the removal of any voids. Generally, thicknesses of 2 mm are often used, and the maximum internal dimensions of the cavity where the adhesive will be deposited are 140 × 65 × 2 mm3 , ensuring the manufacture of defect-free plates. The silicone rubber that is typically used is a room temperature vulcanizing (RTV) silicone with a hardness of about 50 Shore.

2.1.4

Adhesive Application

The adhesive is applied to the cavity formed by the silicone frame in a quantity that is 5% of excess of the volume of the cavity. For two-component adhesives supplied in separate containers, the mixture of the resin and hardener is a very important process because voids can be introduced (Figure 2.8a). To avoid these defects, the mixture should be performed using a high-speed centrifuge machine, an ultrasonic mixer, or with the assistance of a vacuum chamber (Figure 2.8b). Moreover, to weight each component, precision scales should be used because the incorrect ratio of hardener to resin can lead to vastly different mechanical properties after curing or even lead to a situation where the adhesive never fully cures. In some cases, two-part adhesives are supplied in cartridges (one and two components), and the mixture is carried out with the use of an appropriate nozzle (Figure 2.9). The mixed adhesive should be deposited within the silicone frame in a well-controlled manner, with closely clumped beads and then uniformly spread to avoid the introduction of voids.

(a)

(b)

Figure 2.8 Equipment used to ensure proper mixing of adhesives supplied in separated containers: (a) precision scales and (b) high-speed centrifuge. Source: IndiaMART InterMESH Ltd.

2.1 Bulk Specimens

(a)

(b)

Figure 2.9 Two different techniques to apply the adhesive: (a) with the use of a nozzle and (b) with the use of a spatula.

The closure of the mold also requires special attention to reduce the air trapped within the plate. This is achieved by closing the mold in a continuous rotational movement (Figure 2.10). A direct vertical translation of the mold cover should be avoided as it minimizes paths for the release of trapped air.

2.1.5

Curing Procedure

The closed mold is subjected to high pressure, and this can be achieved in a hot plate press, where both the pressure and the temperature can be precisely controlled to reduce the production times and ensure consistency (Figure 2.11). The heating and cooling ramps must also be properly controlled in order to ensure a uniform temperature within the mold, something that is fundamental to reduce the thermal gradients that will generate large residual thermal stresses.

31

32

2 Production and Testing

Figure 2.10

Mold closure process.

Figure 2.11

Application of pressure on the closed mold.

2.1 Bulk Specimens

(a)

(b)

(c)

(d)

Figure 2.12 After curing procedure, the mold is opened (a), unfasted (b), the elements of the mold removed (c), and the plate removed (d). 150 50

50

10

R108

Figure 2.13

Tensile test geometry (dimensions in mm).

After full curing of the adhesive, the metallic frame of the mold is unfastened, the lid is carefully removed, and the adhesive plate is then removed (Figure 2.12). This procedure is easiest if the release agent was properly applied.

2.1.6

Machining Procedure

The cured plate can then be machined according to the dimensions of the tensile specimens that are to be used, e.g. the BS 2782 standard (Figure 2.13). Machining of the cured plates can be easily performed in a computer numerical control (CNC) milling machine using a G-code with the final design of the specimens. The ISO 2818 standard gives precise details on how to machine these adhesive plates. It is important to note that the coolant typically used in the milling process cannot be used as it might diffuse into the adhesive, affecting the mechanical

33

34

2 Production and Testing

Figure 2.14 Tensile specimens after machined.

properties of the adhesive. The cutting parameters and the milling tools used should be carefully selected and adjusted for the type of adhesive that is being used in order to reduce the introduction of micro-cracks in the specimens and to ensure good quality of the specimens (Figure 2.14). However, when very flexible adhesives are used, the specimens cannot be machined and the alternative relies on the use of very sharp dies that use pressure to precisely cut the specimens to the desired shape.

2.2 Thick Adherend Shear Specimens

2.1.7

Testing Procedure

The specimens produced are typically tested using a universal testing machine to determine their strength and stiffness under tensile loads. The load cell being used should have a capacity suitable for testing the type of the material (typically 5 kN). The test speed can vary, but the most commonly used crosshead rate is 1 mm/min. The test specimens should be instrumented with a strain gauge to ensure correct measurement of the deformation during the test. These data can be obtained with the use of mechanical strain gauges or video extensometers (Figure 2.15). Before the start of the tests, it is very important to carefully measure the dimensions of the specimens, especially of the resistant central section. The stresses present in the specimen are determined through the load divided by the initial section of the specimens, and the strain is registered by the extensometer. This curve allows us to determine the tensile stiffness (Young’s modulus), yield strength, maximum strength, and the strain to failure.

2.2 Thick Adherend Shear Specimens 2.2.1

Introduction

The thick adherend shear test (TAST) is typically used to determine the shear strength of the adhesives. This test method allows us to determine the shear stress, shear modulus, and shear strain to failure of the adhesive. While some researchers use the single-lap joint (SLJ) to determine the shear properties of an adhesive, this

(a)

(b)

Figure 2.15 Mechanical tests of tensile tests using extensometer (a) and video extensometer (b).

35

2 Production and Testing

51

42.5

12

0.5 5.75 6.5

(a)

Ø 12 51

57.5

0.5

12

36

5.75 6.5 (b)

110

Figure 2.16 Two methods to manufacture TAST specimens: (a) machined adherends and (b) bonded adherends (dimensions in mm).

procedure is in fact not suitable to determine the true adhesive properties because of the bending moments that cause large peel stresses at the ends of overlap length, and which lead to non-uniform loading of the adhesive layer. The TAST procedure uses a loading mode similar to that of the SLJ, but the geometry of the specimen does not allow for rotation of the adherends, which means that the bondline is subjected to almost uniform shear loads. The geometry of the specimen is essential because the shorter the bond length and the greater the thickness of the adherends, the lower the peel stresses generated at the ends of the joint. The adherends are typically made of steel, ensuring maximum stiffness. The ISO 11003-2 standard proposes two methods to manufacture these specimens, the use of bonded or machined adherends, as shown in Figure 2.16. The bonded adherends consist of four plates bonded together and is more susceptible to misalignment and the formation of adhesive fillets in the center of the joint. Moreover, ensuring uniform adhesive thickness is difficult. In contrast, machined adherends have a much smaller bonded area, and the adhesive fillet is more easily controlled. The adherends obtained with machining show higher bending stiffness than bonded adherends, and this higher bending stiffness leads to a reduction in the peel stresses in the adhesive layer.

2.2.2

Metallic Mold

A metallic mold should be used to ensure the correct alignment of the specimens and allow us to manufacture multiple specimens simultaneously. In these molds, the alignment of the specimens is ensured with pins, while the lateral bars control the overlap length, as can be seen in Figure 2.17.

2.2 Thick Adherend Shear Specimens

(a)

(b)

Figure 2.17

Metallic mold used to manufacture TAST specimens.

The mold should be coated with a mold release agent to ensure that the specimens can be easily removed after curing. Furthermore, the metallic mold plates and the specimens whose surfaces are in contact with the adhesive must be adequately prepared in order to improve the level of interfacial bonding. Contaminants on the metal surface, such as dust, dirt, oil, or oxides, have to be removed to guarantee good adhesion of the release agent. Impurities on the metallic plates must be removed using sanding, followed by degreasing with acetone (Figure 2.18).

2.2.3

Surface Treatment of Adherends

The adherends used for the TAST specimens are typically manufactured from construction steel because of the high stiffness and strength of this material, both of which contribute to a stress state that allows us to characterize the adhesive under

37

38

2 Production and Testing

Figure 2.18

Application of the release agent.

shear loads. The preferred surface treatment for this steel is sandblasting, as it generates a uniform surface roughness, allows for complete removal of any existing oxide, and ensures good wetting (Figure 2.19). Subsequently, the adherends are degreased with acetone so that they are clean and free of any contaminants. In some cases, there may be the need to fabricate specimens out of aluminum, especially for testing procedures where the adhesive is to be subjected to aging. However, interfacial failure can occur because of water ingress through the interface in this situation. To avoid this type of failure, the surfaces of the adherends that will

2.2 Thick Adherend Shear Specimens

Figure 2.19

Sandblasting the area where adhesive will be applied.

Figure 2.20 Anodized adherend surface.

receive the adhesive must be subjected to an active surface treatment (anodization), followed by the application of a primer suitable for the type of adhesive that will be used. A surface treatment that ensures good resistance under aging conditions is the anodization of the adherends, typically following the ASTM D3933 standard. This surface preparation consists of the creation of a stable oxide coating, with many pores, as shown in Figure 2.20, allowing the primer to fully penetrate and interlock with the coating. The phosphoric acid anodization process starts with a surface grit blasting, followed by degreasing with acetone, removing the impurities resulting from blasting. An anodizing solution with 12% of concentration of phosphoric acid and 88% of distilled water should be used according to the aforementioned standard.

39

40

2 Production and Testing

DC power supply

–

+ Anodised part

Electrical wires

Cathodes

Electrolyte

(a)

(b)

Figure 2.21 Laboratorial setup used to anodize adherends (a) and surface to be anodized during the process (b).

The aluminum specimens are immersed in the solution and integrated into an electric circuit as schematically shown in Figure 2.21a, and the surface of the adherends during the anodizing process is shown in Figure 2.21b. This circuit produces a voltage of 16 V, with the terminals being positioned so that the parts are not in contact with the solution. The anodization process should be performed continuously for 20–25 min. Once this stage is completed, the specimens are immediately removed and rinsed with deionized water, fully removing the anodizing solution from the surfaces.

2.2.4

Geometrical Control Using Shims

For this type of specimens, the overlap length and the adhesive fillet are controlled using custom steel shims. The use of these shims allows us to precisely control the location of the adhesive and the shape of bonded area, ensuring that only shear loads are being carried by the adhesive (Figure 2.22). The shims will remain in contact with the adhesive during the curing process and must be removed after the manufacture

2.2 Thick Adherend Shear Specimens

Shims

Adherends

Figure 2.22 fillet.

Metallic mould

Adhesive

Schematic representation of the position of shims to control the adhesive

process is completed. Thus, before manufacturing, it is also necessary to coat the shims with a suitable mold release agent.

2.2.5

Specimen Manufacture

The adherends are placed in the mold, and respective shims are inserted between them (on both sides) to ensure the overlap length and the adhesive fillet (Figure 2.23). For the lower positions, the shims can be slid into place to facilitate manufacture. The adhesive is then applied in the joint, which can be done using a handgun and an application nozzle or manually, using a spatula (Figure 2.24). This part of the process is mostly dependent on the form of the adhesive and how it is made available. After the adhesive is applied, the mold is closed and placed inside a hot press in order to ensure constant pressure and uniform temperature throughout the mold (Figure 2.25). For this type of specimens, pressure control is not as important as is the case for the bulk adhesive plate as pressure is only used to maintain the mold closed and it is not fully transferred to the adhesive layer.

41

42

2 Production and Testing

Figure 2.23

2.2.6

Positioning of the shims and the adherends in the mold.

Final Specimen Preparation

Once the specimens are fully cured, they can be removed from the mold. The excess adhesive (spew fillet) should be removed using a file and coarse sandpaper, and the shims must then be carefully extracted in order to avoid any damage to the bonded area (Figure 2.26).

2.2.7

Testing Procedure

Experimental testing of TAST specimens requires a setup especially developed for use with universal testing machines, as illustrated in Figure 2.27. A machine with a load cell with 10 kN is often suitable to this type of tests. As this type of test evaluates the pure shear strength and stiffness of the adhesives, the relative displacement of the adhesives should be measured using a transducer located

2.2 Thick Adherend Shear Specimens

(a)

(b)

Figure 2.24

Application of the adhesive using a nozzle (a) or a spatula (b).

in the central zone of the samples. This transducer uses two strain gauges, fixed on the opposite faces of the specimen (Figure 2.28). The individual readings from the transducers ensure correct measurement of the strains acting in the bonded region, thus minimizing errors due to bending. To minimize the measurement of adherend deformation, the points of contact should be as close as possible to the bonded area. For the calculation of shear stress and strain in the adhesive layer, it is assumed that the adhesive acts essentially under pure shear and that the shear stress along the overlap length is uniform, although some concentration at the edges is inevitable. The accuracy of the strain measurement depends mainly on the thickness of the adhesive layer and the stiffness of the bond. Thinner adhesive layers have less displacement, so the errors associated with the measured displacement are relatively higher. Finally, to obtain the strain of the adhesive, it is necessary to impose a correction that takes into account the displacement of the adherends. The standard refers that this correction should be made considering the shear deformation in a solid steel

43

44

2 Production and Testing

(a)

(b)

Figure 2.25

Closing of the mold (a) and closing the mold (b).

specimen. However, some authors demonstrate that as the shear stress distribution in the specimen is not uniform, this method is not totally reliable. More accurate results are thus obtained using the elasticity equations and considering only pure shear.

2.3 Fracture Mechanics Specimens 2.3.1

Introduction

The double cantilever beam (DCB) and end notched flexure (ENF) tests are commonly used to characterize the fracture behavior of adhesives and other materials. This type of specimens was first developed with metallic adherends, but because of the growing interest in fiber-reinforced polymer matrix composites, it has been

2.3 Fracture Mechanics Specimens

(a)

(b)

Figure 2.26

Removing the shims (a) and the excess of adhesive from the joint (b).

expanded to composite adherends. The DCB and ENF tests are used to determine the critical strain energy release rate in mode I (GIC for the tensile opening mode) and mode II (GIIC for the shear mode), respectively. The DCB test is the test method most commonly used to determine GIC and is widely accepted by researchers. It is standardized in the ASTM D3433-99 standard for adhesive joints and ISO/DIS 15024 for composite fracture characterization. For GIIC determination, there is less agreement behind the ENF procedure. However, the Japanese (JIS 7086) and European (AECMA prEN 6034) standards for composites have been established and successfully adapted to adhesive joints. There are two ways to obtain DCB specimens, one way consists in the use of two plates with large dimensions that are bonded and subsequently cut into individual specimens (Figure 2.29a). The alternative methodology is to use adherends with

45

46

2 Production and Testing

Figure 2.27 Setup used to test the TAST specimens according to the ISO 11003-2 standard.

Figure 2.28 Assembly of transducers in TAST specimens.

2.3 Fracture Mechanics Specimens

(a)

(b)

Figure 2.29 two beams.

Two methods to obtain DCBs: (a) cutting out specimens and (b) bonding of

the final dimensions and where the specimens are manufactured individually (Figure 2.29b). The manufacture of individual specimens has the disadvantage of having longer production times. However, manufacturing specimens from two large, bonded plates requires them to be cut from single specimens that can lead to introduction of cracks in the adhesive. The DCB tests apply an opening load to both beams, and this requires the introduction of features where this load can be applied. If thick adherends are used, this is usually achieved by drilling holes (Figure 2.30a). However, when thin adherends are used, there is no space to drill these holes, so special blocks or piano hinges are

(a)

(b)

(c)

Figure 2.30 Different techniques to apply loads: (a) drilled holes, (b) bonded piano hinges, and (c) bonded blocks.

47

48

2 Production and Testing

Figure 2.31

Mold used to manufacture the specimens.

attached to the beams (Figure 2.30b,c). These solutions are typically employed with composite adherends.

2.3.2

Metallic Mold

Figure 2.31 shows a drawing and the metallic mold suitable for the manufacture of DCB and ENF specimens. The metallic mold is composed of two flat plates, drilled with holes for pins. These pins serve two different purposes. The first is the alignment of the joints and its adherends and the second is to ensure the correct alignment between both plates of the mold. This mold allows us to manufacture up to six individual specimens in a single step. The surfaces that contact with the adhesive joints must be well prepared in advance in order to ensure that the mold is clean without any burrs and adhesive residues (Figure 2.32). It is also crucial to ensure that both plates are completely flat and parallel, ensuring uniform pressure application as the joints are being cured.

2.3 Fracture Mechanics Specimens

Figure 2.32

Removal of dirt and dust from the mold.

Figure 2.33

Application of release agent in the mold.

The surfaces that contact with the joints should be covered with a mold release agent, selected as a function of adhesive family that will be used. The release agent should be applied carefully, coating all parts of the mold, taking particular attention to the pins that contact with the adhesive (Figure 2.33).

2.3.3

Surface Treatment of Adherends

The surface preparation of the adherends to be bonded is an operation that requires a great deal of care in order to obtain satisfactory results. The aim of this step is to ensure that the adhesion is such that the weakest point in the joint is either the adhesive or the adherend, reducing the chance of breakage at the interface. The available methods for surface preparation include passive mechanical and chemical processes. Passive processes include mechanical abrasion and shot blasting, while active processes involve changes of the chemical composition of the adherend surface. All surface treatments aim to ensure a good degree of adhesion between the adhesive and the adherend and thus guarantee that the adhesion forces are greater than the cohesion forces, thus ensuring that a cohesive failure will occur and never an adhesive failure. When the fracture properties of the adhesive subjected to aging conditions are to be determined, adherends made from aluminum are typically used. In this case, the surface treatment that ensures a good degree of resistance under aging conditions

49

50

2 Production and Testing

Spacer Adherend Razor blade Adhesive Adherend

Spacer (a)

(b)

(c)

Figure 2.34 Technique used to create a pre-crack in the bonded area, (a) schematic representation, bonding of razor blade to the spacer (b) and bonding of top spacer.

is the anodization of the adherends, following the ASTM D3933 standard or similar. The main steps for this treatment are described in Section 2.8.

2.3.4

Adhesive Spacers

In order to ensure uniform adhesive thickness along the joint length, a calibrated steel tape is used, maintaining correct spacing between the adherends. These calibrated steel tapes are also used to introduce a pre-crack in the bondline (Figure 2.34). Fracture tests require the presence of a crack, and the damage initiation is the result of the pre-crack, which is the result of an arrangement composed by a razor blade bonded with a calibrated steel tape to perform the final thickness of the bondline. With the blade placed under a clean surface, small drops of cyanoacrylate are applied to it. The calibrated steel tape is placed in a way that only the edge of the blade is visible and bonded to the steel tape. A few drops of cyanoacrylate are applied to the free surface, and another calibrated steel tape is placed over the blade. In standardized DCB and ENF specimens, the resulting spacer with a razor blade is positioned at 45 mm from the loading holes.

2.3.5

Specimen Manufacture

Surface preparation is an important step in an adhesive bonding process to assure the quality of the joint. The lower arms of the test specimens are first placed between the guide pins and the calibrated steel tapes are placed in the edge positions. The spacer without a razor blade could be made longer, and its positioning is not critical because its main purpose is to control the adhesive thickness (Figure 2.35). The spacer with

2.3 Fracture Mechanics Specimens

Figure 2.35

Positioning of adherends and spacers in the mold.

the razor blade in the middle has two main purposes, to control the adhesive thickness and to introduce an artificial pre-crack. The calibrated steel tapes should be coated with the mold release agent in order to ensure that the adhesion between the adhesive and the spacers is minimum, ensuring that this will not affect the strength of the adhesive during the test. The application of the adhesive in the adherend can be performed manually with a spatula or with a powered gun. When two-part adhesives are used, it is necessary to ensure uniform mixing of the two components. This can be performed using mixing nozzles (when the adhesive is supplied in a cartridge) or using specialized mixing equipment, such as high-speed mixers, and ultrasound and vacuum devices (when the adhesive is supplied in separate containers), as is shown in Figure 2.36. The areas to be bonded are considerable, so when fast-curing adhesives are used, such as acrylic adhesives, the adhesive application needs to be very quick in order to ensure that the adhesive is still in an uncured state when the specimens are assembled and the mold is closed. After depositing the adhesive on the lower arms of the test specimens, the upper arms are placed over the adhesive layer very carefully, taking care to avoid the displacement of the spacers and blades (Figure 2.37). Pressure is carefully applied over the upper arms to ensure the homogenization and the overflow of the excess adhesive. Finally, the upper part of the mold is placed over the specimen assembly and the mold is introduced in a hot plate press, able to control the pressure and temperature. Please note that pressure is not a critical factor in the manufacture of these specimens as the load is supported by the spacers and is not directly transferred to the adhesive layer.

2.3.6

Final Preparation of Specimens

After the adhesive is fully cured, the specimens can be removed from the mold. The excess adhesive should be manually removed using a file and coarse sandpaper (Figure 2.38).

51

52

2 Production and Testing

(a)

(b)

Figure 2.36

Adhesive application using a spatula (a) and a nozzle (b).

There are multiple methods available to measure the critical strain energy release rate, often requiring the direct measurement of crack length as the test progresses. In order to allow for this measurement, the bondline can be polished and painted with a white ink, which will readily show the location of the crack. Moreover, a ruler can also be bonded to the upper adherend to serve as a reference for measuring the crack length during the fracture tests.

2.3.7

Testing Procedure

The DCB or ENF specimens can be tested in a universal testing machine (electromechanical or servo-hydraulic testing machine), recording the load–displacement (P−𝛿) data. The DCB setup consists of two loading supports that connect with each arm of the specimen through a pin (Figure 2.39a). A load cell with 10 kN is suitable for these tests. The ENF setup is based on a three-point bending setup,

2.3 Fracture Mechanics Specimens

Figure 2.37

Placement of upper adherend using a rotational movement.

where the specimen is supported by two rollers and loaded in the middle through a third roller, which is forced to displace vertically. The rollers should rotate freely in order to avoid friction effects during testing. For this type of setup, it is also necessary to use a load cell with a capacity of 20 kN or higher as higher loads are verified, especially when metallic adherends are used. This is because under mode I, the resistant area can be assumed to be a single line, while under mode II, the resistant area is all of the bonded area subjected to an almost uniform shear load. Furthermore, as the specimen bends during the test, the ENF arms tend to contact and slide under friction (Figure 2.39b). This friction will erroneously be added to the measured energy, so, in order to reduce this friction, a Teflon layer with oil should be positioned between both arms.

2.3.8

Data Reduction Schemes

The determination of the critical strain energy release rate is not direct as it requires the data treatment of the recorded data. The data reduction schemes that typically are used to determine the critical strain energy release rate can be divided into two

53

54

2 Production and Testing

Figure 2.38

Removal of excess adhesive using a coarse file.

Load

(a)

Load Load

(b)

Figure 2.39

Schematic representation of DCB test (a) and ENF test (b) loading conditions.

2.3 Fracture Mechanics Specimens

(a)

(b)

Figure 2.40 Setup used to measure the real crack propagation during the DCB test (a) and the ENF test (b).

different groups. The first includes the methods that require the measurement of crack length, and the second includes the methods that do not require the direct measurement of crack length. Examples of methods that require the measurement of crack length during the test are the compliance calibration method (CCM), the direct beam theory (DBT), and the corrected beam theory (CBT). These methods require the implementation of a methodology to record the crack propagation and require a special preparation bondline to facilitate the measurement of the crack length (Figure 2.40). As stated above, this is usually carried out by painting the specimens with a thin layer of white ink, which will allow us to better identify the crack position. This can also be assisted by the use of a ruler to provide a clear reference on the crack length position. However, it is very difficult to precisely measure the position of crack length as there is no guarantee that the crack seen on the outside of the specimen corresponds to the true front of the crack region. Thus, approaches that do not require the direct measure of the crack length have grown in popularity. These include the compliance-based beam method (CBBM) and the beam theory including bending rotations (BTBR).

55

2 Production and Testing

2.4 Single-Lap Joint Specimens 2.4.1

Introduction

The SLJ is one of the most commonly employed bonded joint geometries, being simple and cost-effective. This joint is also extensively used as a validation specimen used to quickly assess the performance of new adhesives and surface treatments at a joint level. Several standards provide guidance on how to prepare joints with flat adherends, such as ASTM D 1002 and ISO 4587. These standards recommend the production of multiple joints from a single set of two large, bonded plates cut and machined into the final desired dimensions. However, this technique does have some disadvantages such as the need for machining operations after curing, which can introduce or create some cracks in the adhesive and introduce temperature changes that can alter the mechanical properties of the adhesive. This can be mitigated with the use of cutting/cooling fluid, but this fluid might be absorbed by the adhesive, changing its mechanical properties. Alternatively, the individual production of complete SLJ specimens avoids the contamination of the joints with cutting fluid and also allows us to eliminate the probability of mechanical degradation of the joints as a result of the heat released during the cutting process. In this case, adherends are pre-cut to the final dimensions before bonding, subjected to a surface preparation procedure (if needed) and then bonded together using some sort of fixing device to control the joint geometry. Figure 2.41 shows the standardized dimensions of SLJs manufactured according to ASTM D 1002.

2.4.2

Surface Treatment of Adherends

Adequate selection and application of surface preparation is crucial to ensure an effective level of adhesion between the adhesive and the adherends. As is the case for other joint geometries, different surface treatments can be used in SLJs to improve the adhesion level using passive or active methods. Fundamentally, the purpose of the surface treatment is to ensure excellent adhesion between the adhesive and 12.7 Grip area 25.4

56

63.5

25.4

101.6 1.62

Figure 2.41

Geometry of SLJs according to ASTM D 1002 (dimensions in mm).

2.4 Single-Lap Joint Specimens

the adherend, guaranteeing that the adhesion forces are stronger than the cohesion forces, ensuring that cohesive, rather than an adhesive, failure will always occur. The selection of the treatment should always be performed as a function of the adherend material. Proper selection of the treatment will lead to good adhesion between adherend – primer and/or adhesive. The following list summarily describes surface treatments and adhesives that are most suitable to bond different types of adherend materials ●

●

●

Metal (steel, titanium, and aluminum): – Surface treatments – mechanical and chemical treatments; – Adhesives – epoxy, modified acrylic, polyaromatics, phenolics, and polyurethanes; Composites (glass fiber and carbon fiber): – Surface treatments – mechanical, chemical, and physical treatments; – Adhesives – epoxy, modified acrylic, polyaromatics, phenolics, and polyurethanes; Polymers: – Surface treatments – chemical and physical treatments; – Adhesives: epoxy → all except PVC (flexible); Modified acrylic → PE, PP, epoxy, and phenolic; Phenolics → PE, PP, PA, and phenolics; Polyurethanes → PVC (flexible and rigid), epoxy, and phenolic.

In Figure 2.42, three different surface treatments are shown, all suitable to increase the adhesion (e.g. sandblasting for metals and plasma for polymers) or to improve the durability (e.g. anodizing for non-ferrous metals).

2.4.3

Joint Manufacture

To manufacture geometrically consistent SLJs, a metallic mold should always be used, properly cleaned, degreased, and coated with the release agent, as described in detail in Section 2.1.3. Adherends and shims should be precisely positioned to define the area where the adhesive will be applied. In fact, shims play three different important roles. They control the overlap length, the adhesive thickness, and the adhesive fillet. Figure 2.43 shows a scheme of adhesive joints being manufactured with the shims, illustrating how the adhesive fillet is controlled. The shims must be coated with mold release agent in order to facilitate their removal after full curing. The use of shims is the most effective way to control the shape of the adhesive fillet in SLJs. Other techniques can also be used to control the adhesive thickness, such as the use of calibrated wires of glass spheres (Figure 2.44). However, this can lead to a decrease of adhesive strength and also requires precise control of both the pressure applied and of the position of the calibrated wire. This is crucial to ensure that thickness control elements remain in position as the excess adhesive flows away from the joint. The process of adhesive application is also complex, requiring special care to avoid the retention of air inside the bondline. This is similar to what was mentioned for

57

58

2 Production and Testing

(a)

(b)

(c)

Figure 2.42 Different surface treatments: (a) sandblasting, (b) anodizing, and (c) atmospheric plasma treatment.

2.4 Single-Lap Joint Specimens

Adherend Shim

Figure 2.43 area.

Shim

Adhesive Adherend

Construction scheme of SLJs using shims to control the parameters of bonded

Pressure

Glass sphere or ballotini

(a)

Calibrated wire

(b)

Figure 2.44 Different techniques to control the adhesive thickness (a) with glass spheres and (b) with calibrated wire.

59

60

2 Production and Testing

(a)

(b)

Figure 2.45

Application of the adhesive using a spatula (a) and with a nozzle (b).

DCBs. Furthermore, the use of shims allows us to clearly identify the area where the adhesive will be applied, optimizing the amount of adhesive to be used (Figure 2.45). After the adhesive is applied, the upper adherend must be carefully positioned over the lower adherend in a smooth rotational movement, continuously providing a path for the escape of trapped air (Figure 2.46).

2.4.4

Final Preparation of Specimens

Once the adhesive is fully cured, joints can be removed from the mold. At this stage, the shims should also be separated from the joints, and the excess adhesive should be removed using a file and sandpaper (Figure 2.47). As the fillet provides additional joint strength, it might not be recommendable to fully remove it, unless the specific testing procedure being followed requires so.

2.4.5

Testing Procedure

SLJs can be tested in a universal testing machine or other type of testing machines (such as drop weight impact testers or servo-hydraulic testing machines). One of

2.4 Single-Lap Joint Specimens

Figure 2.46 technique.

Positioning of upper adherend after applying the adhesive using rotation

Figure 2.47

Removal of excess adhesive from the joint.

Figure 2.48

The misalignment of SLJ during the test.

the main issues associated with testing SLJs is the significant misalignment of one adherend in relation to the other which, during testing, will force the SLJ to rotate in order to align the load transfer through the adhesive layer (Figure 2.48). This misalignment will introduce an important bending moment at the ends of overlap length, eventually leading to premature failure of the joints. One way to minimize the effect of this misalignment is the application of tabs at the ends of the adherends, creating a balanced geometry that will drastically reduce the rotation of the joints during the test. These tabs can be bonded to the adherends or fixed using fasteners. Should a bonded tab be used, this operation can be carried

61

62

2 Production and Testing

(a)

(b)

Figure 2.49 SLJ testing setup using bonded tabs (a) or tightened tabs, both seeking to reduce misalignment.

out either during the manufacture of the SLJ (called co-curing) or at a later stage on already prepared joints. Testing of SLJs can be performed using vice grips or special testing fixtures composed of plates that are firmly tightened together using fasteners (Figure 2.49). If sufficient pre-load torque is applied to these fasteners, any slippage during these tests can be almost completely eliminated. Because of the SLJ geometry, rotation of the specimen will occur during testing, which can be reduced using tabs (Figure 2.50). This is especially important as this rotation introduces bending moments at the ends of overlap length that can cause premature failure of the joints. After failure, it is important to see if the any plastic deformation was present on the adherend as this can mean that the failure was controlled by the adherend and not by the adhesive. The load–displacement behavior of the joints is recorded during testing, with this data being logged until joint failure occurs.

2.4 Single-Lap Joint Specimens

Figure 2.50 Rotation during testing that introduces bending moments at the ends of overlap length.

In order to better understand the failure mechanism of the joint, the resultant fracture surfaces are usually analyzed in detail. This can be performed with a simple visual observation of the surfaces, or it can be carried out in more detail using optical or electronic microscopy.

63

65

3 Laboratorial Activities and Report Examples 3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints 3.1.1

Introduction

Surface treatment plays a key role in the strength of an adhesively bonded joint. If there is low adhesion between the adhesive and the adherend, the adhesive joint will fail prematurely at the interface and will not carry the load effectively between the two bonded adherends. This type of failure is called an adhesive failure and usually occurs when the surface energy of the adherend is lower than that of the adhesive. As adhesives typically have low surface energy, adherends with high surface energy do not exhibit this issue and will create a durable and strong bond with the adhesive layer. The surface energy can be assessed by analyzing the contact angle of a water droplet on a surface. Materials with high surface energy will exhibit good wetting, meaning that the droplet will form a very low contact angle with the surface, while materials with low surface energy will repel the water droplet and lead to large contact angles (Figure 3.1). Most polymers exhibit very low surface energy, which makes them poorly suitable for adhesive bonding processes. However, many different surface treatments can be employed to significantly increase their bondability. Surface treatments can be divided into active and passive processes. Passive processes are those that do not change the material chemically and include all abrasive processes and the use of solvents and detergents to clean the surface from impurities and contaminants. In contrast, active processes will chemically modify the adherend. Among these processes are chemical etching, anodization, flame treatment, corona discharge, and plasma treatment. These processes differ greatly in their action. For example, while abrasion with sandpaper leads to better adhesion by creating a macroscopically rough surface that will enable a larger degree of mechanical interlocking between the adhesive and the adherend, a plasma treatment will instead use highly ionized air to impart energy into polymeric materials and break the polymer chains, creating a new and highly chemically reactive surface that will form bonds with the molecules of the adhesive (known as a functionalized surface). Adhesive Bonding Technology and Testing, First Edition. Ricardo João Camilo Carbas, Eduardo André Sousa Marques, Alireza Akhavan-Safar, Ana Sofia Queirós Ferreira Barbosa, and Lucas Filipe Martins da Silva. © 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.

66

3 Laboratorial Activities and Report Examples

θ θ

Liquid Solid

Liquid Solid

(a)

(b)

Figure 3.1

Example of surfaces with lower wetting (a) and good wetting (b).

3.1.1.1 Joint Strength Prediction

To predict the behavior of a single-lap joint (SLJ), it is essential to understand the stresses that are acting on it. Most analysis assumes that the adherends are rigid and that the adhesive only undergoes deformation because of shear. In this case, the shear stress, 𝜏, for the adhesive joint shown in Figure 3.2 is given by Eq. (3.1): P (3.1) 𝜏= bl where P is the applied load, b corresponds to the joint width, and l is the overlap length. For highly ductile adhesives (those with more than 20% of shear strain), all overlap length undergoes plastic deformation (generalized yield criteria), and the failure load (Pmax ) can be estimated by simply considering the adhesive’s yield strength under shear (𝜏 y ) and the bonded area, calculated as the width (b) multiplied by the overlap length (l) (Eq. (3.2)): Pmax = 𝜏y b l

(3.2)

If the adhesive and the interface are sufficiently strong, the failure can occur entirely outside of the bonded area, only in the adherend. This is the preferred mode of failure when designing an adhesively bonded joint, as it means that the joint is stronger than the material of the adherends and that it does not introduce a weak point to the bonded structure. P b t I

Shear stress on the adhesive (τ)

Figure 3.2

Shear stress on a single-lap joint: most basic analysis.

P

3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints

In this case, the Pmax can be accurately estimated by considering the tensile strength of the adherend material (𝜎 y ), multiplied by the cross section of the adherend, calculated as the adherend width (b) times the adherend thickness (t). This is the adherend yielding criterion (Eq. 3.3): Pmax = 𝜎y b t

3.1.2

(3.3)

Work Description

Manufacture and determine the strength of three different SLJs bonded with a two-component polyurethane adhesive: ● ● ●

an SLJ with untreated polymeric adherends; an SLJ with polymeric adherends abraded with sandpaper; an SLJ with polymeric adherends treated with atmospheric pressure plasma.

The failure load of the bonded joints shall be determined using the generalized yield of the adhesive criterion and the yielding of the adherend criterion.

3.1.3

Materials

3.1.3.1 Adherends (Tables 3.1 and 3.2) Table 3.1

High-density polyethylene (HDPE) properties.

E (MPa)

𝛎

w (MPa)

1500

0.46

22

Table 3.2

Adherend dimensions (HDPE).

Length (mm)

Width (mm)

Thickness (mm)

120

25

2

3.1.3.2 Adhesive (Table 3.3) Table 3.3

Adhesive properties. SikaForce 7817 L7

Young’s modulus, E (MPa) Poisson’s ratio, 𝜐

0.38

Shear yield strength, 𝜏 y (MPa)

20

3.1.4 ● ●

2500

Experimental Work

Lightly abrade the bonded area with sandpaper for a pair of HDPE adherends. Apply the atmospheric plasma treatment (one minute at 1 cm of distance) for another pair of HDPE adherends.

67

68

3 Laboratorial Activities and Report Examples ●

●

●

● ● ● ● ● ● ● ●

Clean the bonding area for the abraded, plasma-treated, and untreated adherends, with a paper cloth embedded with acetone. Determine the water droplet contact angle of the three different adherend preparations using an optical contact angle measurement equipment. Assemble the adherends in the mold and use spacers to obtain the intended adhesive thickness (0.2 mm) and overlap length (25 mm). Mix the two components of the SikaForce 7817 L7 adhesive. Apply the adhesive on the bonding areas of the three specimens. Position the upper adherend over the lower adherend and close the mold. Apply the weights over the mold lid. Remove excess adhesive with a file and sandpaper. Mark and reference each specimen. Measure the final adhesive thickness using a caliper. Test the complete joints in tension using a universal testing machine, registering the failure load and the mode of failure.

3.1.5

Report

The report should include the following: ●

● ● ● ●

●

prediction of the failure loads of the adhesive joints using the generalized yield of the adhesive criterion and the yielding of the adherend criterion; results of the water contact angle test; plot with the failure loads and the failure load predictions; images of all the fracture surfaces and comments; comparison between predicted failure loads and experimental failure loads using the contact angle measurements to better explain the behavior; indication of the better suited surface treatment and the most suitable prediction method for each case under study with proper justification. REPORT 3.1.5.1 Introduction

In this practical activity, the effects of the surface treatment on the mechanical behavior of adhesively bonded joints were studied. In order to allow a better understanding of the practical activity results, brief bibliographic research was performed. In adhesive joints, three main types of failure might occur. These are as follows: ● ● ●

Cohesive failure in the adhesive; Cohesive failure in the adherend; Adhesive failure.

3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints

Cohesive failure in the adhesive occurs when there is rupture of the adhesive without detachment from the adherend. This might happen because of the breakdown of the intermolecular bonding forces in the adhesive substance. Despite not being the ideal failure type, it can be predicted because the adhesive properties are known. Cohesive failure in the adherend is the ideal failure mode as the weak link in the connection is the adherend and not the adhesive itself. This kind of failure happens when the adherend material used is not as strong as the adhesive, for a given applied load. Adhesive failure occurs when an adhesive layer completely separates from the adherend. This means that the adhesion of the adhesive with the adherend is poor. As expected, one should never design the joint for this type of failure. Adhesive failure happens at the interface between the adhesive and the adherend, being the weakest link in that joint. The properties of the interface are not known, and therefore, it is impossible to predict when the joint is going to fail [1]. Adhesive failure usually happens when the surface energy of the adherend is lower than the surface energy of the adhesive. If the surface of the adherend is properly treated before making the joint, the only possible failure types are the cohesive ones, which is acceptable. Moreover, to ensure that this last type of failure does not occur, the surface treatment must be performed correctly. Surface treatments should precede any other step or procedure in the adhesive bonding technology [2]. According to the theory of adhesion, the surface energy can be assessed by analyzing the behavior of a liquid droplet in a surface, especially by the contact angle, 𝛼, between both. Materials with high surface energy will maintain a low contact angle with the droplet, while a low surface energy material will have a greater angle, repelling the droplet, just as shown in Ref. [1]. Wetting can be observed when a drop of liquid spreads over the surface of a solid-state material and is achieved when the surface energy (or tension) of the liquid is lower than the surface energy (or energy) of the solid. In theory, the contact angle ranges from 0∘ (with complete spreading) to 180∘ (with no wetting), as can be seen in Figure 3.3. θ ≈ 90° θ ≈ 0°

Liquid Solid (a)

Figure 3.3

Liquid Solid

(b)

Different types of wetting, spread (a) and poor wet (b).

(Continued)

69

70

3 Laboratorial Activities and Report Examples

(Continued) The surface energy or surface tension is related to the imbalance of attraction forces at the surface of a material. In fact, if we take a liquid as an example, while the molecules in the interior of it are in equilibrium, as the forces in all directions are balanced, the molecules near the surface are not. They are constantly being pulled down as a result of the surface energy. The contact angle is also a measure of wetting as low angles mean good wetting of the surface. If the adhesive can wet the adherend properly, 𝛼 < 90∘ , it means that we are going to have good adhesion between them. Usually, the surface energy of adhesives is much lower than the adherends used. However, in some cases, the opposite occurs. Therefore, surface treatments can be applied to increase the surface energy of the adherend and to ensure a more durable adhesive joint [1]. A surface treatment should produce a clean and stable surface, suitable to be bonded with the primer and/or adhesive. The most important for the joint manufacture process is to remain reliable in a production environment and to be suitable for different process variables. The surface preparation of the adherend is one of the factors with highest impact on the quality and durability of adhesively bonded joints. In order to ensure good surface preparation, different requirements are needed such as follows: ●

●

●

●

●

Surfaces must be cleaned from any undesirable contaminations by removing any dust, rust, oil, and miscellaneous dirt in the surfaces of the adherends; The adhesive must wet the adherend surface to bond. Sometimes, these contaminations can obstruct the desirable wetting of the adhesive on the surface to bond and therefore create a poor adhesive bond, unless the adhesive used in the joint is specifically designed to absorb these contaminants. Organic solvents, water-miscible detergents, polar hydrocarbons (e.g. esters, ketones, and alcohols), and acetones are some common cleaning agents; Surface preparation must enable the formation of strong bonds (chemical or physical) between the primer or the adherend and the adhesive and, if necessary, between the primer and adherend interfaces; The interfaces must ensure a strong and stable bond for all the lifetime of the bonded structure; The treated surfaces must be reproduced with same characteristics, independently of materials variables (i.e. surface contamination, mill scale, and even alloy or heat treatment).

The surface treatment of adherends removes or delays the formation of a weak layer on the adherend surface, promotes the molecular interaction between the adhesive or the primer and the adherend, develops a desired surface microstructure of the adherend, and optimizes the adhesion forces (between adhesive and/or primer and adherend interfaces) to ensure sufficient initial and long-term joint strength. Surface treatments can be classified as passive or active. In a passive treatment, chemical alteration of the adherend

3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints

surface (mechanical treatments) does not occur, and in passive treatment, some sort of chemical alteration occurs (chemical and physical treatments). 3.1.5.2 Experimental Procedure 3.1.5.2.1 Surface Treatment The first step of this experimental procedure is to

apply the surface treatments to each of the two sets of adherends. The first treatment was only to clean the surfaces that will be bonded with acetone, while the second included an atmospheric plasma treatment. Subsequently, to assess and compare the surface energies of the three types of treated surfaces and to predict the adhesion effect, the water droplet contact angle was measured with an optical contact angle measurement equipment using ethylene glycol as the liquid, and an approximation of the value was made through the use of Dyne pens. 3.1.5.2.2 Joint Assembly After the surface treatment, the lower adherends were

assembled in a mold, adequately set up for an overlap length of 25 mm and an adhesive thickness of 0.2 mm. With this, the adhesive was applied. The adhesive compound used was a SikaForce 7817 L7 adhesive, of ductile characteristics, stored in a container and ready to use. Its application was done manually with a spatula, uniformly all the overlap length. Afterward, the upper adherends were placed, followed by the mold’s cover. 3.1.5.2.3 Tensile Tests Following the cure of the adhesive, the three joints were subjected to tension testing, with a universal testing machine, in order to obtain the load–displacement graphs and the failure load. A visual inspection was also performed to better understand the failure mode. 3.1.5.3 Materials

In order to facilitate the analysis of the experimental results and justify the values used while applying the theoretical failure criteria, Table 3.4 shows the properties of the adherends and adhesives used in this first practical activity. Table 3.4 Mechanical properties of the polymeric adherend and of the adhesive used. Adherend PE–HD

Adhesive SikaForce 7817 L7

Young’s modulus (E)

1500 MPa

2500 MPa

Poisson’s ratio (𝜈)

0.46

0.38

Yield strength (𝜎 y )

22 MPa

20 MPa

(Continued)

71

72

3 Laboratorial Activities and Report Examples

(Continued) 3.1.5.4 Failure Load Prediction 3.1.5.4.1 Generalized Yield of the Adhesive When working with ductile adhe-

sives – with more than 20% of shear deformation, it is acceptable to estimate the breaking force, P max , using Eq. (3.4), as the generalized yield of the adhesive criterion usually results in a good approximation. 𝜏y =

Pmax ⟺ Pmax = 𝜏y b l bl

(3.4)

Ductile Adhesive – L7818 L7

𝜏y = 20 MPa b = 25 mm ●

For l = 25 mm → P max = 12500 N

3.1.5.4.2 Adherend Yield The yielding of the adherend criterion was applied. According to Adams et al (1997), the failure load can be given by Eq. (3.5). The factor k accounts for the effect of the rotation of the joint in the relation between the applied load and the generated bending moment. For small loads and overlaps, k tends to 1, whereas for bigger overlap lengths when compared to the thickness of the adherend, k tends to 0, which translates in a complete plastic deformation of the adherend. 𝜎y b t (3.5) Pmax = 1 + 3k l l < 20 ⇒ k = 1, ≥ 20 ⇒ k = 0 t t Lower failure loads will be achieved mainly because of the usage of polyethylene adherends with limited shear yield strength. High-Density Polyethylene

𝜎y = 22 MPa b = 25 mm ●

t = 2 mm

For l = 25 mm → P max = 790 N 25 l = = 12.5 ⇒ k = 0.131 t 2

3.1.5.5 Experimental Results and Discussion

In this section, the experimental results regarding the effect of surface treatment will be shown and discussed. Two surface conditions were analyzed, untreated, and atmospheric pressure plasma treated. A solvent, acetone, was the only process applied to the untreated surface. The results drawn from this experience are the contact angle between a droplet of ethylene glycol and the surface of the bonding area of the PE adherend, the testing pens’ markings, and the tensile test plots, as well as images of the fracture surfaces.

3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints

3.1.5.5.1 Contact Angle Measurement In the following figures, the contact

angles between a droplet of H2 O over the PE adherend are shown for the two surface treatments. The results are also summarized in Table 3.5. Table 3.5

Contact angles after three distinct surface treatments.

Left contact angle Right contact angle

Untreated (acetone)

Plasma treatment

95.8∘ 95.6∘

24.1∘ 25.1∘

As PE is a polymer, it is expected for its surface to have a low surface energy, which will cause a low wetting of the surface and, consequently, poor spreading of the droplet, resulting in a large contact angle. This can be seen in Figure 3.4, which shows that the contact angles of the untreated surface are 95.8∘ and 95.6∘ . As a rule of thumb, contact angles of less than 90∘ will, most of the times, guarantee a strong and durable bonding between the adhesive and the adherend. In this case, as the contact angle is higher than the threshold, one can predict that the adhesion of the adherend will be very limited.

Figure 3.4

Contact angles between the ethylene and PE surface untreated (acetone).

In some cases, when the surface energy of the adherend is low and the droplet makes a contact angle being higher than 90∘ , mechanical passive treatments are used. However, in some cases, they might have the opposite effect, reducing adhesion levels even further. Plasma treatment, being an active process, is expected to generate the highest surface energy from the cases under study. By treating the surface with (Continued)

73

74

3 Laboratorial Activities and Report Examples

(Continued) atmospheric pressure plasma, the chemistry of the surface alters, as the surface is oxidized, and polar groups are formed. This results in a huge increase of the surface energy and adhesion. Accordingly, the contact angles are shown to be low, registering at 24.1∘ and 25.1∘ . Moreover, as can be seen in Figure 3.5, the droplet wets the surface properly, assuring a good starting point for the adhesion process.

Figure 3.5 Contact angles between the ethylene and PE surface treated with atmospheric pressure plasma.

From these images, it is possible to conclude that a surface treatment will increase the surface energy, as the contact angles decrease in comparison with the untreated surface. Moreover, the plasma treatment is shown to be the one that affects the surface energy and contact angle the most as it changes the chemical properties of the adherend, making it highly advantageous for these endeavors. Dyne Pens Another way of roughly measuring the surface energy is using Dyne pens. In this testing method, a more uniform spread of ink means a more consistent adhesion between the test ink and the analyzed surface and, of course, a more accurate approximation to the adherend’s surface energy, in mN/m (Dyne). This method is a more practical and cheaper alternative to the one explored before, but it falls short on providing exact values. With this method, it is possible to confirm what has been assessed in the previous section. By visual inspection of Figure 3.6, it is perceptible that while the pens with surface energy 30 and 32 can wet the adherend relatively well, the one with 34 cannot. This means that the surface energy of the untreated adherend is somewhere between 30 and 32 Dynes.

3.1.5.5.2

3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints

Figure 3.6

Testing pens with untreated bonding area.

In Figure 3.7, it can be seen that the 44 Dyne pen can properly wet the surface. This means that the surface energy of the adherend is higher than that value. Another limitation of this test is that the user needs to use a large number of pens to accurately evaluate the surface energy. For the plasma-treated surface, we know that the surface energy is higher than 44. However, one should have testing pens with higher Dyne value to be able to determine more precisely the range in which the surface energy value is as its value can be higher than 44.

Figure 3.7

Testing pens with plasma-treated bonding area.

These results match the ones previously obtained, in which the untreated adherend had the lowest surface energy and plasma treatment provided the highest. (Continued)

75

3 Laboratorial Activities and Report Examples

(Continued) Adhesive Joint Tests After the curing of the adhesive, joints were tested and submitted to a tensile test. The results of this process can be seen in Figures 3.8 and 3.9.

3.1.5.5.3

1.2

Load (kN)

1 0.8 0.6 0.4 0.2 0

0

2

4

6

8

10

12

Displacement (mm) Plasma treatment

Figure 3.8

Acetone cleaning

Load–displacement curves of plasma treated and cleaned joints.

1.2 1 0.8 Load (kN)

76

0.6 0.4 0.2 0

0

50

100

150

200

250

300

350

400

Displacement (mm) Plasma treatment

Figure 3.9 plasma.

Acetone cleaning

A typical load–displacement curve of the joints with surfaces treated with

In this graph, it is possible to have an overview of the comparison between the different surfaces. This allows us to conclude that higher surface energy conducts to a better adhesive joint, which consequently elevates the failure load.

3.1 Effect of Surface Treatment on the Mechanical Behavior of Adhesively Bonded Joints

The failure load of the untreated surface (acetone) is the lowest one, with a value of approximately 0.1 kN. A higher contact angle results in less contact and proximity between the adhesive and the surface, which means that there are less Van der Waals or attraction forces at play. With a plasma-treated surface, failure of the joint was not registered, and the load and displacement achieved until plastic deformation of the adherend were significantly higher than those from the other joints, as it can be seen in Figure 3.9. These values can be explained by the high surface energy, as the adhesive is very well spread, properly wetting the adherend and originating a strong adhesion. In addition, with this active surface treatment, the chemical composition of the adherend changes because the molecules break and react with oxygen, forming polar groups. This way, it is possible to obtain primary bonds, which are the strongest forces according to the adhesion theory. Until the adherend yielding occurred, a relatively large load of 0.9 kN was registered. These tests show that joint strength can be improved by 200% just by treating the surface properly, without having to change the type of adhesive, type of adherend, or overlap length. Surface treatment is something that cannot be disregarded and plays a crucial role in adhesive bonding. 3.1.5.5.4 Failure Mechanism In this final section, the mechanical behavior of the joints in the tensile tests is analyzed. Through visual inspection and data analysis from the previous section, the types of failure in the joints are also investigated. For the untreated surface (acetone), one can observe that the adhesive completely disbonded from one of the adherends (Figure 3.10). This fact indicates that the type of failure was adhesive in nature. In fact, in this case, the interface between the adherend and the adhesive was the weakest link in the joint. As explained before, this type of failure is the worst-case scenario as it cannot be reliably predicted, and the interface properties are generally unknown. Usually, this kind of failure is due to inappropriate surface treatment selection.

Figure 3.10

Fracture of the untreated bonded area.

(Continued)

77

78

3 Laboratorial Activities and Report Examples

(Continued) The two tests performed to measure the surface energy had already indicated that the surface was not ready for the bonding process, and this fact is confirmed by premature failure in the tensile test, concluding that this kind of passive surface treatment is not enough to improve the adhesion in polymeric adherends. For the plasma-treated surface, the mechanical behavior of the joint was very different. Through the analysis of Figure 3.11, it can be seen that the joint did not fail. In this case, the polymeric adherend started to deform plastically before the adhesive joint failed, which means that the adhesive joint is even stronger than the material it is bonding. The type of failure identified is cohesive failure of the adherend.

Figure 3.11

Mechanical behavior of the plasma-treated bonded area.

From the adhesive perspective, this is the perfect case because the mechanical failure occurred not on the adhesive itself but on the adherend instead. 3.1.5.6 Conclusions

The main objective of this work was to understand the influence that surface treatment has on joint strength. By performing the tensile test, it was possible to obtain the experimental failure load for both joints performed (treated with plasma and untreated), as well as the failure mechanism. The best surface treatment, from the studied ones, is the chemical plasma treatment as it significantly increases the surface energy of the adherend. It was evident once again that correct surface treatment selection is imperative when adhesive joints with excellent properties are to be obtained. In fact, surface treatment of the adherend enables the strength of the adhesive joint to be increased. In the absence of surface treatment, failure tends to occur at the interface between the adhesive and the adherent. This is a consequence of reduced adhesion and contributes to the greater difficulty in predicting the failure load of the joint (as the failure occurs in a zone where the properties neither match those of the adhesive nor those of the adherend). Regarding the best method to predict the behavior on the joints, for the abrasion and solvent treatment, there is no criterion to best describe the behavior of the joints as the fracture is not cohesive; hence, it is unpredictable, and there is no way to know the properties of the interface between the adhesive and the adherend. In the case of the plasma treatment,

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

there is only plastic deformation of the adherends, which means it is possible to predict its behavior, as the material properties are known. The most appropriate prediction method proved to be the yielding of the adherend method as the yield strength of the adherend is much lower than the one from the adhesive. These conclusions are in accordance with the surface energy predictions and experimental results.

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints 3.2.1

Introduction

Volkersen in 1938 introduced the concept of differential straining or shear lag analysis to this problem [3]. Volkersen stated that if the adherends are elastic and continuity is assured on the adhesive/adherend interface, the parallelogram shown in Figure 3.12a) for the unloaded joint will become distorted under load, as shown in Figure 3.12b). The analytical solution found by Volkersen for the shear stress is schematically shown in Figure 3.12c. The value of shear stress in the adhesive is maximum at the edges of the overlap area and minimum in the center of the overlap area. Considering Figure 3.13, the analytical solution of the Volkersen model can be obtained by balancing the loads in the upper adherend: 𝜎1 bt1 + 𝜏bdx = (𝜎1 + d𝜎1 )bt1 ⇒

d𝜎1 𝜏 = dx t1

(3.6)

Unloaded

Loaded P P τ

Shear stress distribution

Figure 3.12 Volkersen analysis, joint unload (a), joint loaded (b), and shear stress distribution (c).

79

80

3 Laboratorial Activities and Report Examples

E1

P

t1

Ga

ta

b

E2

P

t2 dx

x I σ1bt1

(σ1+dσ1)bt1 τbdx

τbdx σ2bt2

(σ2+dσ2)bt2

Figure 3.13 Single-overlap joint as analyzed by Volkersen (1938). Source: Adapted from Volkersen [3].

Balancing the loads in the lower adherend: 𝜎2 bt2 = (𝜎2 + d𝜎2 )bt2 + 𝜏bdx ⇒

d𝜎2 𝜏 =− dx t2

(3.7)

Balancing the complete joint: P = 𝜎1 bt1 + 𝜎2 bt2

(3.8)

Considering the shear deformation in the adhesive: } ) ( ( ) 𝛾 = G𝜏 𝜎2 d𝛾 1 𝜎1 1 d𝜏 1 du1 du2 a = − = = − ⇒ 𝛾 = t1 (u1 − u2 ) dx Ga dx ta dx dx ta E1 E2 a

(3.9) After combining Eqs. (3.6)–(3.9): G t d2 𝜎 1 t2 d2 𝜎2 d2 𝜎 1 1 d𝜏 = 1 = − ⇒ t = a 1 Ga dx Ga dx2 Ga dx2 ta dx2

(

𝜎1 𝜎 − 2 E1 E2

) (3.10)

Taking into account Eq. (3.8): 𝜎2 =

t P − 𝜎1 1 t2 bt2

(3.11)

Substituting Eq. (3.11) into Eq. (3.10): d2 𝜎1 − 𝜆2 𝜎 1 + C 0 = 0 dx2

(3.12)

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

where

( ) Ga 1 1 𝜆 = + ta E1 t1 E2 t2 G P C0 = a ta bE2 t2 t1 2

The solution of Eq. (3.12) is of the type: C0 𝜆2

𝜎1 = A cosh(𝜆x) + B sinh(𝜆x) +

(3.13)

With boundary conditions: P P 𝜎1 = 0, 𝜎2 = for x = 0 𝜎1 = 0, 𝜎2 = for x = 0 bt2 bt2 P 𝜎1 = 0, 𝜎2 = for x = 0 bt2 P P 𝜎2 = 0, 𝜎1 = for x = l 𝜎2 = 0, 𝜎1 = for x = l bt1 bt1 P 𝜎2 = 0, 𝜎1 = for x = l bt1

(3.14)

(3.15)

Using Eqs. (3.13), (3.14) and (3.12) to determine the A and B constants: C0 𝜆2 C0 cosh(𝜆l) − 1 P 1 + B= 2 sin(𝜆l) bt1 sinh(𝜆l) 𝜆

A=−

The solution for Eq. (3.12) is: } { C C0 sinh(𝜆x) C0 P + 2 𝜎1 = − 20 cosh(𝜆x) + [cosh(𝜆l) − 1] + 2 wt1 sinh(𝜆l) 𝜆 𝜆 𝜆

(3.16)

From Eq. (3.6): 𝜏 = t1

Ct d𝜎1 = − 02 1 𝜆 sinh(𝜆x) + dx 𝜆

{

C0 t1 P [cosh(𝜆l) − 1] + w 𝜆2

}

𝜆 cosh(𝜆x) sinh(𝜆l) (3.17)

Replacing C0 and 𝜆2 and using 𝜏 =

P : bl

𝜆l 𝜏 = [(𝜑 − 1) cos h(𝜆(l − x)) + cosh(𝜆x)] 𝜏̄ 𝜑 sin h(𝜆l)

(3.18)

where 𝜙=

E1 t1 +1 E2 t2

The Volkersen analysis considers a fully elastic behavior for the adherends and the adhesive. However, in practice, both the adhesive and the adherend can deform plastically. In the case where the adherends undergo plastic deformation, the yielding of the adherends causes an additional deformation of the adhesive, which precipitates its yielding and failure. Adams et al. have proposed a simple methodology to evaluate

81

82

3 Laboratorial Activities and Report Examples

the effect of the overlap length on the performance of bonded single-lap joints, assuming that both the adhesives and the adherends can undergo plastic deformation [4]. In the case where adherends which do not suffer plastic deformation are combined with ductile adhesives, the maximum load supported by the joint can be predicted using the generalized yield of the adhesive criterion. However, if the adherends are also ductile, their deformation cannot be ignored and must be included in this type of analysis. Using the classical beam theory equations, the maximum flexural stress, considering elastic deformation, can be written as 6M (3.19) 𝜎f = 2 f bt Employing the theoretical formulation proposed by Goland and Reissner [5], the bending moment at the edges of the overlap area can be given by kPt (3.20) 2 where k is a factor (0 ≤ k ≤ 1) that includes the effect of joint rotation in the relationship between the bending moment and the applied load, P. The maximum flexural stress can then be written as 3kP 𝜎f = (3.21) bt On the other hand, a normal tension stress is also present on the adherends because of the applied load Mf =

P bt which creates a maximum normal stress, given by 𝜎t =

(3.22)

P(1 + 3k) (3.23) bt By equating 𝜎 n to the yield stress of the adherend, it is possible to determine the maximum load that this can sustain before it yields 𝜎n =

Pmax ́ =

𝜎y bt (1 + 3k)

(3.24)

For reduced loads and small overlap lengths, it is possible to consider k ≈ 1, which leads to 𝜎y bt Pmax (3.25) ́ = 4 However, for larger overlap lengths, the value of k gradually reduces. When the l/t ratio is lower than 20, k ≈ 0, making Pmax ́ = 𝜎y bt

(3.26)

corresponding to the complete plasticization of the adherend section. The authors have applied the model schematically in Figure 3.14 and predicted the strength of

ra l

ize

dy

iel d

ing

Failure load

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

Ge

ne

Adherend yielding

I/t = 20

Overlap length

Figure 3.14 Joint strength prediction method based on the adhesive and adherend yielding. Source: Adapted from Rudawska [2].

steel–steel joints bonded with and epoxy adhesive, considering three different types of steel. A good agreement was reached between predicted and experimental results.

3.2.2

Work Description

Manufacture and determine the strength of ● ● ● ● ● ●

an SLJ, with 25 mm of overlap, bonded with a ductile adhesive; an SLJ with 50 mm of overlap, bonded with a ductile adhesive; an SLJ, with 25 mm of overlap bonded with a brittle adhesive; an SLJ, with 50 mm of overlap bonded with a brittle adhesive. The ductile adhesive to be used is SikaForce L7818 L7, manufactured by Sika. The brittle adhesive to be used is AV138/HV998, manufactured by Huntsman.

The failure load of the bonded joint shall be determined using both the Volkersen model and the generalized yield of the adhesive and the adherend yield criteria.

3.2.3

Materials

3.2.3.1 Adherends

High-strength steel adherends – 2 mm thick (Table 3.6) Table 3.6

High-strength steel properties (DIN 55 Si 7 treated).

E (GPa)

𝝂

𝝈 y (MPa)

210

0.3

1100

83

84

3 Laboratorial Activities and Report Examples

3.2.3.2 Adhesives (Table 3.7) Table 3.7

Adhesive properties.

Young’s modulus, E (MPa) Poisson’s ratio, 𝜐 Shear yield strength, 𝜏 y (MPa)

3.2.4 ● ● ●

●

● ● ● ● ● ● ● ●

L7818 L7

AV138/HV998

2500

4590

0.38 20

0.35 25

Experimental Work

Lightly abrade the bonded area with sandpaper. Clean the bonded area with a paper cloth embedded with acetone. Assemble the adherends in the mold and use spacers to obtain the intended adhesive thickness (0.2 mm) and overlap length (25 and 50 mm). Weight and mix the two components of AV138/H998 adhesive. Use 3.6 g of AV138 (resin) and 1.4g of HV998 (hardener). Mix the two components of the SikaForce 7817 L7 adhesive. Apply the adhesives on the bonded area. Join the adherends and close the mold. Apply the weights over the mold lid. Remove excess adhesive with a file and sandpaper. Mark and reference each specimen. Measure the final adhesive thickness using a caliper. Tensile test the complete joints in a universal testing machine, registering the failure load and the mode of failure.

3.2.5

Report

The report should include the following: ●

● ● ● ●

prediction of the failure loads of the adhesive joints using the Volkersen model, the generalized yield of the adhesive, and the adherend yield criteria; plot with the failure loads and the failure load predictions; images of all the fracture surfaces and comments; comparison between predicted failure loads and experimental failure loads; indication of the better suited prediction method for each case under study and its justification. REPORT 3.2.5.1 Introduction

The effect of adhesive properties on the joint performance can be observed by the stress distribution analysis of a simple lap shear or peel joint. Although joint strength depends on multiple factors, adhesive stiffness is often one of the most important. Generally, joints with flexible adhesives have a more uniform stress

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

Shear stress

distribution and high elongation, while brittle adhesives lead to a joint with high stress concentration at the ends of overlap and lower elongation (Figure 3.15).

Stiff adhesive Flexible adhesive Overlap length

Figure 3.15 Stress distribution along the overlap for brittle and flexible adhesives. Source: Adapted from Adams et al. [1].

Flexible adhesives have better resistance to peel, fatigue, crack propagation, and impact loads but show reduced heat resistance and a lower cohesive strength. Brittle adhesives have lower toughness and higher modulus. These adhesives are more densely cross-linked and are usually employed in structural applications to resist elevated temperatures and aggressive environmental conditions, leading to joints with high stress concentrations. Assemblies bonded with ductile adhesives are less affected by the most important limitations of adhesive joining as considerable ductility is associated with shear response of typical adhesives, which is beneficial in minimizing the effect of shear stress joint strength. In fact, peak stresses at the overlap edges smoothen as a result of adhesive plasticization while stresses are redistributed toward the lightly loaded inner regions of the overlap [6]. The joint strength increases accordingly with the adhesive’s allowable ductility. The design of a bonded joint considering solely the adhesive elastic response deprives the application from a significant amount of additional structural capability. Under normal operation, it is advisable to keep the applied load in the joint low enough to ensure purely elastic response for most practical situations where time-varying loading is encountered. Some damage to the adhesive probably occurs in the ductile regime, which would degrade the long-term response. The benefit of ductile behavior is to provide increased capacity for peak loads and damage tolerance with regard to flaws (e.g. voids and porosity) in the adhesive layer. The effect of the overlap length depends on the type of adhesive and on the adherend yielding. For elastic adherends and ductile adhesives (more than 20% shear strain to failure), the analytical failure criterion to be used is the global yielding of the adhesive. Joint strength is approximately proportional (Continued)

85

86

3 Laboratorial Activities and Report Examples

(Continued) to the overlap length because the adhesive can deform plastically along the whole overlap. For elastic adherends and brittle adhesives, joint strength is not proportional to the overlap length as the stress is concentrated at the ends of the overlap. For adherends that yield, failure occurs when the yield point of the adherend is exceeded [1]. 3.2.5.2 Materials

In order to facilitate the comprehension of the experimental results and justify the values used while applying the theoretical failure criteria, the following tables summarize the properties of the adherends and adhesives used in this practical activity (Tables 3.8 and 3.9). Table 3.8

Mechanical properties of the high-strength steel adherend.

DIN 55 Si7 treated

Young’s modulus (E)

210 GPa

Poisson’s ratio (𝜈)

0.3

Yield strength (σy )

1100 MPa

Table 3.9

Mechanical properties of the applied adhesives.

Adhesive

L7818 L7

AV138/HV998

Young’s modulus (E)

2500 MPa

4590 MPa

Poisson’s ratio (𝜈)

0.38

0.35

Shear yield strength (𝜏 y )

20 MPa

25 MPa

3.2.5.3 Prediction of the Failure Loads

In this section, a prediction of the failure loads of the different samples is made using different analytical criteria. The following schematic representation of a single-lap joint contains the nomenclature that shall be used from this point forward regarding its geometry (Figure 3.16).

b t

ta I

Figure 3.16

Schematic representation of a single-lap joint.

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

3.2.5.3.1 Generalized Yield of the Adhesive The most simplistic analysis of the

stress distribution along a single-overlap joint admits that the adherends are rigid and that the adhesive only deforms because of shear. This assumption leads to a constant shear stress along the entire length of the overlap. Hart-Smith studied the benefits of plastic deformation on adhesives, and when analyzing the behavior of extremely ductile adhesives, we encounter an extreme situation where it is plausible to consider a uniform and simultaneous plastic deformation along the entire length of the overlap. Thus, when working with ductile adhesives, with more than 20% of shear deformation, it is acceptable to estimate the failure load, P max , using Eq. (3.27), as the generalized yield of the adhesive criterion usually results in a good approximation. Pmax (3.27) ⇔ Pmax = 𝜏y b l bl Considering the simplicity of the equation and having all the necessary data in order to estimate the failure load, the prediction of this parameter for each joint follows. Taking into account the enunciated assumptions of this criterion, it is expected that this method will only be successful for joints where the more ductile adhesive is used. The results will only depend on the geometry of the overlap and the adhesive properties. 𝜏y =

Ductile Adhesive – L7818 L7

𝜏y = 20 MPa ● ●

b = 25 mm

For l = 25 mm → P max = 12500 N For l = 50 mm → P max = 25000 N

Brittle Adhesive – AV138/HV998

𝜏y = 25 MPa ● ●

b = 25 mm

For l = 25 mm → P max = 15625 N For l = 50 mm → P max = 31250 N

3.2.5.3.2 Volkersen Model The Volkersen model analyzes the elastic behavior of the adherends and considers interface continuity along the adhesive bond. This method does not account for any deformation in the plastic domain and is thus more suitable for brittle adhesives and adherends. With the application of a continuity constraint, differential straining and shear lag between the adhesive and the adherend occur. Thus, the deformation of the adhesive will not be constant along the overlap length, also resulting in a non-uniform shear stress distribution – higher in the ends and lower in the middle. The failure load is then predicted considering the behavior on the ends,

(Continued)

87

88

3 Laboratorial Activities and Report Examples

(Continued) and it is thus expected that this criterion will return a lower value than the previous method. 𝜆l 𝜏 = [(𝜙 − 1) cosh(𝜆 (l − x)) + cosh(𝜆 x)] 𝜙 sinh(𝜆 l) 𝜏

(3.28)

where 𝜆, 𝜑, and 𝜏 are given by √ ( ) Ga 1 1 𝜆= (3.29) + ta E1 t1 E2 t2 E t (3.30) 𝜙= 1 1 +1 E2 t2 P 𝜏= (3.31) bl Additionally, the shear modulus of the adhesive, Ga , can be obtained as follows: Ea (3.32) Ga = 2 (1 + υ) The equation through which a load, P, can be calculated according to the Volkersen model appears as follows: P=

b 𝜏 𝜙 sinh(𝜆 l) 𝜆 [(𝜙 − 1) cosh(𝜆 (l − x)) + cosh(𝜆 x)]

(3.33)

With all the necessary equations enunciated, it is now possible to estimate the failure load for each of the samples. As the interest relies in the rupture of the joints, 𝜏 = 𝜏 y is considered. Ductile adhesive – L7818 L7

𝜏 = 𝜏y = 20 MPa

b = 25 mm

Ea = 2500 MPa

𝜈 = 0.38

Ga = 905.8 MPa The calculated P max represents the load that can be applied to these joints without failure – which corresponds to the minimum value verified in the previous graphics. ● ●

For l = 25 mm → P max = 6471.5 N For l = 50 mm → P max = 6800.6 N

Brittle adhesive – AV138/HV998

𝜏 = 𝜏y = 25MPa Ga = 1700 MPa

b = 25mm

Ea = 4590MPa

ν = 0.35

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

The calculated P max represents the maximum load that can be applied to these joints without failure – which corresponds to the minimum value of load when observing the graphs in 4. ● ●

For l = 25 mm → P max = 6132.4 N For l = 50 mm → P max = 6212.6 N

3.2.5.3.3 Adherend Yield Both the adhesive and the adherend can deform in

the plastic domain, and not considering this behavior could lead to inaccurate results, especially when dealing with ductile joints. This plastic domain is mainly critic on the adherend because according to Adams et al., the yielding of the adherend aggravates the deformation of the adhesive, causing it to fail prematurely [4]. Nonetheless, this particular criterion determines the failure load according only to the strength of the adherend, being completely dependent on its yielding – which will lead to significant errors if the failure occurs in the adhesive. The failure load can be given by Eq. (3.34). The factor k accounts for the effect of the rotation of the joint in the relation between the applied load and the generated bending moment. For small overlaps, tl < 20, k tends to 1, whereas for bigger overlap lengths, tl ≥ 20, k tends to 0, which translates into a complete plastic deformation of the adherend [4]. 𝜎y b t

(3.34) 1 + 3k In this case, the results will not vary with the properties of the adhesives, but with the properties and geometry of the adherends, as well as with the overlap length. Thus, if we have the same adherend and geometrical parameters, it is a function of adherend material and not of adhesive. Pmax =

High strength steel DIN 55 Si 7 treated

𝜎y = 1100MPa ●

b = 25mm

t = 2mm

For l = 25 mm Pmax = 93.5 N

●

For l = 50 mm Pmax = 55.0 kN

3.2.5.3.4 Prediction Using Joint Designer Software The analytical analysis can be easily achieved using the Joint Designer software (www.jointdesigner.pt) that allows us to effectively predict the failure load of adhesive joints. Its intuitive user interface makes it a quick and simple alternative to conventional analytical calculation. The analytical results obtained using the Joint Designer software were the same as was obtained in previous sections.

(Continued)

89

3 Laboratorial Activities and Report Examples

(Continued) 3.2.5.4 Experimental Results 3.2.5.4.1 Tensile Tests Through the realization of tensile testing of the four

single-lap joints, the obtained values for strength versus displacement appear in Figure 3.17. In this plot, it becomes clear that even though the brittle adhesive (AV138/HV998) has a higher yield strength than the ductile adhesive (L7818 L7), the joints with ductile adhesive present better performance for a 25 mm overlap length and much higher strength for a 50 mm overlap. 30 25

Load (kN)

90

20 Brittle_25mm

15

Brittle_50mm Ductile_25mm

10

Ductile_50mm

5 0

0

0.5

1

1.5

2

2.5

3

Displacement (mm)

Figure 3.17

Tensile tests of the single-overlap joints.

In fact, the impact of the properties of the adhesive on the behavior of the joint is not only measured by its strength but also by its ductility, as it became clear in this case.4 As it is clear in Table 3.10, for these overlap lengths, the joint with a ductile adhesive always presents a better behavior, and the difference becomes higher with its increase. It should be noted that as the brittle adhesive shows almost Table 3.10

Brittle Ductile

Failure load of the different samples. Failure load

Increase

25 mm

7.5 kN

23 %

50 mm

9.2 kN

25 mm

13.3 kN

50 mm

25.1 kN

89 %

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

no evolution with the increase in overlap, the ductile joint presents a close to linear evolution as by doubling its overlap length, the failure load went up 89%. The predicted behavior for the joints with fragile behaviors is based on the stress distribution expected on these. With fragile adhesives, we can observe a stress concentration on the edges of the joints. In other words, the shear yield stress will be superior on these edges because of the fact that the force necessary for the joint to break in these specific locations will be smaller than that on the remaining overlap length. Therefore, an increase on the overlap length will have no considerable practical effects. This matter was already treated on the application of the Volkersen criterion. 3.2.5.4.2 Fracture Surfaces The visual evaluation of the fracture surfaces can

be indicative of the quality of the surface treatment applied. An adhesive failure signal improves joint preparation as there was a disconnection of the adhesive on its interface with the adherend. A theoretically perfect joint should always show cohesive failure of the adherend, but cohesive failure of the adhesive is also completely acceptable as the properties of the adhesive are known. In summary, one looks for damage mechanisms that we can predict, such as adherend or cohesive failure. The cohesive failure is shown in Figure 3.18a and adhesive failure in Figure 3.18b.

(a)

(b)

Figure 3.18

Failure types on adhesive bonding – cohesive (a) and adhesive (b).

SikaForce 7817 L7 25 mm Overlap In this experiment, and as will be verified

later on the remaining samples, we can observe the existence of, at least, a thin film that remains attached to both surfaces (Figure 3.19). Therefore, we can conclude that failure occurred mainly on the core of the ductile adhesive. This is enough to guarantee that failure was cohesive and not adhesive in nature, which shows the good quality of this adhesive joint preparation, even though this is not a stellar example as it is quite close to the interface. Moreover, there were no signs of plastic deformation on the steel adherend throughout the experiment. (Continued)

91

92

3 Laboratorial Activities and Report Examples

(Continued)

(a)

(b)

Figure 3.19 Fracture surface of the SikaForce 7817 L7’s joint with 25 mm of overlap length – side (a) and top (b) view.

SikaForce 7817 L7 50 mm Overlap Once again, although not a perfect example, there is a clear presence of adhesive on both surfaces, which can lead us to assume a correct surface preparation process for this adhesive joint (Figure 3.20). Consequently, we can state that the failure was cohesive and not adhesive.

(a)

(b)

Figure 3.20 Fracture surface of the SikaForce 7817 L7’s joint with 50 mm of overlap length – side view (a) and top view (b).

AV138/HV998 25 mm Now, with the joint being prepared using a different

adhesive (more brittle), with higher values of strength and rigidity, can

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

conclude that there is a difference in coloring on the whole surface of one of the plates on the overlap zone, which means that at least an adhesive film is present on both adherends. Thus, we can state that the failure was cohesive. If the failure had been adhesive, if we looked through the microscope, we would only see the exposed metallic adherend (Figure 3.21). Once again, the steel adherends did not suffer from any kind of substantial plastic deformation.

(a)

(b)

Figure 3.21 Fracture surface of the AV138/HV998’s joint with 25 mm of overlap length: side view (a) and top view (b).

AV138/HV998 50 mm Finally, with a 50 mm overlap length and the brittle adhesive, there are, again, signs of cohesive failure, with the presence of adhesive on both adherends (Figure 3.22).

(a)

(b)

Figure 3.22 Fracture surface of the AV138/HV998’s joint with 50 mm of overlap length: side view (a) and top view (b).

(Continued)

93

94

3 Laboratorial Activities and Report Examples

(Continued) The visual inspection of these samples proved that there was good joint preparation that should improve the value of the obtained results regarding the failure load of these joints and their comparison with the prediction by theoretical models. 3.2.5.5 Discussion

This section provides the analysis of the applicability of the theoretical criteria used for failure load prediction in comparison with the experimental results. Table 3.11 shows the values obtained up to this point for the failure load of every sample. Table 3.11 Failure loads obtained through different analytical criteria and from the experimental results. Ductile

Generalized yielding Volkersen model

Brittle

25 mm

50 mm

25 mm

50 mm

12.5 kN

25 kN

15.6 kN

31.3 kN

6.5 kN

6.8 kN

Adherend yielding

39.5 kN

55 kN

Experimental

13.3 kN

25.1 kN

6.1 kN

6.8 kN

39.5 kN

55.0 kN

7.5 kN

9.2 kN

In summary, as expected, the generalized yield criterion is found to be the most appropriate for joints with ductile adhesives, as the Volkersen model is presented as the best solution with brittle adhesives. We should now look on how to justify these conclusions. As mentioned before, when applying the generalized yield criteria, accurate predictions can be attained for joints with ductile adhesives – something that became clear by observing the plot in Figure 3.23, which demonstrates that, when considering a rigid adherend with mainly shear deformation of the adhesive is a good approximation to this case. The error produced is quite small and incomparable with the one obtained through other methods. This criterion clearly follows the linear evolution of the strength with the increase in overlap length, as it was clear by its formulation. The fact that the predictions coming from the Volkersen model are low compared with the experimental results is also in accordance with the expectations. By considering a stress concentration on the edges, we have a joint more susceptible to failure than with a uniform distribution of load along its overlap length.

3.2 Effect of Adhesive Type and Overlap Length on the Failure Load of Adhesively Bonded Joints

60

Failure load (kN)

50 40 30 20 10 0

0

10

20

30

40

50

Overlap length (mm) Generalised yielding criteria

Volkersen model

Adherend yielding

Experimental failure load

Figure 3.23 criteria.

Failure load for joints with ductile adhesives according to different

As Figure 3.23 shows, by increasing the overlap length, the error associated with utilizing the Volkersen model increases significantly as there is no consideration of the influence of the entire overlap length in the strengthening of the joint, continuing to consider a high stress concentration on the joint’s edges. The assumptions that go along with this criterion result in a limited rate of improvement in the failure load with a higher overlap. By applying the Adherend Yield Criteria, we incur in a very significant error, for both brittle and ductile adhesives, as it becomes clear in the plots in Figure 3.23 and Figure 3.24, with the sole exception of the joint with a ductile adhesive and a 25 mm overlap. This criterion relies heavily on the plastic deformation of the adherends, meaning that it is highly dependent on their yield strength. The mechanical properties of steel that composes the adherends are vastly higher than the ones of the adhesives, making it inevitable that the results of this criterion are far from those obtained experimentally. For this reason, it is our understanding that the good approximation in the mentioned sole case must be the resultant of pure chance, as there is no logical correlation for it. The high strength steel that composes the adherends presents a yield strength (𝜎 y = 1100 MPa) that is many times higher than the shear yield strength of the adhesives (𝜏 y = 20 MPa and 25 MPa), which justifies the error associated with the application of this method. Considering the known properties of the components of this joint, it is expected that the failure will occur on the adhesive before there is any significant plastic deformation of the adherends. (Continued)

95

3 Laboratorial Activities and Report Examples

(Continued) 60 50 Failure load (kN)

96

40 30 20 10 0

0

10

20

30

40

50

Overlap length (mm) Generalised yielding criteria

Volkersen model

Adherend yielding

Experimental failure load

Figure 3.24

Failure load for joints with brittle adhesives according to different criteria.

Figure 3.24 shows that, while utilizing adhesives of brittle nature, the Volkersen model is, by far, the most appropriate for the prediction of the joint failure load, at least among the studied criteria. The assumptions made around the generalized yield criteria and the Volkersen model in the analysis of the results with ductile adhesives translate in the opposite direction to their application with ductile adhesives, meaning that we are now in the condition where the Volkersen model is the most appropriate for the exact same reasons that it was not in the previous case. The increase of the failure load experimentally verified was slightly higher than expected, increasing the prediction error with the higher overlap length. In spite of that, the Volkersen model still presents the best approximation, by far, with errors that vary from 18% to 26.3%. Besides that, the experimental values confirm that there is not a linear increase in strength with the evolution of the overlap length, as commented before and in accordance with the initial expectations. In this case, considering that the entire overlap length is under a constant shear stress and contributes evenly to the joint’s strength leads to predictions made by the generalized yield criteria that are much higher than what was experimentally found. In line with the formulation of this criterion, with the increase in overlap length, there is a larger error in the failure load prediction – going from 52% to 71% – because, as referred to, brittle adhesives in joints show a clear stabilization of their failure load because of the larger influence of the edges in its mechanical behavior, something that is not compatible with the assumptions surrounding the generalized yielding criteria.

3.3 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints

3.2.5.6 Conclusions

In this work, regarding the influence of the overlap length, it became clear that, especially for ductile adhesives, higher overlaps usually lead to a greater joint strength. As seen up to this point, when using brittle adhesives, this evolution is not linear – opposite to what happens with the ductile ones – and it even becomes slightly constant as the overlap grows larger. With brittle adhesives, the stagnant behavior that occurs at a certain overlap length needs to be taken into account in joint design as the waste in the added material and weight might not be worth the very limited increase in performance. Considering both types of adhesives utilized, the differences were quite clear. The theoretical models that produced predictions for the failure load were quite successful if applied to joints that match their assumptions. In the case of brittle adhesive, the prediction that comes closest to the experimental result is the one provided by Volkersen’s criterion. This is due to the properties of the adhesive, whose brittleness (low maximum shear deformation) leads to an uneven distribution of stresses along the joint. As the adhesive presents a low plastic deformation during loading, there is a concentration of shear stresses at the extremes of the bonded area and in the middle there is very little (or almost no) stress. For joints bonded with a ductile adhesive, its higher maximum shear deformation leads to the adhesive being equally loaded, in which case, the generalized creep criterion proves to be the most accurate. For the ductile adhesive, there is only a small difference between the value estimated by the generalized yield criterion and the value obtained experimentally. Besides that, it became clear that the adhesive’s strength does not translate inevitably into a stronger joint, as its ductility is also crucial to the mechanical behavior of the joint. As can be seen in this case, even though the ductile adhesive had a lower strength, it produced better performing joints because of its ductility. Although it has difficulty to ascertain initially, joint strength does in fact rely heavily on strength, stiffness, and ductility of the adhesive, all of these being key parameters that must be understood to enable higher failure loads.

3.3 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints 3.3.1

Introduction

Adhesive thickness is one of the geometrical characteristics of the joint that must be considered during the design phase. The optimal thickness for the adhesive layer (around 0.1–0.5 mm) must be guaranteed to ensure good joint performance. For most applications, adhesive manufacturers recommend a thickness of 0.1–0.2 mm so that maximum strength can be obtained. The reasons behind the decrease in joint

97

98

3 Laboratorial Activities and Report Examples

strength for layers thinner than 0.1 mm are not yet fully understood. It is believed that extremely thin layers increase the risk of bonding failure and lead to weakening defects. On the other hand, it has been experimentally verified that joint strength is also reduced if thicknesses above 0.1 or 0.2 mm are used. Analytical or numerical models, based purely on elastic analysis, predict a reduction in the normal and shear stresses as the thickness increases [3–5]. This results in the estimation of higher failure loads, something that is not confirmed with experimental testing. Crocombe used the global yield criterion to demonstrate that in single-lap joints, the complete yield of the adhesive layer occurs prematurely in thicker joints [7]. The author concluded that a non-linear analysis is therefore fundamental to understand the influence of the adhesive layer. Adams and Peppiatt have forwarded three possible explanations for the reduction in joint strength with increasing thickness: higher stress concentrations, higher probability of internal defects (voids, porosities, and micro-cracks due to thermal stresses introduced during the curing process), and dissimilar strain rates [8]. The authors then concluded that the presence of internal defects is the most important of these factors. More recently, Grant et al. demonstrated that the effect of thickness could also be explained by the increased bending moment, for the case of joints using soft steel as the adherend [9]. For the cases where the adherends undergo plastic deformation, the adherend yielding causes an additional adhesive deformation that ultimately leads to its failure. The Volkersen analysis assumes a fully elastic behavior of the adherends. However, both the adhesives and the adherends of a real joint can suffer plastic deformation. Modern structural adhesives such as toughened epoxies can exhibit large plastic deformation before failure. The adherends can also suffer deformation. Hart-Smith [10] improved on Volkersen theory by also taking into account the plastic behavior of the adhesive. The adhesive is modeled as being an elastoplastic material, with the area below the shear stress curve equivalent to the area of the real curve and the same stress and strain at failure (see Figure 3.25). Hart-Smith has shown that the adhesive plasticity increases the joint resistance when compared to a purely elastic analysis. A ductile adhesive can yield plastically, and this allows an additional load to be supported before the failure strain is encountered. This is shown in Figure 3.25. It is therefore generally preferable to have a joint with a ductile adhesive such as a modified epoxy. Not only the adhesive is stronger but it also fails in a safer manner, deforming plastically before the joint fails, redistributing and reducing peaks in the shear stress distribution.

3.3.2

Work Description

Manufacture and determine the strength of ● ● ●

a single-lap joint with a thickness of 0.2 mm; a single-lap joint with a thickness of 1 mm; a single-lap joint with a thickness of 2 mm;

The same adhesive will be used for all the joints (SikaForce®-7818 L7). The failure load will be predicted using the Volkersen analytical model, the adhesive generalized yield, the adherend yield criterion, and the Hart-Smith criterion.

3.3 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints

P P

Increases in strength

C

Shear stress

Yield stress

B C

B

A

Yield strain

A

C

Shear strain

Increases in strength

B A

Figure 3.25 Schematic explanation of the plastic shear on the adhesive. Source: Adapted from Hart-Smith [10].

3.3.3

Materials

3.3.3.1 Adherends:

Mild steel (DIN St33 with 𝜎 y = 185 MPa) – 2 mm thick 3.3.3.2 Adhesives:

3.3.4 ● ● ●

● ● ● ● ● ● ● ●

Experimental Work

Lightly abrade the bonded area with sandpaper. Clean the bonded area with a paper cloth embedded with acetone. Place the adherends in a mold and adjust the spacers to obtain thicknesses of 2, 1, and 0.2 mm with an overlap length of 25 mm. Mix the two components of the SikaForce 7817 L7 adhesive. Apply the adhesives on the bonded area. Join the adherends and close the mold. Apply the weights over the mold lid. Remove excess adhesive with a file and sandpaper. Mark and reference each specimen. Measure the final adhesive thickness using a caliper. Tensile test the complete joints in a universal testing machine, registering the failure load and the mode of failure.

3.3.5

Report

The report should include the following: ●

Prediction of the failure loads of the adhesive joints using the Volkersen analytical model, Hart Smith analytical model, the adhesive-generalized yield, and the adherend yield criteria;

99

100

3 Laboratorial Activities and Report Examples ● ● ● ●

Plot with the failure loads and the failure load predictions; Images of all the fracture surfaces and comments; Comparison between predicted failure loads and experimental failure loads; Indication of the better suited prediction method for each case under study and its justification. REPORT 3.3.5.1 Introduction

The thickness of the adhesive layer is an important factor for joint strength. Adhesive thickness is controlled by the pressure applied, the viscosity of the adhesive, and the presence of scrim or carrier layers in the case of film adhesives [4]. An optimum thickness for the adhesive should be ensured to obtain the best performance of the joint. The most widely used and recommended thickness by manufacturers is between 0.1 and 0.2 mm. It appears that the experimental strength of a joint decreases with increasing thickness of the adhesive from 0.1 to 0.2 mm. For thickness below 0.1 mm, there is also a sudden drop of the joint strength probably because of a crack of adhesive. Common practice involves the use of film adhesives containing scrim cloth, some forms of which help to maintain bond thicknesses. It is also common practice to use mat carriers of chopped fibers to prevent a direct path for access by moisture to the interior of the bond [1]. The decrease of joint strength by increasing the adhesive thickness can be explained by the following: ●

●

●

For high adhesive thicknesses, there is the risk of introducing defects, such as air bubbles and micro-cracks; As the thickness of the adhesive increases, the bending moment increases at the end of the overlap, resulting in a decrease of joint strength; The interface stresses (peel and shear) increases with increasing bondline thickness and the failure load of a bonded joint decrease.

3.3.5.2 Experimental Details 3.3.5.2.1 Adherend Table 3.12 shows the mechanical properties of the high

strength steel used in this study. Table 3.12

Properties of treated DIN St33 steel.

E (GPa)

𝝂

𝝈 y (MPa)

210

0.3

1100

3.3 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints

3.3.5.2.2 Adhesive Table 3.13 shows the properties of the adhesive used. The

shear modulus was obtained using Eq. (3.35). E = 2G (1 + ϑ) Table 3.13

(3.35)

Properties of the adhesive BETAMATETM 2098. BETAMATETM 2098

Young’s modulus, E (GPa)

0.93

Shear modulus, G (GPa)

0.34

Shear strength, 𝜏 y (MPa)

18.1

Maximum shear deformation, 𝛾 f (%)

56

Poisson’s coefficient, 𝜈 (−)

0.35

3.3.5.2.3 Geometry The single-overlap joints were manufactured with the following dimensions: ● ● ● ●

Adherend width: 25 mm Adherend thickness: 2 mm Adhesive thickness: 0.2 and 1 mm Overlap length: 25 mm

3.3.5.3 Prediction

Predictions were made regarding the force required for joint failure by the Volkersen, generalized yielding, and adherend yielding criteria. The results of these predictions are presented in Table 3.14. An additional criterion was also developed starting from the adherend yielding criterion. Knowing that the bending stress in the adherend is given by Eq. (3.36) 𝜎f =

6 Mf

(3.36)

b t2

Table 3.14 Predictions of joint failure forces according to the generalized yielding, Volkersen, and adherend yielding. Criteria

Thickness of 0.2 mm

Generalized yielding criteria (kN)

Thickness of 1 mm

11.3

11.3

Volkersen model (kN)

8.1

10.4

Adherend yielding (kN)

4.4

4.4

(Continued)

101

3 Laboratorial Activities and Report Examples

(Continued) The normal stress due to tensile load can be obtained using Eq. (3.37). P (3.37) bt The maximum normal stress can be obtained by adding these two components. In the adherend yielding criterion, the bending moment given by the method of Goland and Reissner is considered. In the proposed criterion, the bending moment determined by Grant et al. [1] was used to determine the normal stress because of bending to which the adherend is subjected. Thus, it was possible to take into account the effect of increasing thickness, which causes an increase in the bending moment (Mf ); this increases correspondingly to a decrease in joint strength. Table 3.14 presents the predictions of the failure forces of the joint, taking into account several criteria. 𝜎t =

3.3.5.4 Experimental Results

The tensile test of the two joints was performed, and the following values were obtained at failure (Table 3.15). In Figure 3.26, the predictions were compared with the experimental results. Table 3.15

Experimental results for adhesive bonds of 0.2 mm and 1 mm thickness. Thickness of 0.2 mm

Thickness of 1 mm

Failure load (kN)

9.5

8.9

Displacement at failure (mm)

4.3

3.2

12 10 Failure load (kN)

102

8 6 4 2 0

0

0.2

0.4

0.6

0.8

1

1.2

Adhesive thickness (mm) Generalised yielding criteria [kN]

Experimental failure load [kN]

Volkersen model [kN]

Adherend yielding [kN]

Figure 3.26

Comparison of experimental and predicted failure load.

3.3 Effect of Adhesive Thickness on the Failure Load of Adhesively Bonded Joints

3.3.5.5 Failure Surfaces

To better understand the effect of adhesive thickness, the failure mechanism was evaluated. In the case of the 0.2 mm joint, the rupture was cohesive with an adhesive layer on both adherends (Figure 3.27a). In the case of 1 mm thick joint, there seems to be adhesive breakage at the edges of the joint (Figure 3.28a). This may be due to poor preparation, together with the pull-out caused by the plastic deformation of the adherends (Figure 3.28b), which is much greater than that observed in the 0.2 mm joint (Figure 3.27b). The higher deformation in the 1 mm joint can be explained by the higher bending moment because of the additional adhesive thickness.

(a)

Figure 3.27 0.2 mm.

(b)

Failure surface (a) and adherend yielding (b) of adhesive joints with

(b)

(a)

Figure 3.28

Failure surface (a) and adherend yielding (b) of adhesive joints with 1 mm.

3.3.5.6 Conclusion

For the 0.2 mm-thick joint, the criterion that comes closest to the experimental result is the Volkersen method. However, this should be seen as a coincidence as the Volkersen criteria only consider elastic deformations in the adhesive, when in reality, a ductile adhesive with high shear deformation is being used. One would expect the adherend yield strength criterion to be the closest to reality, as (Continued)

103

104

3 Laboratorial Activities and Report Examples

(Continued) a soft metal was used together with a high-strength ductile adhesive. However, the adherend yielding criterion does not effectively reflect the maximum value of joint resistance and predicts an extremely low resistance because it does not take into account the hardening of the test metal, the value corresponding to the beginning of plasticization. In the case of a 1 mm-thick joint, the Volkersen criterion predicts an increase in joint resistance, which does not occur. The generalized yielding criterion does not take into account the thickness of the adhesive, so it was predictable that it would not be a correct approximation. The adherend yielding criterion provides a lower prediction of the breaking strength than the experimental one as the equation assumes that after plastic deformation of the adherends occurs, the strength remains constant. This fact does not occur in reality because after exceeding the yield stress and with increasing applied load, the steel undergoes strain hardening. By taking into account the thickness of the adhesive, the proposed criterion should be a good approximation. The additional strength of the adhesive may be due to the fact that the additional thickness allows accommodating a greater normal deformation before rupture (due to peel stress). If finite element simulation of the joint is performed and the bending moment of the joint at failure is obtained, the design can be more accurate.

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints 3.4.1

Introduction

Structural adhesives are the most efficient method to join composite materials. In the case of composite laminates, the peel stresses can cause interlaminar failure of the composite adherend by means of localized delamination near the singularity point (see Figure 3.29). Therefore, the design of joints containing composites must not only be concerned with adhesive failure but also with the transversal failure (in the direction of the thickness) of the composite. The Goland and Reissner model allows the design of an adhesive joint taking into account these two failure modes [5]. The adhesive failure criterion can be based on a maximum value of stress (for fragile adhesives) or global yielding of the adhesive (for ductile adhesives). The adherend failure criterion is based on the peel stress on the adherend near the interface, assuming that there is continuity in the peel stress between the adhesive and the adherend. For highly ductile adhesives (those with more than 20% of shear strain), all overlap length undergoes plastic deformation (generalized yield criteria) and the failure load (Pmax ) can be estimated by simply considering the adhesive’s yield strength (𝜏 y ): Pmax ́ = 𝜏y bl

(3.38)

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints

Stress concentration

Transversal rupture (through thickness) of the composite

Figure 3.29

3.4.2

Interlaminar failure of composite adherends.

Work Description

Manufacture and determine the strength of ● ●

an SLJ, with 12.5 mm of overlap; an SLJ, with 25 mm of overlap.

A ductile adhesive to be used is SikaForce L7818 L7, manufactured by Sika. The failure load will be predicted using the Goland and Reissner analytical model and the adhesive generalized yield.

3.4.3

Materials

3.4.3.1 Adherends:

High-strength steel adherends – 2 mm thick (Table 3.16) Table 3.16

Properties of the glass fiber composite.

E 11 (GPa)

E 22 (GPa)

𝝂 12

𝝂 23

𝝈 22 , 𝝈 33 (MPa)

84

7

0.3

0.02

65

3.4.3.2 Adhesives (Table 3.17) Table 3.17 L7.

Adhesive properties of SikaForce®-7818

E (MPa)

𝝂

𝝉 y (MPa)

𝜺r (%)

2500

0.35

18.0

2.5

3.4.4

Experimental Work

● ●

Lightly abrade the bonded area with sandpaper. Clean the bonded area with a paper cloth embedded with acetone.

105

106

3 Laboratorial Activities and Report Examples ●

● ● ● ● ● ● ● ●

Assemble the adherends in the mold and use spacers to obtain the intended adhesive thickness (0.2 mm) and overlap length (12.5 and 25 mm). Mix the two components of the SikaForce 7817 L7 adhesive. Apply the adhesives on the bonded area. Join the adherends and close the mold. Apply the weights over the mold lid. Remove excess adhesive with a file and sandpaper. Mark and reference each specimen. Measure the final adhesive thickness using a caliper. Tensile test the complete joints in a universal testing machine, registering the failure load and the mode of failure.

3.4.5

Report

The report should include the following: ●

● ● ● ●

Prediction of the failure loads of the adhesive joints using the Goland and Reissner and the generalized yield of the adhesive criteria. Plot with the failure loads and the failure load predictions. Images of all the fracture surfaces and comments. Comparison between predicted failure loads and experimental failure loads. Indication of the better suited prediction method for each case under study and its justification. REPORT 3.4.5.1 Introduction

The properties of the adherends are known to have significant influence on the stress distributions in bonded joints. Non-uniform stress distributions along the overlap in adhesive layers is mainly caused by the relative displacement of the adherends because of the strain in the adherends, and thus, the Young’s modulus and the adherend thickness of each adherend are important factors in the shear stress distribution. Adherend yielding can cause a premature failure of the joint when the stress reaches the yield point of the adherend, creating a plastic hinge at the edge of the overlap. For composite adherends, it is recommended to have the outer layers oriented in a direction parallel to the loading direction to avoid intralaminar failure of these layers. In fact, the major problem of composite laminate adherends is its low transverse strength (through the thickness direction). The transverse strength of composites is of the same order or lower than the adhesive tensile strength. Adhesive joints with composites tend to fail in an interlaminar manner because of the high peel stresses at the edge of the overlap (Figure 3.30).

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints

Figure 3.30

Failure of the composite in adhesive joints.

Adhesive joints with composites have lower interlaminar shear stiffness and shear strength than metals. Deformations in an adhesive joint with a composite laminate and a metallic adherend under tension loading are shown in Figure 3.30. Failure tends to initiate in the composite at the ends of overlap. 3.4.5.2 Characterization of the Tested Joints

The tested adhesive joints were SLJs, such as the one schematically represented in Figure 3.31:

b t

ta I

Figure 3.31 Single-overlap joint (b – adherend width; t a – adherend thickness; l – overlap length; and t a – adhesive thickness).

The dimensions of the joints tested in the two cases only varied in the overlap length (l), whose value was 12.5 and 25 mm for each of the tested adhesive joints. Table 3.18 shows the dimensions of the adhesive joints used. Table 3.18

Dimensions of the tested specimens.

Dimension

Value (mm)

Adherend width (b)

25

Adhesive thickness (t a )

0.2

Adherend thickness (t)

2

(Continued)

107

108

3 Laboratorial Activities and Report Examples

(Continued) The material constituting the adherends of the adhesive joint is a fiberglass composite, which is an anisotropic material, whose mechanical properties are found in Table 3.19. It should be noted that as we are in the presence of an anisotropic material, its properties vary with the direction and thus are presented as a function of the main directions: direction 1 represents the direction of the fibers, direction 2 is transversal to the first direction, and, finally, direction 3 is normal to the first. Table 3.19

Mechanical properties of fiberglass composite.

Property

Value

Young’s modulus (E 11 )

84 GPa

Young’s modulus (E22 )

7.0 GPa

Poisson’s coefficient (𝜈 12 )

0.3

Poisson’s coefficient (𝜈 23 )

0.02

Ultimate strength (𝜎 22 /𝜎 33 )

65 MPa

As for the adhesive, SikaForce −7818 L7, from the manufacturer SIKA, was used, whose properties are found in Table 3.20. Table 3.20

Mechanical properties of SikaForce®-7818 L7 adhesive.

Property

Value

Young’s modulus (E)

2.5 GPa

Poisson’s coefficient (𝜈)

0.35

Shear yield stress (𝜏 y )

18.0 MPa

Extension after failure (𝜀r )

2.5 %

3.4.5.2.1 Experimental Results Force–Displacement Curves

– 12.5 mm overlap length The failure load obtained for the joints with an overlap length of 12.5 mm was 6.4 kN and at a displacement of 2.2 mm (Figure 3.32). – 25mm overlap length Figure 3.33 shows the failure load obtained for the joints with an overlap length of 25 mm that was 9.5 kN, occurring at a displacement of 4 mm.

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints

7

Load (kN)

6 5 4 3 2 1 0 0

0.5

1

1.5

2

2.5

Displacement (mm)

Figure 3.32

Load–displacement curve for 12.5 mm overlap length.

10

Load (kN)

8 6 4 2 0

0

1

2

3

4

5

Displacement (mm)

Figure 3.33

Load–displacement curve for 25 mm overlap length.

3.4.5.2.2 Failure Surface

– 12.5 mm overlap length In this case, the failure was found to have occurred in the adhesive layer, thus corresponding to cohesive failure in the adhesive (Figure 3.34). – 25 mm overlap length For the larger overlap length, failure occurred within the adherend, and this mechanism is designed as cohesive failure in the adherend (Figure 3.35). More specifically, this failure took the form of delamination. (Continued)

109

110

3 Laboratorial Activities and Report Examples

(Continued)

Figure 3.34

Joint with an overlap length of 12.5 mm, after failure.

Figure 3.35

Joint with an overlap length of 25 mm, after failure.

Analysis of the Results Obtained The experimentally obtained force–displacement curves generally show the behavior of a brittle polymeric material. In the sample with an overlap length of 12.5 mm, the failure was cohesive, which means that failure occurred because of the failure of the adhesive. This information can be confirmed by observing the presence of adhesive on both sides of the specimen. In the sample with an overlap length of 25 mm, on the other hand, failure already occurs because of the failure of the adherend, more precisely delamination. In Figure 3.36, it can be seen that the fibers of the composite are exposed because of the breakage thereof. In this second case, it is not the adhesive that controls the strength of the joint but the adherend itself.

3.4.5.2.3

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints

12.0

Failure load (kN)

10.0 8.0 6.0 4.0 2.0 0.0

0

5

10

15

20

25

30

Overlap length (mm)

Figure 3.36

Failure load as a function of overlap length.

Of the two, the most desirable situation is the one where failure occurs in the adherend, such as for 25 mm overlap length, because it means that the joint no longer fails because of the adhesive. However, delamination failure is still relatively undesirable as it occurs at a generally low load because if the low transverse strength of the composite. The analysis of the influence of the overlap length can be carried out with the aid of Figure 3.36, which shows the variation of the failure load as a function of the overlap length of the joint. For an overlap length of 12.5 mm, the experimentally verified failure load was 6398 N. Using a joint with the same adhesive, the same adherend, and double the area, it could be expected that the failure load would be twice the previous one, which is not the case, being only 9502 N. The behavior described, in reality, could not be verified because, in this second joint, the joint failure is not a result of the failure of the adhesive but of its adherend. As it can be seen from the above graph, the failure load varies non-linearly with the overlap length; this behavior is due to the occurrence of adhesive delamination from a certain overlap length. There is then an increase in the failure load with the increase in the resistant area of the adhesive (consequence of the increase in the overlap length), keeping this relatively constant from the point at which the adherend delamination occurs. This behavior is in accordance with what is theoretically predicted and represented in the graph of Figure 3.37. From this observation, it can also be concluded that an increase in the overlap length of the joint from 25 mm is not advantageous, as this does not translate into an increase in its resistance. (Continued)

111

3 Laboratorial Activities and Report Examples

(Continued) Figure 3.37 Failure load as a function of overlap length.

Joint strength

112

Overlap length

3.4.5.3 Theoretical Prediction of Failure Load

There are several analytical models suitable for the analysis of stresses in an adhesive joint, such as those of Volkersen, Goland and Reissner, or Hart-Smith, all of which have different assumptions, so the choice of the most appropriate model should be made taking into account the specificities of the situation under analysis. The same applies to the choice of failure criterion, and for this aspect, it is extremely important to understand if we are in the presence of a ductile or fragile adhesive [4]. In this work, the prediction of failure load is performed using the model of Goland and Reissner and the generalized yield criterion [11]. Goland and Reissner Model The Goland and Reissner model, unlike the Volkersen model – which only accounts for the shear forces in the adhesive joint – takes into account the rotation of the adhesive joint because of the bending moment introduced by the misalignment of the traction forces acting on the adhesive joint through the introduction of a factor k that allows to relate the bending moment with the applied load, see Eq. (3.39) [1, 4].

3.4.5.3.1

P⋅t (3.39) 2 This k factor will present values between 0 and 1, being approximately 1 when we are in the presence of sufficiently low loads. There is, then, the consideration not only of the shear stresses but also of the peel forces because of bending. In Figure 3.38, this situation is schematically represented and the resulting peel stress distribution is also shown. As it can be seen from the analysis of Figure 3.38, there is a greater concentration of stresses at the ends of the adhesive joint, which is why this is the area where the adherend failure is expected to occur. It should be noted that this model does not take into account the plasticity of the materials, being therefore a purely elastic model. Using this model as a failure criterion, the specimen failure occurs when the maximum peel stress at the ends of the joint exceeds Mf = k

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints

the value of the failure stress in the direction normal to the fibers (𝜎 33 ) of the material of the adherend. It is therefore assumed that the stress values obtained through this model will be the same for the adhesive and the adherends, which is a very reasonable approximation. Unloaded

Stress

Loaded

Shear stress Peel stress

0

Overlap length

Figure 3.38 Distribution of peel stress in the adhesive joint determined by the model of Goland and Reissner.

3.4.5.3.2 Failure Load Prediction The calculation of the failure load by the

model of Goland and Reissner was carried out using a software developed for this purpose – Jointdesigner (www.jointdesigner.pt). Introducing in the above-mentioned program all the data necessary to obtain the values of peel stress, namely, the behavior of the adhesive and adherends (which is considered elastic, according to this model), the dimensions of the tested specimens, and the properties of those already presented for the materials that compose them, the value of applied load was then determined, which would result in a peel stress at the ends of the adhesive joint equal, in the limit, to the failure stress of the adherend (65 MPa). Table 3.21 presents the calculation carried out for the two different overlap lengths, and the following values were obtained for the failure load. Table 3.21 Failure load predicted by the Goland and Reissner model for different overlap lengths. Overlap length (mm)

Failure load (kN)

12.5

6.1

25

8.1

(Continued)

113

3 Laboratorial Activities and Report Examples

(Continued) The graphs obtained for the distribution of the peel stresses along the overlap length of the joint for each of the above cases are also presented in Figure 3.39. 70 60

Peel stress (MPa)

50 40 30 20 10 0 –10 –20 –7

–6

–5

–4

–3

–2

–1

0

1

2

3

4

5

6

7

Overlap length (mm)

Figure 3.39

Peel stress along the overlap length (l = 12.5 mm).

Generalized Yield Criterion The generalized yield criterion, initially proposed by Crocombe (1989), is based on a simple analysis of the shear stress distribution, considering that it is uniform along the entire overlap length [1]. This failure criterion is applicable to ductile adhesives and accounts for their plasticity, assuming that the failure of the adhesive joint only occurs after the complete yielding of the adhesive layer. Thus, the maximum load that the adhesive joint can withstand will correspond to the total plasticization of the adhesive, in Eq. (3.40) (Figure 3.40).

3.4.5.3.3

Pmax = 𝜏y ⋅ b ⋅ l

(3.40)

70 60 50 Peel stress (MPa)

114

40 30 20 10 0 –10 –20 –12.5

–7.5

–2.5

2.5

7.5

Overlap length (mm)

Figure 3.40

Peel stress along the overlap length (l = 25 mm).

12.5

3.4 Effect of Overlap Length on the Strength and Failure Mechanism of Composite Adhesive Joints

3.4.5.3.4 Failure Load Prediction Using the previous expression to calculate the failure load predicted by the generalized yielding criterion (P max ) and having the values of the necessary parameters (𝜏 y , b, and l), the failure load predictions were obtained for each overlap length in Table 3.22. Table 3.22 Failure load prediction by the generalized yield criterion for different overlap lengths. Overlay length (mm)

Failure load (kN)

12.5

5.63

25

11.25

3.4.5.4 Comparison with Experimental Results

In order to analyze and compare the experimental values obtained for the failure load of the two adhesive joints tested against the predicted values, the plot shown in Figure 3.41 was prepared, which represents the failure load (both predicted and experimental) as a function of the overlap length of the adhesive joints. 14.0

Failure load (kN)

12.0 10.0 8.0 6.0 4.0 2.0 0.0

0

5

10

15

20

25

30

Overlap length (mm) Goland and reissner

Figure 3.41

Generalized yield criterion

Experimental

Predicted and experimental failure load as a function of overlap length.

The relative error of the value of the failure load predicted by each model for the two test pieces was also calculated (Table 3.23). Table 3.23

Relative error of the breaking force predicted by the different models.

Overlap length (mm)

12.5

Model

Goland and Reissner Generalized yield criterion

25

Relative error (%)

4.4 12.1

Goland and Reissner

14.6

Generalized yield criterion

18.4

(Continued)

115

116

3 Laboratorial Activities and Report Examples

(Continued) Table 3.23 allows us to immediately conclude that the best approximation of the theoretically predicted failure loads of experimental joints occurs for the situation of lower overlap length (l = 12.5 mm) and for the model of Goland and Reissner, presenting an error of 4.4% in relation to the experimentally obtained values. However, this model no longer predicts the failure load so accurately when there is an increase in the joint overlap length, presenting an error of 14.6% in this case. It should be noted that as in the case of the overlap length of 12.5 mm there was a cohesive failure of the adhesive, the generalized yield criterion should present a value for failure load closer to the real one than that predicted by the model of Goland and Reissner as the failure is not caused by excessive peel stresses (above that which the adherend can withstand), but instead by yielding of the adhesive. In fact, although the generalized yield criterion is more suitable for ductile adhesives, even if a brittle adhesive is used, the short overlap length and the limited ductility of the adhesive allows the adhesive to plasticize completely before breaking. However, this was not the case for the test being performed as the failure load predicted by this criterion presents an error of 12.1% in relation to the experimental one, substantially higher than that given by Goland and Reissner model. Analyzing the results obtained for an overlap length of 25 mm, it can be seen that the generalized yield criterion predicts failure load that is significantly higher than that found experimentally – about 1.75 kN higher – with a relative error of 18.4%. From this, it can be concluded that the criterion of generalized yielding is not applicable to the latter case. This result was expectable as this criterion does not account for the possibility of adherend delamination. This criterion predicts a linear increase of the breaking force with an increase of the overlap length of the joint, which in fact does not take place from the moment when adherend delamination occurs as failure will occur always before complete plasticization of the adhesive. Despite presenting an error that is also relatively high for the case of the 25 mm overlap length, the model of Goland and Reissner is still the most suitable for this situation as it allows for predicting, albeit conservatively, the failure of the adherend in a direction normal to the composite fibers. It is for this reason that the Goland and Reissner model is one of the models that is better suited for the analysis of adhesive joints bonded with brittle adhesives. 3.4.5.5 Conclusions

By carrying out this work, two major conclusions were drawn in relation to the variation in the behavior of an adhesive joint loaded under tension. The first is that, as the overlap length increases, the strength of the joint also increases, ceasing to fail because of the failure of the adhesive and

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

starting to fail instead because of the failure of the adherend. The second situation, observed for an overlap length of 25 mm, constitutes a more advantageous solution for the manufacture of structural joints, as in addition to greater strength, it reveals that the joint is not in fact the weak point of the structure. Secondly, it is concluded that the predicted failure loads, determined according to the Goland Reissner method, present values closer to reality, when compared with those obtained by the generalized yielding method. This result was expected for the joint with an overlap length of 25 mm, but not for a joint of 12.5 mm, which suffered cohesive failure that should have been better approximated by the generalized yielding method. In this specific case, this did not occur, most likely because of an inadequate curing time.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling 3.5.1

Introduction

Modeling a bonded joint using finite element analysis is nowadays a very popular procedure in the automotive industry as it allows for a very agile design process and can include highly complex material behavior. Furthermore, these models can also include cohesive zone modeling, whereupon a model will be able to precisely reproduce the failure mechanism of a bonded joint and determine its failure load.

3.5.2

Work Description

Modeling of an adhesive joint using cohesive zone modeling of ● ● ● ●

an SLJ, with 25 mm of overlap an SLJ, with 25 mm of width adhesive thickness of 0.2 mm a brittle adhesive

3.5.3

Materials

3.5.3.1 Adherends:

High-strength steel adherends – 2 mm thick (Table 3.24) Table 3.24

Properties of the high-strength steel.

E (GPa)

𝝂

𝝈 y (MPa)

210

0.3

1200

117

118

3 Laboratorial Activities and Report Examples

3.5.3.2 Adhesives (Table 3.25) Table 3.25 E (MPa)

G (MPa)

𝝈 y (MPa)

𝝉 y (MPa)

G I (N/mm2 )

G II (N/mm2 )

4890

1560

41.0

30.2

0.35

4.91

3.5.4 ● ● ● ● ● ● ●

Properties of brittle adhesive.

Modeling Procedure

Construction of model using ABAQUS/CAE. Including of all mechanical property, elastic fracture properties. Creation of all sections. Applying all boundary conditions. Generate the mesh in all models. Running the model. Visualization of the response of the model at loading constrains applied.

3.5.5

Report

The report should include the following: ● ● ●

The definition of the important material properties. Definition of boundary conditions and meshing. Finally, the extraction of the necessary load–displacement data from the model. REPORT 3.5.5.1 Introduction

Modeling of bonded joints using finite element analysis (FEA) is now a commonplace procedure and has almost completely replaced the use of analytical models for the design of bonded joint in many applications. This process is mainly carried out in commercial FEA packages, such as Ansys and ABAQUS. In this chapter, the process for modeling an SLJ using a commercial FEA package is described in detail, showing the steps and the properties necessary to reach a simulated load–displacement curve. In this chapter, we will consider the use of ABAQUS. ABAQUS is a proprietary software widely used for FEA and Computer Aided Engineering (CAE). ABAQUS includes the following: ● ● ●

Abaqus/Standard (the solver for quasi-static loads) Abaqus/Explicit (the solver transient loads/impact) Abaqus/CAE (a visual interface)

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

An example of the Abaqus CAE interface is shown in Figure 3.42, showing an SLJ bending under the effect of tensile load.

Figure 3.42 Inc.

Example of a single-lap joint being modeled in Abaqus. Source: Abaqus

As is the case for most commercial FEA packages, ABAQUS does not include preset units of measurement. As such, it is the user’s responsibility to select and use a consistent set of units. For use with adhesively bonded joints, it is common to operate with dimensions in the order of millimeters and loads in Newton’s. This means that stresses will be considered to be in MPa (N/mm2 ). A list of consistent units based on this approach are summarized in Figure 3.43. Figure 3.43 Suggested unit system for modeling of adhesively bonded joints.

Length: mm (milimeters) Force: N ( Newtons) Stress/Pressure: MPa or N/mm2 Time: s (seconds) Mass: tonne (1E^3 kg) Temperature: °C (Celsius) Energy: mJ (milijoule) Thermal expansion: 1/°C (α) Thermal power: W (Watt) Thermal conductivity: W/mm.K

There are two main methods of operation with ABAQUS (and for most other FEA packages). We can follow an approach that uses the CAE visual interface, which allows us to build the model and analyze it directly with the tools in (Continued)

119

120

3 Laboratorial Activities and Report Examples

(Continued) the visual interface. Alternatively, the model can be created as an input file, with the properties, materials, and boundary conditions being programmed in an input file (.inp). This file can then be loaded to the appropriate solver, which will run it and output the result data. These two approaches are summarized in Figure 3.44.

CAE only approach

Command line

Model is built and analyzed inside Model is generated externally Abaqus/CAE and “.inp” file is run in Abaqus command

Figure 3.44

Possible modeling approaches in Abaqus.

In this specific example, we will follow an approach based on the ABAQUS/CAE visual interface as it provides a clearer process path for the beginner. We will also follow a design approach that will include the use of cohesive modeling. Cohesive zone models are able to combine the strength of materials approach with a fracture mechanics approach. In this model, cohesive elements are placed in the adhesive layer and are able to model crack propagation and thus can accurately reproduce the behavior of a joint all the way until its failure. Such approach is often complex to implement as it requires tensile and shear data as well as fracture energy values, but it has important advantages, allowing to precisely determine the location of failure within an adhesive layer, as shown in Figure 3.45.

Cohesive elements

Elastic elements

Figure 3.45 Use of multiple cohesive element layers in a single-lap joint, with damaged and failed elements shown in the inset.

It has been used successfully to model adhesive joint behavior by several authors, although its use is still relatively recent when compared to stress analysis-based methods. A cohesive model will employ a traction separation law to model the opening of the cohesive element, which represents the stress supported by the adhesive as a function of the relative displacement of the upper and lower nodes of the element being opened. The simplest of these laws is the triangular cohesive

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

law, which shows a linear variation in the elastic region (corresponding to the element stiffness) and then another linear variation in the damage region, after the peak stress is reached. Cohesive laws must be defined for at least two modes of loading (tensile and shear) with the behavior under mixed mode being interpolated by the FEA software. Examples of two triangular separation laws for mode I and mode II are shown in Figure 3.46. tn,s

E = 2270 M Pa

59.1 MPa

Gic = 1.25 N/mm

Model 1 Model 2

34.1 MPa

Giic = 2.5 N/mm

Model II

G

=

81

4

M Pa

Model I

(b)

δn,s

(a)

Figure 3.46 Triangular cohesive law for two loading modes (a) and examples of these loading modes (b).

Most authors use a triangular law or a combination between a bilinear (triangular) and trilinear (trapezoidal law) law. The laws can be modified to simulate the degradation of the adhesive caused by external factors. To determine these two laws, six different properties are required. For mode I, the properties are the Young’s modulus, tensile strength, and fracture energy in mode I. For mode II, the properties are the shear modulus, shear strength, and fracture energy in mode II. These properties and the experimental methodologies that are required for their determination are summarized in Figure 3.47. Test

Specimen shape

Property

Bulk tensile tests

Young modulus, tensile strength, deformation

Thick adherent shear tests

Shear modulus, shear strength, deformation

Double cantilever beam

Fracture energy (mode I)

End notched flexure

Fracture energy (mode II)

Figure 3.47 Summary of experimental testing methodologies used for determining the key mechanical properties of adhesives.

(Continued)

121

3 Laboratorial Activities and Report Examples

(Continued) Having established the necessary background, we can now advance to the modeling process description. As, as stated before, we will use the ABAQUS/ CAE-based approach, we will follow a well-defined workflow that is integrated in the software interface. We will first define a Part geometry, define the properties of the materials being modeled, create an assembly, define the simulation steps, create interactions and loads, mesh the model, and finally run the job and visualize the results. This workflow is summarized in Figure 3.48. •Part •Property Work flow

122

•Assembly •Step •Interaction •Load •Mesh •Job •Visualization

Figure 3.48

Workflow through the different Abaqus modules. Source: Abaqus Inc.

3.5.5.2 Module/Part

The model in ABAQUS/CAE can be constructed in two or three dimensions, and the modeling process starts with the definition of a sketch with well-defined dimensions. For a 2D analysis, this sketch is sufficient. However, for a 3D analysis, this sketch can be extruded in the normal direction, as shown in Figure 3.49.

Figure 3.49 Design and extrusion of a single-lap joint shape (left) and the Create Part menu (right). Source: Abaqus Inc.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

Alternatively, we can import an external sketch (in .dxf file format), create a part, and then apply the sketch to the drawing area, a process that is shown in Figure 3.50.

Figure 3.50 Abaqus Inc.

Process for importing a sketch created in an external software. Source:

As the 3D modeling capabilities of ABAQUS/CAE are limited, models can also be imported from other CAD software using the formats SAT, IGES, VDA, STEP, CATIA, Parasolid, ProE, among many others. This import process is shown in Figure 3.51.

Figure 3.51 Abaqus Inc.

Process for importing a model created in an external software. Source:

One important note regarding the design of the part is to explore the use of symmetry to simplify the model. By using symmetry to reduce the part size, one can significantly decrease the number of elements in the model and thus reduce (Continued)

123

124

3 Laboratorial Activities and Report Examples

(Continued) the calculation times, especially in 3D modeling. An example of a simplification procedure based on symmetry is shown in Figure 3.52. Modelled area

Figure 3.52 Example of a simplified model geometry, using symmetry to analyze one quarter of the specimen.

3.5.5.3 Module/Property

As state above, for cohesive modeling of an adhesive layer, six different properties are required (Young’s modulus, tensile strength, fracture energy in mode I, shear modulus, shear strength, and fracture energy in mode II). However, modeling materials with ABAQUS is of course much vaster and can take into account factors such as material density (important for dynamic analysis), thermal conductivity, and specific heat (useful for thermally coupled analysis), as summarized in Figure 3.53. Density∗

Conductivity∗

Young’s modulus

Poisson’s ratio

Specific heat∗ Stress strain curve ∗ - For thermal or dynamical models

Figure 3.53

Key properties used for modeling of adhesives.

Looking again at the specific case of the cohesive element, we need to introduce in ABAQUS the elastic properties the tensile properties and the fracture properties. 3.5.5.3.1 Elastic Properties for Tensile Separation Laws In a traction separation law, the initial elastic portion is always kept linear, defined by a constitutive matrix (K), containing the stiffness parameters and relating current stresses (t) and strains (𝜀) in the three loading modes across the interface.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

A suitable approximation for a thin adhesive layer is given by the following K parameters: K nn = E, K s1s1 = K s2s2 = G, and K ns1 = K ns2 = K s1s2 = 0 (E and G are the Young’s and shear moduli, respectively). The basic stiffness matrix is given by Eq. (3.41). ⎧t ⎫ ⎡K K K ⎤ ⎧𝜀 ⎫ ⎪ n ⎪ ⎢ nn ns1 ns2 ⎥ ⎪ n ⎪ t = ⎨ts1 ⎬ = Kns1 Ks1s1 Ks1s2 . ⎨𝜀1 ⎬ = K𝜀 ⎪ts2 ⎪ ⎢⎣Kns2 Ks1s2 Ks2s2 ⎥⎦ ⎪𝜀2 ⎪ ⎩ ⎭ ⎩ ⎭

(3.41)

In the ABAQUS/CAE interface, this is defined by creating a material that has a traction-type elastic behavior and defining the tensile modulus E nn and the shear moduli E ss and E tt . This is shown in Figure 3.54.

Figure 3.54 Setting elastic properties for cohesive modeling in the Edit Material menu. Source: Abaqus Inc.

(Continued)

125

126

3 Laboratorial Activities and Report Examples

(Continued) 3.5.5.3.2 Damage for Traction Separation Laws The Quads damage initiation criterion is used to predict damage initiation in cohesive elements where the cohesive layers are defined in terms of traction separation. The criterion evaluates the stress ratios between a given stress value and the peak nominal stress value in each of the three directions. The Quads criterion is based on a quadratic combination of all the three ratios. Damage initiation can be specified by different criteria. In many cases, the simple quadratic nominal stress criterion is considered, as shown in Eq. (3.42). { }2 { }2 } { ts1 ts2 ⟨tn ⟩ 2 + + =1 (3.42) 0 0 tn0 ts1 ts2

To introduce the stresses necessary to induce damage in a cohesive element, one should define a Quads Damage condition and apply the tensile strength of the material to the normal-only mode nominal stress. The shear strength can then be used for the nominal stress in the first direction and the second direction, as shown in Figure 3.55.

Figure 3.55 Setting stress limits for cohesive modeling in the Edit Material menu. Source: Abaqus Inc.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

After the mixed-mode cohesive strength is attained (t m 0 ) by the fulfillment of the quadratic nominal stress criterion, the material stiffness is degraded. Complete separation is predicted by a linear power law form of the required energies for failure in the pure modes, as defined in Eq. (3.43). Gn Gs1 Gs2 + + =1 Gcn Gcs1 Gcs2

(3.43)

To define this degradation process, one should access the Suboption Editor available within the Edit Material form. In this case, we can control different aspects of the second part of the traction separation law, such as the type of damage evolution, the softening process, the degradation, and the mixed mode behavior. One can also define the mode mix ratio. By defining the damage evolution type as energy, it is possible to introduce the fracture energy in mode I as normal mode fracture energy and the fracture energy in mode II as shear mode fracture energy first direction and second direction (Figure 3.56).

Figure 3.56 Defining damage evolution parameters in the Suboption Editor within the Edit Material menu. Source: Abaqus Inc.

The damage evolution defines how the material degrades after one or more damage initiation criteria are met. Multiple forms of damage evolution may act on a material at the same time – one for each damage initiation criterion that was defined. Energy damage evolution defines damage in terms of the energy required for failure (fracture energy) after the initiation of damage. This type corresponds to the fracture energy field in the data table. Linear softening specifies a linear softening stress–strain response for linear elastic materials or a linear evolution (Continued)

127

128

3 Laboratorial Activities and Report Examples

(Continued) of the damage variable with deformation for elastic–plastic materials. Linear softening is the default method. The maximum degradation form indicates that the current damage evolution mechanism will interact with other damage evolution mechanisms in a maximum sense to determine the total damage from multiple mechanisms. Maximum is the default selection. Power law mixed mode behavior specifies the fracture energy as a function of the mode mix by means of a power law mixed mode fracture criterion; it is available only when you select the energy type with cohesive elements. The fracture energy field in the data table is replaced by the fracture energy in the normal mode and first direction and second direction shear mode components. Energy mixed mode ratio defines the mode mix in terms of a ratio of fracture energy in the different modes. This definition is the default, and it must be used when one selects power law or the BK law for the mixed mode behavior. Finally, an example of material properties necessary for cohesive modeling of a brittle epoxy adhesive is shown in Table 3.26. Table 3.26

Example of cohesive properties for a brittle epoxy adhesive.

Property

Test

AV 138

Young’s modulus E (MPa) Shear modulus G (MPa) Tensile strength tn0 (MPa) Shear strength ts0 (MPa) Mode I fracture energy GIC (N/mm2 ) Mode II fracture energy GIIC (N/mm2 )

Failure strength test Thick adherend shear test (TAST) Failure strength test Thick adherend shear test (TAST) Double-cantilever beam (DCB) test End notched flexure (ENF) test

4890 1560 41.0 30.2 0.35 4.91

3.5.5.4 Module/Section

The properties we have defined must then be assigned to sections, and those sections are themselves assigned to the geometrical locations within the actual model. Different types of sections can be defined. As an example, one can use Solid Homogeneous sections to define elastic materials, while Cohesive sections are fundamental for use of cohesive zone modeling. The Section Assignment Manager and some example sections are shown in Figure 3.57.

Figure 3.57 Section Assignment Manager showing correlation between Sections, Materials, and Regions. Source: Abaqus Inc.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

To create sections, the Create Section menu is used. Here, one can select the “Solid/Homogenous” or “Other/Cohesive” sections, among many others. For 2D models, it is important to enter the out-of-plane thickness of each section in the Edit Section menu as this will allow the correct calculation of the failure load for the model under analysis, accounting for material thickness. The Edit Section menu is also where the material properties (previously defined) are attributed to a given section. This process is shown in Figure 3.58.

Figure 3.58 Creation of Sections and assigning them to previously defined materials. Source: Abaqus Inc.

Once the sections are created, they must be matched to geometrical location on the model. This is done by picking the geometrical location within the model in the Edit Section Assignment menu, as shown in Figure 3.59

Figure 3.59 Defining the section assigned to a specific model region. Source: Abaqus Inc.

3.5.5.5 Module/Step (First Phase)

Once the sections are created with a given material property set and correctly assigned to model locations, one can create the analysis Step where the analysis will be performed. Multiple steps can be created for a single model. For example, one can sequentially subject the model to different types of loads by creating a sequential set of steps. In this case, we will look at a (Continued)

129

130

3 Laboratorial Activities and Report Examples

(Continued) single-step creation. There are multiple types of Step procedures available. Among the most basic are the Static, General steps for stress and displacement analysis and “Coupled temp-displacement” if one wishes to add temperature distributions. The creation of a new Step is shown in Figure 3.60. Figure 3.60 Create Step menu, showing some of the many model steps permitted. Source: Abaqus Inc.

The Step definition also includes the incrementation controls that govern the progression of the model. The increment size corresponds to a percentage of the

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

simulation duration. An increment size of 1 represents 100% of the simulation and 0.01 represents 1%. The initial, minimum, and maximum step size should be defined to determine how the simulation progresses, avoiding extremely large jumps. The Edit step menu is shown in Figure 3.61.

Figure 3.61

Edit Step menu, showing the incrementation control. Source: Abaqus Inc.

3.5.5.6 Module/Load

The Load module is used to introduce loads and boundary conditions that act over the model for each step. Loads can be pressures, forces, moments, etc., while the boundary conditions are mostly restrictions on movement and rotation, defined for in one or multiple axis. Typical boundary conditions for an SLJ are shown in Figure 3.62, where one can see an imposed displacement of 5 mm and limited movement in the vertical direction on one side of the joint, as well as pinned behavior of the opposite side of the joint. y

Ux = 0 Uy = 0

x

δ = 5 mm

Uy = 0

Figure 3.62 Example of boundary conditions suitable for modeling a single-lap joint under tension.

Many options for mechanical loads and boundary conditions are available, as show in the Create Load and Create Boundary Condition menus shown in Figure 3.63. (Continued)

131

132

3 Laboratorial Activities and Report Examples

(Continued)

Figure 3.63 Create Load and Create Boundary Condition menus, listing the types of loads and boundary conditions allowed. Source: Abaqus Inc.

3.5.5.7 Module/Mesh

In ABAQUS/CAE, meshing is carried using a set of tools that allows us to apply “seeds” that will serve to guide the division of the model into a well-defined mesh of elements. The main tools available for this purpose are shown in Figure 3.64. •Apply seeds (base points for the mesh) •Apply mesh controls •Select the appropriate element type •Create the mesh

Figure 3.64

Main meshing tools available in Abaqus.

Use the seed controls to reduce the number of elements and use structured mesh to guarantee alignment of elements. Once the mesh seeds are created, one can then pick the most appropriate element type for each case. Examples of element types are plain strain (CP8R) plain stress (CPS8R), coupled temperature displacement (C3D8T), and cohesive (COH3D8).

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

One important additional meshing tool is the Mesh Control menu (Figure 3.65), which allows us to define both the element shape (Quad, Quaddominated or Tri) and a specific meshing technique that is required for the use of cohesive elements. This is the Sweep type of mesh technique.

Figure 3.65

Mesh Control menu. Source: Abaqus Inc.

The use of Sweep meshing with cohesive elements requires the correct definition of the Sweep Path. This is done using the Redefine Sweep Path control that will ask the user to verify the orientation of a red arrow placed in the periphery of the meshed area, allowing the cohesive element to define their expected opening direction. As shown in Figure 3.66, a correctly defined Sweep Path will show the arrow to be perpendicular to the adhesive layer. If this is not the case, the user must select a new orientation and even position of this arrow.

Figure 3.66 Correct redefinition of a Sweep Path, showing the arrow (red) perpendicular to the adhesive layer (purple).

(Continued)

133

134

3 Laboratorial Activities and Report Examples

(Continued) 3.5.5.8 Module/Step (Second Phase)

Before calculating the model results, it is now important to return to the step module, which will place data requests on the model, generating the necessary data that will allow us to extract meaningful load–displacement curves from the model results. These curves can then be compared against the experimental data for model validation. Within the step module, two different types of data can be requested. These are ● ●

history outputs (such as forces and displacements) field outputs (field values, such as degradation of cohesive elements)

The Field Outputs Requests are accessible from the Output dropdown menu and provide access to the Manager and the Create menus, where one can manage and create new requests. This is shown in Figure 3.67.

Figure 3.67

Field Output Request tools.

However, to create a History Output, one must first have created a node set where the history output data will be requested. This will correspond to the selection of specific nodes for displacement and force. In the case of an SLJ, a displacement node set will be created by grouping the nodes subjected to the imposed displacement, while the load node set will be defined as the nodes subjected to the reaction forces that hold the specimen in place. This is summarized in Figure 3.68.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

y

Ux = 0 Uy = 0

(a) Force nodes

(b)

x

δ = 5 mm

Uy = 0

Displacement nodes

(c)

Figure 3.68 Process of creation of two sets of nodes, one for extracting displacement data and another for extracting load (reaction force) data. Node set manager menu (a), definition of boundary conditions Source: Abaqus Inc. (b), and the selection of the respective force and displacement node sets (c).

Once the node sets are defined, the load and displacement values (and their evolution as a function of time increments) are requested in the History output menus. These procedures require the definition of the domain whereupon the calculation is made (pointing to the correct node set) and the output variables that are to be recorded by the software. Furthermore, it is also possible to reduce the amount of data being recorded by defining a specific recording frequency, which can be defined to record at specific increment intervals. This process is summarized in Figure 3.69.

Figure 3.69 History Output Requests Manager, showing the possibility to request values of displacement for a given node set. Source: Abaqus Inc.

(Continued)

135

136

3 Laboratorial Activities and Report Examples

(Continued) The same process can be done for the Field Output request, which is used to request the stiffness degradation (SDEG) of the cohesive elements and can even be used to command their deletion when the damage reaches 100% (using the STATUS command). In this case, the request is made for the whole model, and there is no need to define a specific node set for the calculation of the data (Figure 3.70).

Figure 3.70 Field Output Requests Manager, showing the possibility to request values for the whole model. Source: Abaqus Inc.

3.5.5.9 Module/Job

Finally, the model is submitted and monitored in the Job Manager dialog box, as shown in Figure 3.71. Using the Edit button is possible to adjust the processor and memory usage to optimize model performance.

Figure 3.71 Job Manager menu, showing the Write Input, Data Check, Submit, Monitor, and Result controls. Source: Abaqus Inc.

3.5 Modeling a Single-Lap Joint Using Finite Element Analysis and Cohesive Zone Modeling

3.5.5.10 Module/Visualization

Lastly, we arrive at the visualization module that can show the stresses, displacements, or other parameters acting on each element. Pictured in Figure 3.72 is the SDEG (stiffness degradation of cohesive elements).

Figure 3.72 Visualization of stiffness degradation within the adhesive layer of a single-lap joint.

The load–displacement curve is also an important output as it enables the direct comparison between experimental and numerical data. An example of a load–displacement curve is shown in Figure 3.73, also showing experimental data and how failure mode is predicted with element damage (SDEG). 9 Experimental Numerical

8

95% confidence interval of experimental failure load

7

Load (kN)

6 5 4 3 2 1 0 0.0

0.2

0.4

0.6

0.8

1.0

Displacement (mm)

Figure 3.73 Example of a load–displacement curve generated using history output data compared with an experimental curve, compared with experimental data.

(Continued)

137

138

3 Laboratorial Activities and Report Examples

(Continued) This curve is obtained by combining the time variation of the reaction forces and the imposed displacement. The first step is extracting the sum of the reaction force data, which we have previously requested for the nodes where the boundary conditions restrict the model. This is done from the Create XY data menu, sourcing data from the OBD (output database) history output. All the data available, for all nodes, will be displayed in a single window. The user can then combine the data in a single value but summing it, using the Save XY Data As menu. The process of extracting the Reaction Forces and the Load acting on the model is shown in Figure 3.74. .

Figure 3.74

Process for saving the reaction forces as a XY dataset. Source: Abaqus Inc.

This is followed by spatial displacement of a single node identified as the imposed displacement, again using the previously requested data. This process is similar to that shown in Figure 3.74, but it only requires the selection of a single data point, as displacement of a single node is the same as the displacement of all nodes at the displacement boundary condition. This is shown in Figure 3.75.

Figure 3.75 Abaqus Inc.

Process for saving the spatial displacement as a XY dataset. Source:

3.6 Case Study in Joint Design for a Structural Automotive Application

The XY datasets for displacement and load are at this stage independent and listed as a function of the increment time. To obtain the load–displacement curve, they must be combined into a single XY dataset, using the operate XY data window. In the case, the Operate on XY data option must be selected, which allows us to use a special function (combine) to join them, providing the Load as a function of the Displacement (Figure 3.76). The resultant data can then be plotted directly on ABAQUS/CAE or extracted to an external software for processing.

Figure 3.76 Combining the displacement and load using the Operate on XY data menu. Source: Abaqus Inc.

3.6 Case Study in Joint Design for a Structural Automotive Application 3.6.1

Introduction

In the automotive industry, joint design is a crucial activity, which must balance diverse aspects such as strength, durability, and manufacturability. Thus, both the materials being used and the joint geometry must be carefully assessed to reach a desired joint behavior.

3.6.2

Report

The report should include the following: ● ●

Numerical modeling of a typical geometry for the automotive industry; Experimental validation activities that are representative of those used in the automotive industry.

139

140

3 Laboratorial Activities and Report Examples

REPORT 3.6.2.1 Introduction

The use of adhesive bonding in the automotive industry has significantly expanded in the past decades because of its advantages when compared with traditional joining methods. Despite this, it can still be difficult to define a consistent approach for the design of adhesively bonded joints, as joint performance can depend on the combination of several factors. This work defines and lists the steps related to the design process of a bonded joint, summarizing the necessary information and describing the decisions and their underlying rationale. Figure 3.77 summarizes a typical joint design process, as implemented in the automotive industry.

Define the design brief

Perform adhesive selection

Select surface treatement

Validate the design

Design and predict joint strength

Determine all relevant material properties

Implement in manufacturing

Figure 3.77

Joint design workflow.

3.6.2.2 Design Brief

The objective of this work is to study the design of a joint commonly used in the automotive industry. In this case, we will consider the sill side beam joint of a vehicle built from extruded sections of a high-strength aluminum alloy. These extruded sections must be joined at key locations within the vehicle body, something that can be achieved using multiple methods, such as riveting, bonding, or welding. Figure 3.78 exemplifies the location of this type of joint within a vehicle frame.

Figure 3.78

A vehicle sill side beam joint.

3.6 Case Study in Joint Design for a Structural Automotive Application

To simplify the analysis, we will reduce the complexity of this joint by considering a simplified T-joint geometry, connecting two aluminum beams with a box section, as shown in Figure 3.79. Aluminium pillar

Aluminium cross-beam

Figure 3.79

The configuration of the bonded area under analysis.

3.6.2.2.1 Why Use Adhesives? In this initial work, we are purposely exploring

the use of adhesives, but how can we ensure that adhesive bonding the optimal solution? In fact, there is no perfect joining method. Obviously, different joining processes will have different advantages and limitations. For adhesive bonding, in specific, one can refer the advantages and disadvantages shown in Figure 3.80. Advantages: Improved stress distribution; Good fatigue performance; Minimizes secondary operations; Allows bonding of dissimilar substrates.

Disadvantages: Requires curing time; Requires careful surface preparation; Cannot be easily dismantled.

Figure 3.80 Relative advantages and disadvantages of adhesives in automotive applications.

We do have very significant and useful advantages in this list, but how can we mitigate the disadvantages? There are different ways to do so, as we will see throughout this work. The requirement of short curing time can be minimized with the use of heat to accelerate cure. The requirement for a careful surface preparation can be overcome by selecting a simple but effective surface treatment. What about the difficulty associated with the dismantlement of the joint? This is harder, but we can design the full structure with recycling into account and even consider adhesives that have self-dismantling capabilities. 3.6.2.2.2 Loading Conditions As a worst-case scenario, we will consider a rollover case, with a peak of 3G’s acceleration on the vertical direction. If we

(Continued)

141

142

3 Laboratorial Activities and Report Examples

(Continued) consider that 25% of the vehicle’s loaded weight (1900 kg) is carried by this beam, this corresponds to a load on the joint of around 14 kN. In a real-world application, strain rate effects should also be considered, but this will not be approached in this case study for the sake of simplicity. 3.6.2.2.3

Adherend Dimensions For this example, we will consider the geometry

of the extruded aluminum beams as defined by the vehicle structural loads and therefore will not be changed during joint design. The dimensions of the bonded area are shown in Figure 3.81. Bonding area

w = 80 mm

t = 2 mm

Figure 3.81

h = 50 mm

Dimensions of the bonded area under consideration.

3.6.2.2.4 Environmental Conditions Besides the actual load, any joint used in the automotive sector must also withstand harsh environmental conditions. These include the following: ●

● ● ●

Extreme temperatures: −40 to 80 ∘ C (typical range for the automotive industry); High relative humidity: 0–100% (corresponding to dry and wet environments); Fatigue loadings: cyclic loads over vehicle lifetime; Long bonded joint lifetime: 30 years (as defined by the manufacturer).

These appear to be impossible demands, but in fact, the correct selection of an adhesive and proper joint design can overcome many of these requirements. We will start by looking at adhesive selection in the following section. 3.6.2.3 Adhesive Selection

Although there are a large number of adhesive formulations available for structural use, for this application, the four adhesive formulations summarized in Figure 3.82 are typically used. Two-part acrylic adhesives

Conventional onepart epoxies

Two-part polyurethanes

Hybrid adhesives

Figure 3.82 Main families of structural adhesives used in the automotive industry.

3.6 Case Study in Joint Design for a Structural Automotive Application

We can look at each of these options in more detail in order to better understand the most adequate for this specific application. 3.6.2.3.1 Two-Part Acrylics Two-part acrylics (methyl methacrylate) have, as

advantages, high cure through depth, room temperature cure, high peel and impact strength, good environmental resistance, bond to moderately contaminated surfaces, and their cure can be accelerated with heat. However, their use is limited by slow fixture times (5–30 min), the generation of waste associated with static mix process, the strong odor they emit, and the low strength and stiffness, when compared with epoxy-based adhesives. This adhesive is interesting for some automotive structural application curing process because of its very fast curing, which has the potential to greatly accelerate the manufacture process. However, the relatively low strength and stiffness are somewhat undesirable for automotive applications. 3.6.2.3.2 Conventional One-Part Epoxies One-part epoxies are very popular adhesives, with a wide range of formulations available. They provide a very high level of adhesion to diverse adherends as well as excellent mechanical strength, durability, and environmental resistance. These are, of course, highly desirable features for automotive structural construction. The cure process can be accelerated with the use of heat, but fixturing systems are essential. They can be quite brittle, especially in an unmodified state, which limits some of their effectiveness under impact loading conditions. 3.6.2.3.3 Two-Part Polyurethanes Two-part polyurethanes are quite flexible and tough. They are also strong, especially under impact conditions and operate well under very low temperatures. The cure process of these adhesives is not very fast, especially for very large adhesive thicknesses such as those that are usually found in automotive applications. Often times, the use of a primer is mandatory with these adhesives, which requires the addition of costly additional stages to the manufacturing process. 3.6.2.3.4 Hybrid Adhesives Finally, we can consider the use of hybrid adhesives,

which are usually highly modified epoxies. By combining different chemical formulations in the same product, they provide high toughness and strength. In practice, these adhesives benefit from the strength of the epoxies and from the ductility of the polyurethanes while retaining good resistance to solvents and environmental conditions. They excel under impact conditions. However, the use of these materials is somewhat limited by their high temperature of cure and high cost. 3.6.2.3.5 Selected Adhesive For the intended application, it is clear that a high-ductility hybrid adhesive is the most well-suited adhesive type. This can be supported by looking at the chart shown in Figure 3.83, which shows that

(Continued)

143

3 Laboratorial Activities and Report Examples

(Continued) these adhesives provide a well-rounded performance in all of the key metrics under consideration. Stiffness 5 4 3 2

Hardening time

Strength

1 0

Two-part acrylic adhesives One-part epoxy adhesives Two-part polyurethanes Hybrid adhesives

Solvent resistance

Figure 3.83 industry.

Toughness

Relative performance of four different adhesives used in the automotive

An example of a hybrid adhesive is the Nagase-Chemtex XNR6852E-3 adhesive. This adhesive has a very high failure load (similar to that of pure epoxies), but it also fails in a much safer manner, deforming plastically before joint separation occurs, redistributing and reducing peaks in the shear stress distribution. This is evident in the plot shown in Figure 3.84. 60 AV138 50 (epoxy) Tensile stress (MPa)

144

40

XNR6852 E-3 (crash resistant adhesive)

30 20 10 0

0

2

4

6

8

10

12

14

16

Strain (%)

Figure 3.84 Comparison between the tensile curves for a crash-resistant adhesive and a conventional one-component epoxy formulation.

3.6 Case Study in Joint Design for a Structural Automotive Application

Stress (MPa)

3.6.2.3.6 Adherend Material Selection As we have defined that we would consider an aluminum-bodied vehicle for this application, the adherends used in this work are composed of 6063 T6 alloy, a medium strength alloy with excellent corrosion resistance. The aluminum stock used for this example is in the tempered state (T6). Figure 3.85 shows the stress–strain curve of this material as well as the samples used in this work. 400 350 300 250 200 150 100 50 0

0

0.1

0.05

0.15

ε

Figure 3.85 Stress–strain curve for the 6063 T6 aluminum alloy under consideration (left) and the beams used in the T-joint construction (right).

3.6.2.4 Surface Treatment Selection 3.6.2.4.1 Available Surface Treatment Processes The use of a surface treatment is

critical for avoiding the highly undesirable adhesion failure. It also has a major role in improving long-term durability, especially in adverse environmental conditions. We have diverse surface treatment options for our aluminum adherends. We can ● ● ● ● ●

leave in as supplied state; perform solvent cleaning; use shot blasting/sandblasting; perform chemical etching; carry out an anodization process.

3.6.2.4.2 Identification of the Most Effective Surface Treatment The use of contact angle measurement provides a quick way to assess the suitability of a surface treatment. Surfaces with low surface energy necessary present high contact angles and bad adhesion properties. Effective surface treatments will convert high contact angles to low contact angles, representing an increase in surface energy. Durability is significantly increased with the use of an adequate surface treatment. The literature shows the most adequate treatment for aluminum, considering both adhesion and durability, is the phosphoric acid anodization (PAA) process. This process uses electrical current to drive the formation of a thick and stable oxide structure in the surface of the aluminum adherends, which

(Continued)

145

146

3 Laboratorial Activities and Report Examples

(Continued) will provide a mechanical anchoring point to the adhesive. This process is shown in Figure 3.86

Figure 3.86

Image of an anodization process (right) and the end result (left).

3.6.2.5 Material Properties

Joint design is based on the determination of stresses acting upon the adhesive and the adherends, which are then analyzed using a failure criterion. 3.6.2.5.1 Mechanical Characterization of Adhesives What adhesive properties are required for strength prediction in a joint design context? At the most basic, the main properties necessary to enable this calculation are the stiffness (in tensile and shear directions) and the yield stresses (also in tensile and shear). Some more powerful criteria or methodologies also require the toughness of the material, assessed by its fracture toughness (measured in mode I and mode II loads). Examples of tests suitable for determining all these mechanical properties are shown in Figure 3.87. Test

Specimen shape

Property

Bulk tensile tests

Young modulus, tensile strength, deformation

Thick adherent shear tests

Shear modulus, shear strength, deformation

Double cantilever beam

Fracture toughess (mode I)

End notched flexure

Fracture toughess (mode II)

Figure 3.87 Key mechanical testing procedures used for the characterization of structural adhesives.

3.6 Case Study in Joint Design for a Structural Automotive Application

The bulk tensile test and the thick adherent shear test (TAST) allow us to determine stress–strain curves that contain information on the peak stresses supported by the adhesive, as well as the stiffness for both tensile and shear conditions (Figure 3.88).

50

Shear stress (MPa)

Tensile stress (MPa)

60

40 30 20 10 0

50 45 40 35 30 25 20 10 5 0

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

0

1

2

3

4

5

6

7

8

9

10 11

Strain (%)

Strain (%)

Figure 3.88 Tensile stress–strain curve (left) and shear stress–strain curve (right) for the hybrid adhesive under consideration.

The curves shown indicate that the selected adhesive has a significant amount of ductility and a high yield stress, giving it an excellent mechanical behavior. The complete mechanical characterization data of the adhesive as well the tests used to characterize it are shown in Table 3.27. Table 3.27

Key mechanical properties of the XNR6862-E3 adhesive.

Property

Value

Test

Young’s modulus E (MPa)

1728

Failure strength test

Shear modulus G (MPa)

665

Thick adherend shear test (TAST)

Tensile strength tn0 (MPa)

51.5

Failure strength test

ts0

45

Thick adherend shear test (TAST)

Mode I fracture energy GI (N/mm2 )

6.37

Double cantilever beam (DCB) test

Mode II fracture energy GII (N/mm2 )

51 (*)

End notched flexure (ENF) test

Shear strength

(MPa)

3.6.2.5.2 Data Required for Durability Prediction Prediction of fatigue behavior is usually based in experimentally obtained S–N or Paris law curves. These are usually determined for joint level specimens. The assessment of aging effects can be conducted using artificially aged bulk specimens of adhesive. These specimens enable the determination of moisture uptake and the measurement of changes in properties as a function of humidity. For example, one can subject bonded double cantilever beams (DCBs) to aging conditions and then assess the influence of these parameters on the mode I fracture energy. The creep behavior of the adhesive is experimentally assessed with strain–time curves obtained at different temperatures. These curves can be

(Continued)

147

3 Laboratorial Activities and Report Examples

(Continued) assembled into a master curve to predict long-term behavior of a joint using the time–temperature superposition principle, as shown in Figure 3.89. 0.16 log10 (a100 °C) = 2.93

0.14 0.12 0.1 Strain

148

100 °C 0.08 75 °C

0.06

log10 (a75 °C) = 1.14

0.04 60 °C

0.02

log10 (a60 °C) = 0.78 Room temperature

0 1

0

2

3

4

5

6

log10 time (hours)

Figure 3.89 Adhesive creep curves obtained at different temperatures being used to construct a master curve, suitable for the long-term prediction of the adhesive creep behavior at room temperature.

3.6.2.6 Joint Design

Joint design is the process of defining the geometrical configuration of the joint to achieve the intended mechanical performance. The geometrical design possibilities for bonded joints are almost endless, which requires a designer to precisely define the key points that will be considered and optimized. For the specific application approached in this work, one can simplify the analysis by considering the parameters shown in Figure 3.90. Note that as we have locked the adherend dimensions, we will not consider the effect of adherend thickness of shape. Adhesive layer dimensioning

Overlap length

Adherend thickness

Overlap width

Adherend shaping

Adhesive thickness Adhesive layer shaping

Figure 3.90

Limited in our application example

Main design parameters under consideration.

3.6 Case Study in Joint Design for a Structural Automotive Application

Overall, the target of a joint process design should also be able to maximize the shear loads carried by the joint while minimizing the peel and cleavage loads. To increase the strength, one can increase the overlap length or joint width to increase the joint strength. However, an increase in overlap does not guarantee an increase in failure load. It is important to take into account yielding or failure of the adherends. In many cases, it is better to increase the bondline width, as this also requires wider, and stronger, adherends. What about increasing the adhesive layer thickness? This has a negative effect on joint strength as large bondline thicknesses reduce joint strength. In fact, various research studies have demonstrated that for structural adhesives, the optimal bondline thickness is around 0.2 mm. We will use this value for our design. Also of note is the fact that spew fillet tends to increase the joint strength. If possible, we will not remove it from our joint. Adhesive joint design is typically performed using analytical or numerical models, as shown in Figure 3.91. Figure 3.91

Joint design approaches. Failure load prediction

Analytical models

Numerical models

Analytical models are simple to use and faster and for many decades have been the sole methodology to design bonded joints. However, their scope is limited, being restricted both in geometrical configurations and in material behavior. Examples of geometries that are typically studied with analytical models are the single-lap joints, butt joints, and tubular joints, which provide well-defined geometrical parameters and constraints. In contrast, numerical models are much more flexible and powerful but are also more complex and resource intensive. They allow the study of any kind of geometry and very complex material behavior. Their use in adhesive joint design is now extensive, with wide application in automotive and aerospace industries. 3.6.2.6.1 Joint Design Using Analytical Models We will start our analysis by adapting four different analytical models to study the performance of our T-joint. Please note that this requires some important simplifications to be made. The four analytical models to be used in this work are shown in Figure 3.92.

(Continued)

149

150

3 Laboratorial Activities and Report Examples

(Continued)

Failure load prediction

Analytical models

Generalized yielding of the adhesive

Figure 3.92

3.6.2.6.2

Yielding of the substrate

Volkersen model

Hart Smith model

Analytical models used for strength prediction in this work.

Adhesive Generalized Yielding Criterion For extremely ductile adhesives

(those with more than 20% of shear strain), all of the overlap length is subjected to plastic deformation (global yielding criterion), and the failure load of the joint (P max ) can be estimated simply from the shear yield stress of the adhesive (𝜏 y ), the joint width (p), and the overlap length (l). This is shown in Eq. (3.44). Pmax = 𝜏y bl

(3.44)

Given the relatively large strain of the adhesive we have selected, we can use the adhesive generalized yielding criterion to quickly determine the joint strength if all the available bonding area is used and the adherend does not yield. If we assume ● ●

b = 50 mm (width of the beam) 𝜎 y = 45 MPa We can then state that Pmax = 45 MPa × 50 mm × 50 mm

(3.45)

and thus, Pmax = 112.15 kN A value that is eight times higher than the target load of 14 kN.

(3.46)

3.6 Case Study in Joint Design for a Structural Automotive Application

3.6.2.6.3 Adherend Yielding Criterion (SLJ) For single-lap joints, simple equations are available to calculate the failure of the joint driven by the failure of the adherend (Eq. 3.47).

Pmax =

𝜎y bt (1 + 3k)

(3.47)

For small loads and short overlap lengths, we reach a special case where k ≈ 1, which gives 𝜎y bt

(3.48) 4 However, for successively higher overlap lengths, the value of k diminishes. When the ratio l/t ≥ 20, k ≈ 0, we reach Pmax =

Pmax = 𝜎y bt

(3.49)

In our case, because of the high stiffness of the adherends and the very high loads, we will assume that the lower adherend will be loaded evenly in a shear direction. If we assume ● ●

a = 284 mm2 (cross section of beam) 𝜏 y = 152 MPa (yield stress of 6061 T6) We can find the load sustained when then adherend yields Pyield = 504 mm2 × 152 MPa = 43.16 kN

(3.50)

3.6.2.6.4 Adherend Yielding (Torsion + Transverse Shear Effect) A more accurate approach would consist of determining the stress present on the critical section of the specimen. In our case, because of the asymmetrical loads, we can expect a combination of torsion and shear applied near the supports. To calculate the torsional and shear stress on the lower beam, we can consider ● ● ● ● ●

A = 284 mm2 (cross-sectional area) Am = 1104 mm2 (average area) 𝜏 max = 152 MPa (shear stress of 6063 T6) Mt = P yield × l = P yield × 12 bmin = 2 mm (thickness) In this case, the torsional stress can be assumed to be Mt f = 2 × Am

(3.51)

(Continued)

151

152

3 Laboratorial Activities and Report Examples

(Continued) The maximum shear stress is thus shown in Eq. (3.52): 𝜏max =

f

(3.52)

bmin

which leads to a predicted yield strength of Pyield = 24.00 kN

(3.53)

3.6.2.7 Numerical Models

Numerical models, as stated above, are quite powerful and flexible. However, the range of design processes they offer can be quite daunting as many multiple possible exist. Classically, to analyze bonded joints, two different approaches have been followed. These are the strength of material-based approaches and the fracture mechanics-based approaches. The relative advantages and disadvantages of these procedures are summarized in Figure 3.93. Stress or strain criteria are established to identify when the adhesive fails. Strength of materials approach

Not well suited for adhesive joints, as stress concentration points prevent the successful use of these criteria. Uses the onset of crack propagation in the singularities to predict failure.

Fracture mechanics approach

It is difficult to assess stress intensity factors and in many cases there is no crack present in the model.

Figure 3.93 Relative advantages and disadvantages of the strength of materials and fracture mechanics approaches.

Numerical models can also be calculated using implicit or explicit code. Most simulations for automotive structural applications are based on explicit FE codes as these are capable of simulating fast events (such as impacts) within a reasonable execution time. Explicit FE codes do not require the solution of systems of equations as in implicit FE codes. 3.6.2.7.1 Cohesive Zone Modeling We have seen that both the strength of material approach and fracture mechanics-based approach have important disadvantages. Cohesive zone models can combine these two approaches in a single model. Cohesive elements are placed in the adhesive layer and can model adhesive damage, failure, and the subsequent crack propagation. These models

3.6 Case Study in Joint Design for a Structural Automotive Application

require tensile and shear data as well as fracture energy values. They have been successful to model the performance of bonded joints under multiple loading conditions, including the effects of thermal loads and strain rates. Most authors use a triangular law or a combination between a bilinear (triangular) and trilinear (trapezoidal law). We will use the properties of the adhesive to form a triangular cohesive law. Model Configuration To construct our model, all the available overlap will be considered as effectively bonded (50 × 50 mm). A simple plastic behavior (without damage) will be used to model the aluminum beams, while the adhesive layer is modeled with cohesive elements using a triangular cohesive zone law. The general construction of the model is shown in Figure 3.94.

3.6.2.7.2

Substrates

Adhesive layer

Figure 3.94

General construction of the model.

The boundary conditions used in this model will approximate a three-point bending condition. The lower adherend is supported along a line, allowing for bending and approximating the behavior of a longer beam, while the upper adherend is compressed toward the lower beam. Symmetry conditions are used for reduction of computational loads, which means that only half the model is simulated. Boundary conditions are shown in Figure 3.95. (Continued)

153

154

3 Laboratorial Activities and Report Examples

(Continued) Figure 3.95

Boundary conditions of the model.

A uniform mesh of 1.25 × 1.25 mm was then used, with a total of 37773 elements, of which 36914 are linear hexahedral C3D8R elements and 820 linear hexahedral COH3D8 elements. The mesh was of the structured type for the C3D8R elements while the COH3D8 elements used a sweep type of mesh. The mesh is shown in Figure 3.96. Figure 3.96

Mesh of the model.

3.6 Case Study in Joint Design for a Structural Automotive Application

Numerical Model Results Once the model was meshed, it was run and the load, displacement, and damage data were extracted. These results show that the joint suffers significant plastic deformation on the horizontal adherend, with significant deformation on the horizontal direction, shown in Figure 3.97. According to the FEA, the peak load sustained by the structure is around 21 kN. The maximum sustained load occurs at the onset of plastic deformation of the adherends. The damaged model shape is shown in Figure 3.97.

3.6.2.7.3

Figure 3.97 Deformed model shape.

If one isolates the adhesive layer, it becomes possible to understand what has occurred within the adhesive layer. In fact, although there is massive deformation of the beams, the adhesive only suffers partial failure, achieving 100% of cohesive degradation only near the bottom corners. The evolution of the damage within the adhesive layer as a function of the model time is shown in Figure 3.98.

Figure 3.98 Evolution of the stiffness degradation damage parameter (SDEG) within the adhesive layer. Red color represents a fully damaged element, while blue represents an undamaged element.

(Continued)

155

156

3 Laboratorial Activities and Report Examples

(Continued) Taking into account the results of the numerical analysis, the joint can be safely designed with the conditions listed on Table 3.28. Table 3.28

Key joint design parameters.

Overlap length

Joint width

Adhesive thickness

50 mm

50 mm

0.2 mm

Surface treatment

Adhesive

Phosphoric acid anodization

Hybrid adhesive

One can also compare the obtained result with some of the analytical results presented above in Table 3.29. The result returned by the FEA is reasonably close to the analytical study of the yielding of the aluminum adherend as this is the actual mode of failure of the joint. Any criterion or analysis that assumes the failure of the adhesive will grossly overstate the predicted failure load as the adhesive layer is fundamentally undamaged. Table 3.29

Comparison of predicted failure load values.

Design brief

Global yielding of adhesive

Yielding of aluminum substrate

FEA predicted load

Failure load

14 kN

112.5 kN

24 kN

21 kN

Type of failure

n/a

Cohesive in adhesive

Adherend yielding

Adherend yielding

3.6.2.8 Design Validation

For this case study, a specimen was manufactured to experimentally validate the calculated loads. This joint is tested using a three-point bending experimental setup, under a load rate of 2 mm/min, considered as quasi-static load. The loading setup is shown in Figure 3.99. 3.6.2.8.1 Results Load displacement curves were registered from this test, as shown in Figure 3.100. These curves show that, although the yielding of the aluminum adherends defines the maximum failure load, the final failure was due to adhesive failure caused by the massive peel loads applied to the adhesive layer. This is shown in Figure 3.101. The model predicts the cohesive failure of the adhesive as it assumes perfect adhesion. In our case, the adhesive is much stronger than the adherend. More complex surface treatments might be required for exploiting the full energy absorption capability of the joint.

3.6 Case Study in Joint Design for a Structural Automotive Application

Figure 3.99 Threepoint bending setup used for testing the T-joint under analysis.

Substrate yielding @ 22.7 kN

25

Massive plastic deformation

Load (kN)

20 15 10

Failure of adhesive layer

5 0

0

2

4

6

8

10

12

Displacement (mm)

Figure 3.100

Three-point bending setup used for testing the T-joint under analysis.

(Continued)

157

158

3 Laboratorial Activities and Report Examples

(Continued)

Figure 3.101

Failure mode of the tested joint.

3.6.2.8.2 Comparison of Predicted and Experimental Loads A comparison between the predicted design loads and the experimental load is shown in Table 3.30. Table 3.30 load.

Comparison between the predicted design loads and the experimental

Design brief

Global yielding of adhesive

Yielding of lower substrate

FEA predicted load

Experimental load

Failure load

14 kN

112.5 kN

24 kN

21 kN

22.7 kN

Difference to exp.

—8.7 kN

+89.8 kN

+1.3 kN

–1.7 kN

n/a

These results show that we have designed a joint that successfully meets the target design load and that does so with and appreciable safety margin. Furthermore, we have seen that both the analytical study of the yielding of the lower adherend and the FEA predicted load are quite accurate, given that the peak load is predicted by the plastic deformation of the aluminum adherends. This shows that these two approaches are quite valid, although their use demands solid knowledge of the joint failure mechanisms and of the key properties.

3.6 Case Study in Joint Design for a Structural Automotive Application

3.6.2.8.3 Additional Test – Performance Comparison with Another Adhesive To conclude the validation tests, it was decided to create a new set of joints using a one-part epoxy adhesive (Henkel 5089). The results of this test are shown in Figure 3.102. In this case, the yielding of the aluminum adherends defines the failure load, being slightly lower because of the different adhesive layer stiffness. Slightly less stiffness

25

Nagase-Chemtex XNR6852 E-3

Load (kN)

20

Henkel 5089

15 10 Early failure of adhesive layer

5 0 0

2

4

6

8

10

12

Displacement (mm)

Figure 3.102 Comparison between the experimental load–displacement curves of obtained for joints bonded with XNR6852 E-3 and joints bonded with Henkel 5089.

As shown in Figure 3.103, the adhesive fails cohesively for much lower levels of adherend deformation, which indicates that the surface treatment is not an issue for this case.

Figure 3.103 adhesive.

Fracture surface of the alternative joint bonded with Henkel 5089

(Continued)

159

160

3 Laboratorial Activities and Report Examples

(Continued) 3.6.2.9 Design for Manufacturing

Finally, one will conclude this work by considering the manufacturability of the proposed joint. Design for manufacturing should be a transversal component of adhesive joint design. This process is composed of four different key points, summarized in Figure 3.104. Each one of these points will be discussed in detail in this final section. Fixture design and adhesive dispensing

Curing process

Quality control

Health and safety

3.6.2.9.1

Figure 3.104 Key design points in the design for manufacturing of bonded joints.

Fixture Design and Adhesive Dispensing Gap filling, thixotropic adhe-

sives systems such as epoxy considered for use in this work are very tolerant of geometrical variations in the adherends to be bonded. However, joint strength might be changed. The pressure should be applied uniformly on the whole surface of bonding, and point pressure clamping should be used carefully on bonding fixtures. Presses and autoclaves are ideal solutions as they provide very even pressure distribution. The loads applied by the bonding fixture should be set so as not to cause deformation of the adherends. Adhesive dispensing is strongly dependent on adhesive form and paste adhesives use pressurized guns (automated) or spatulas (manual). In this application, the adhesive is applied directly using a spatula and spread evenly through the joint. Excess adhesive is allowed to flow out freely. The bonding fixture should accommodate the free thermal expansion and contraction of the adherends. If heating elements are included, they should be designed to ensure uniform temperature distribution over the whole bonding area. An uneven temperature distribution will lock in thermal stress after curing. The tolerances of the fixture should be regularly controlled with coordinate measuring machines. The fixtures should be designed with adjustability in mind, featuring shims or other mechanisms, ensuring that distortions are not introduced in the joint. The fixture used to produce the joints studied in this work is shown in Figure 3.105.

3.6 Case Study in Joint Design for a Structural Automotive Application

Release agent Frekote NC-770

Aluminium substrates

Adjustable screws

Shims for adjusting adhesive thickness

Aluminium fixture

Figure 3.105 Fixture used to simultaneously produce two T-joints, showing the use of shims and adjustable screws for the definition of key geometrical parameters of the joint.

3.6.2.9.2 Adhesive Hardening Hardening of most structural adhesives takes

place as the result of a chemical reaction. For industrial application of structural adhesives, thermal curing is generally preferred if the adherends can withstand the heat. The cure can be performed using convection ovens or infrared ovens. These are slow but ensure uniform temperature distribution. In practice, adhesives for the automotive industry are designed to cure during the electrocoating (E-coat) process that provides corrosion resistance to the metallic car body frame. This process includes a stage between 150 and 190∘ C, and thus, many automotive industry adhesives are designed to cure within this specific temperature range. The selected adhesive (XNR6852E-3) is one of these cases and therefore does not require a specific oven or dedicated curing stage. Of course, in a laboratorial activity, it is necessary to create the necessary curing conditions, and in this case, the curing of T-joint specimens was performed in a hot plate press, loaded with the previously presented mold. This procedure is shown in Figure 3.106. Mould

Specimens

Lower heated plate

Figure 3.106

Curing of the T-joint specimens in a specialized hot plate press.

(Continued)

161

3 Laboratorial Activities and Report Examples

(Continued) To match the E-coat process, the curing cycle for the T-joints includes a 3 hour stage at 150 ∘ C, as shown in Figure 3.107. 3 hours, 2 bar 150

Figure 3.107 Curing cycle of the T-joint specimens in a specialized hot plate press.

T (°C)

162

0 Time

Because of the thermal differential present, it is fundamental to control local temperature of the adhesive layer. This control is typically performed with thermocouples. Curing pressure, if possible, to control, must be the minimum necessary to ensure removal of excess adhesive. However, if the adhesive thickness is controlled by shims or glass beads, excessive pressure can damage the adherend. If no method for controlling the adhesive thickness is present, most adhesives will flow out of the joint even with limited pressure application. 3.6.2.10 Quality Control Techniques

Multiple methods can be applied to our example joint after bonding to ensure the quality of the adhesion. These include visual inspection, tap test, ultrasonic testing, X-ray analysis, and acoustic emission. The pitch of the sound emitted during a tap test can enable the detection of voids and other defects in bonded joints. Ultrasound analysis, although similar in principle to the tap test, enables higher detail in the analysis and allows the distinction of several types of defects. X-Ray analysis enables the detection of voids and discontinuities, and the quality of the image can be improved by imbedding metallic dust in the adhesive. Acoustic emission can be considered as a semi-destructive test, as the joints must be mechanically loaded during test. Stress waves are caused by micro-crack propagation and recorded by piezoelectric transducers. It is thought to be the only method able to effectively detect disbonds (kissing bonds). Pull-off adhesion testers are especially suitable for large, bonded areas (panels). This is a semi-destructive testing procedure as it removes a portion of the sample. A pull-off pressure is locally applied on the surface of the bonded area and the response is recorded and analyzed. The use of this equipment might not be suitable for high-performance adhesives and thick adherends as is the case of the joint under analysis.

References

3.6.2.11 Health and Safety Concerns

Four principal safety factors must be taken into account handling adhesives: ● ● ● ●

Toxicity of the adhesives, Flammability of adhesives, Dangerous chemical incompatibilities that may cause harm, Risks associated with the dispensing and curing systems.

In the specific case of epoxies, the hardener component is usually the most hazardous. Amine-based hardeners are especially corrosive and toxic. Epoxy resins in uncured liquid or paste form are thus known to cause eye and skin irritation, mainly as the presence of these components. Epoxy residues are also harmful for aquatic environments, requiring extreme care in the disposal process. Also of note is the process of sensitization: If repeated skin contact with epoxies occurs, an allergic reaction will gradually develop. This type of reaction can occur several days after contact and will appear as dermatitis on the affected areas and is one of the major known risks associated to work with epoxy areas, increasing in intensity with further exposure. The risks inherent to adhesive joining with epoxy-based adhesives in the automotive industry can be mitigated by consulting safety datasheets to select adhesives and materials with low toxicity, defining manufacture processes that reduce dangerous operations, providing adequate personal protection to the workforce, and ensuring that all necessary technical equipment follows the relevant safety standards and is maintained in optimal condition. Finally, it is crucial to invest in constant training of the workforce, alerting to dangers of these materials.

References 1 da Silva, L.F.M., Öchsner, A., and Adams, R.D. (2018). Handbook of Adhesion Technology, 2e. Berlin Heidelberg: Springer-Verlag. 2 Rudawska, A. (2019). Surface Treatment in Bonding Technology. Academic Press. 3 Volkersen, O. (1938). Die Nietkraftverteilung in zugbeanspruchten Nietverbindungen mit konstanten Laschenquerschnitten. Luftfahrtforschung 15: 41. 4 Adams, R.D., Comyn, J., and Wake, W.C. (1997). Structural Adhesive Joints in Engineering, 2e. Londres: Chapman & Hall. 5 Goland, M. and Reissner, E. (1944). The stresses in cemented joints. J Appl. Mech. 11: A17. 6 Hu, F.Z. and Soutis, C. (2000). Strength prediction of patch-repaired CFRP laminates loaded in compression. Compos. Sci. Technol. 60 (7): 1103–1114. 7 Crocombe, A.D. (1989). Global yielding as a failure criterion for bonded joints. Int. J. Adhes. Adhes. 9: 145.

163

164

3 Laboratorial Activities and Report Examples

8 Adams, R.D. and Peppiatt, N.A. (1974). Stress analysis of adhesive bonded lap joints. J. Strain Anal. 9: 185. 9 Grant, L.D.R., Adams, R.D., and da Silva, L.F.M. (2009). Experimental and numerical analysis of single lap joints for the automotive industry. Int. J. Adhes. Adhes. 29: 405. 10 Hart-Smith L.J. (1973). Adhesive-bonded single-lap joints. NASA Contract Report, NASA CR-112236. 11 Goland, M. and Reissner, E. (1944). The stresses in cemented joints. J. Appl. Mech. 11 (1): A17–A27.

165

4 Essay and Multi-choice Questions 4.1 Essay Questions In this chapter, some essay questions on structural adhesive joints are proposed, accompanied with answers. It is recommended to read the book “Introduction to Adhesive Bonding” as an auxiliary material for the preparation of further questions. 1

An adhesive bonding process can be divided into five stages. Identify and briefly describe each of the steps.

2

Define the concept of pot life and shelf life. Explain the main differences between these two concepts.

3

Identify three families of adhesives that you are familiar with, identifying the main characteristics, curing process, and typical applications for each family.

4

Name three types of joints that you know of, identifying the industries that use them and present the main advantages and disadvantages.

5

The most widely used adhesive joint configuration is the single overlap joint, but these have large concentrations of stresses at the ends of the overlap length. Identify three techniques to reduce the stress concentration that occurs at the ends of the bondline.

6

Surface treatments are divided into active and passive. Identify two active and passive surface treatments and give examples of materials to which this treatment can be applied and the type of change that this treatment causes on the adherend.

7

The thickness of an adhesive is known to have a significant effect on adhesive joint performance. Explain how varying the thickness affects the strength of the joint and why this occurs.

Adhesive Bonding Technology and Testing, First Edition. Ricardo João Camilo Carbas, Eduardo André Sousa Marques, Alireza Akhavan-Safar, Ana Sofia Queirós Ferreira Barbosa, and Lucas Filipe Martins da Silva. © 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.

166

4 Essay and Multi-choice Questions

8

When designing a joint, it is essential to know the mechanical properties of the adhesive. Identify four of the most relevant properties in joint design, explain their significance, and how they can be determined.

9

Testing procedures for adhesive joints can be divided into two groups, please identify them. Give three examples of tests for each of the groups and briefly describe what each test consists of.

10

After performing any destructive testing procedure, it is important to assess what type of failure was obtained. Describe two types of post-fracture tests that you know of and describe briefly the main purpose of this analysis.

4.2 Multi-choice Questions This chapter presents multiple choice questions about the structural adhesive joints. Answers to these questions can be found at the end of this section. Only one answer is correct. 1

Adhesive bonding has many advantages and some disadvantages. Identify one advantage of using adhesive joints when compared to other joining methods such as the use of screws, rivets, or welding? A Do not require careful surface preparation B Consistent mechanical properties for a long period of time C Almost uniform distribution of stress D Consistent mechanical properties for a wide temperature range

2

The performance of an adhesive bonded can be quantified: A by making complex mathematical calculus to predict their behavior B by valuing their strength and their durability C using simple rules of threes based on similar previous cases D by exclusively determining their strength

3

Identify one key advantage of using adhesive joints: A good resistance to cleavage loading B good fatigue resistance C good resistance to pull-out loads D good resistance to high temperatures

4

Identify one disadvantage associated with the use of adhesive bonding: A poor resistance to temperature and humidity B poor fatigue strength

4.2 Multi-choice Questions

C limited resistance to shear stress D unsuitability for the connection of different materials 5

Adhesives are usually A polymeric materials B ceramic materials C metallic materials D shape memory materials

6

Adhesive technologies are used in several industries, including the aeronautical and automotive sectors. These industries use adhesive to A reduce the weight of the structure B increase the production cadence C join structures without the use of anchoring systems D shorten production times

7

Which equipment is mostly used to apply adhesives in industrial environments? A Spatula B Hand gun C Electric or pneumatic gun D Tube

8

An adhesive joint under service should be submitted mainly to A tensile loads B shear loads C peel loads D cleavage loads

9

The contact angle that the adhesive makes with the adherend to be bonded is 90∘ . This means that A the adhesion between the adhesive and the adherend will be poor B the cure of adhesion will be insufficient C the adhesion of the adhesive will be good D the bonded area will have the presence of air trapped in the adhesive

10

The aeronautical industry was the pioneer in the use of structural adhesives. The most used adhesives in these applications are epoxides, being chosen for the following reasons: A they are the cheapest adhesives B high stiffness and shear strength

167

168

4 Essay and Multi-choice Questions

C cure at low temperatures D high toughness 11

The construction sector uses structural adhesives to join different types of materials. When it is necessary to bond wood panel structures to form the base of a building, what type of adhesive would you choose for this application: A an rubber B a bismaleimide C a phenolic D an aromatic

12

The cure of adhesives via condensation: A leads to a high glass transition temperature B does not release water during curing C releases water during curing D occurs at cryogenic temperature

13

Joints bonded with phenolic adhesives exhibit a large amount of porosity. Identify one action to reduce this problem. A Increase the cure temperature B Increase the curing time C Increase the pressure during the curing process D Increase the pressure after the cure is completed

14

What factor influences the wettability of adherends by an adhesive? A The mechanical properties of the adherends B The shape of the adherends C The thickness of the adherends D The surface energy of the adherends

15

The stress–strain curve of an adhesive allows us to determine A the contact angle of the adhesive B the stiffness of the adhesive C the degree of cross-linking resulting from the adhesive cure process D the glass-transition temperature

16

To avoid premature failure by delamination when composite adherends are used in an adhesive joint, it is very important to A use a very rigid adhesive B know the transverse (thickness direction) strength of the composite

4.2 Multi-choice Questions

C reduce stress concentrations in the fiber direction D perform a careful selection of the surface treatment 17

Epoxy adhesives are commonly used to bond large areas in aeronautical structures. Which adhesive form is better suited for this type of application? A Paste B Film C Liquid D Solid

18

Consider the following joint geometries and assume that the loading is as indicated in the figure. Please identify which of the following statements is true:

Single lap

Double lap

Scarf

Bevel

Step

A the single-lap joint is the less strong joint B the outer chamfer joint (scarf) is the strongest joint C the inner chamfer joint (bevel) is the strongest joint D the step joint is the cheapest joint 19

The single-lap joint is known to be the most widely used joint geometry. When under load, a single-lap joint exhibits the maximum shear stress value: A at the extremities of the bonded area B in the middle of the bonded area

169

170

4 Essay and Multi-choice Questions

C is uniform through the overlap length D is null because there is no shear stress 20

If one desires to manufacture a joint with good resistance to peel, that will work at low temperatures, an adhesive thickness that can vary from 1 to 3 mm and should cure in less than 24 hours, which adhesive is better suited for this application? A an epoxide film B a moisture curing polyurethane C a two-component polyurethane D a cyanoacrylate

21

Large stress concentration on the overlap ends of joints can be reduced: A using a brittle adhesive B using an adhesive that cures with temperature C using a pressure-sensitive adhesive D using a functionally graded adhesive with higher stiffness in the middle of the overlap

22

What is a primer? A A product, applied to the adherend to improve the bonding performance B A coating that improves the finish and aesthetic of the pieces to be bonded C A process in which the weak layers of the adherend are eliminated D An agent that allows cleaning the contaminant surfaces

23

Dyne pens use a special ink that can be used for A the determination of the maximum viscosity of an adhesive suitable to bond adherends that have been tested with dyne pens B the determination of the surface energy of adherends, e.g. for reasons of quality or material control C the determination of the cohesion of adherends to ensure a sufficient rigidity of the bonded end product D labeling of test samples

24

Regarding adhesive viscosity, which statement is true? A Low-viscosity adhesives are less likely to penetrate a porous oxide layer than high-viscosity adhesives. B High-viscosity adhesives are more likely to penetrate a porous oxide layer than low-viscosity adhesives

4.2 Multi-choice Questions

C Low-viscosity adhesives are more likely to penetrate a porous oxide layer than high-viscosity adhesives D The penetration of a porous oxide layer is not related to the viscosity of the adhesive 25

What type of surface treatment would you recommend for bonding a joint with aluminum adherends, aiming to ensure a good durability? A Abrasion using sandpaper B Cleaning and degreasing with acetone C Anodizing D None of the above

26

Surface treatments can be classified as active or passive processes. Identify, from the following list, the process that is classified as active. A shot blasting B sanding C solvent degreasing D flame treatment

27

Flame treatments and corona discharge are physical treatments applied to polymers to A warm up the surface of the adherend before bonding B create free radicals that react rapidly with oxygen to form polar chemical groups C increase the cross-linking degree of the adhesive D melt the surface of the adherend

28

In order to ensure good wetting, it is crucial to ensure that the adhesion forces are superior to cohesion forces. This implies that A if the surface is correctly prepared, the failure will be cohesive B the failure is always cohesive in the adhesive if a weak layer exists on the adherend surface C the failure is always cohesive in the adherend D the failure is always mixed (adhesive and cohesive in adhesive)

29

For two-part adhesives, a proper mixing technique is very important because A it can change the color of the adhesive B it may increase the contact angle of the adhesive C it can avoid the need to use surface treatments D it can introduce porosity into the adhesive

171

172

4 Essay and Multi-choice Questions

30

To manufacture wind turbine blades, which type of adhesive would you use: A film adhesive B adhesive in liquid form C thixotropic adhesive D adhesive with a pot life of a few seconds

31

Adhesives that cure by condensation require special care during cure because they A have a high glass transition temperature B release water during curing C have a pot life of only a few seconds D cure at room temperature

32

The surface of steel adherends was prepared by three different processes: shot blasting, solvent, and chemical attack. The joint strength was identical in all the three cases. This indicates that A adhesive failure was obtained B cohesive failure on the adhesive was obtained C delamination in the adherend was obtained D the joint did not fail

33

Adhesives in the liquid form can be used to A bond sandwich panels B bond complex forms with small gaps between the parts to be bonded C bond large adhesive thicknesses D bond large components used in trains

34

Adhesive joints with a bondline thickness of 0.1 mm, created using a film adhesive were produced and, after testing, the failure strength was lower than expected despite being cohesive. These results can be explained due to A excessive adhesive porosity B poor surface preparation C inadequate adherend selection D excess adhesive on the edges of the overlap

35

From the following list, identify the equipment that can be used to perform a localized cure of a bonded area, typically used to repair a small damage in an airplane. A Heated blanket B Hot press machine C Convection oven D Autoclave oven

4.2 Multi-choice Questions

36

Identify the types of defects that can be detected using ultrasound analysis? A Inadequate surface treatment B Air bubbles trapped in the adhesive C Insufficient curing time D Adhesive discoloration

37

The effect of overlap length on composite adhesive joints (glass fiberreinforced epoxy) was evaluated experimentally. It was found that the failure strength is not proportional to the overlap length and that the failure mechanism is interlaminar in the composite. This indicates that A the adhesive is too stiff B the stress distribution is uniform along the overlap length C the surface treatment was not suitable D the adhesive is excessively tough

38

Consider a joint where good wetting was ensured, leading to a case where the adhesion forces are greater than the cohesion forces. This implies that A if the surface is well treated, the failure will be cohesive B the failures are always cohesive on the adhesive in the case of a weak layer on the surface of the solid C the failure mechanisms are always cohesive on the adherend D the failure mechanisms are always adhesive at the interface

39

When the contact angle of an adhesive with a metal adherend is low (e.g. 40∘ ), this indicates that there is A a good wettability B a bad wettability C no spreading of the adhesive D shape of the adhesive on the adherend is a sphere-like

40

What influences the wettability of adherends by an adhesive? A The surface energy of the adherends B The shape of the adherends C The thickness of the adherends D The cohesion (inner strength) of the adherends

41

Of the following statements, identify the one which is true: A We can classify as a “wettable” surface one with a low-surface energy B We can classify as a “wettable” surface one with a high-surface energy C Wettability is not related to the surface energy D All are wrong

173

174

4 Essay and Multi-choice Questions

42

An adhesive with low surface energy will be used to bond adherends with a high surface energy level. When load is applied to the joint until failure, which type of behavior do you anticipate being exhibited? A B C D

43

Identify the failure mode that is exclusive of composite joints: A B C D

44

cohesive failure in the adhesive adhesive failure interfacial failure none of the above

cohesive failure in the adhesive adhesive failure cohesive failure in the adherend delamination

Identify the main objective of undertaking surface preparation of adherends before the adhesive is applied. A decrease the bonding area B simplify the process of adhesive application C increase the adhesion of surfaces and the mechanical strength of the joints D reduce the cure time

45

In general, polymers are very difficult to bond, such as polyethylene and polypropylene. Identify, from the list, the technique that can be used to improve both wetting and adhesion: A using a detergent B increasing roughness by shot blasting C using a physical treatment such as flame or corona discharge D reducing roughness by sandpaper

46

Composite materials (polymeric materials reinforced with carbon or glass fibers) show low transverse strength (through the thickness). When composite adhesive joints are manufactured, delamination of the composite may occur. Delamination can be avoided A using stiff adhesive B using excess adhesive to reduce the presence of voids in the bondline C using an adhesive fillet D preparing the surface by plasma

47

Considering adhesive joints with aluminum adherends that will work for 30 years under high levels of humidity, identify the type of surface treatment that is more adequate for this application:

4.2 Multi-choice Questions

A shot blasting and cleaning with acetone B anodizing treatment C corona discharge treatment D ultrasound 48

When it is necessary to join two components made of polypropylene materials, which adhesive should be selected in order to dispense with surface preparation? A a modified acrylic B a bismaleimide C an epoxy D one silicone

49

A single-lap joint bonded with a stiff adhesive shows the maximum shear stress value in the adhesive layer: A near the edge of the bonded area B in the central part of the bonded area C there are no shear stresses D there is no peak, stresses are uniform along the length of the overlap

50

Which of the following statements is true? A Only the thermal properties of the adherend will influence the bond characteristics B The thermal properties of the adherend will influence the internal stresses of the bond C The thermal properties of the adherend do not influence the bond D Only the thermal properties of the adherend will influence the bond characteristics

51

An epoxy adhesive has a tensile strength of 30 MPa, a strain at failure of 5% at room temperature, and a glass transition temperature of 80 ∘ C. For which application is this adhesive suitable? A Panels used in aircraft engines B Manufacture of sandwich structures for car exterior panels C Bonding of car windshields D Panels used in spacecraft

52

The Volkersen analytical model allows for A the determination of the shear stresses in the adhesive B the determination of the peel stresses in the adhesive

175

176

4 Essay and Multi-choice Questions

C design of joints with ductile adherends and ductile adhesives D design of joints with composite adherends 53

Identify the false statement from the following list. The finite element method: A allows to determine the von Mises stress in the adhesive B allows to simulate more complex joints and materials than analytical models C allows designing joints with composites and ductile adhesives D does not require prior knowledge of the mechanical properties of materials

54

Identify the follow statement that is true. A Cleavage stresses are preferable to tensile stresses B Peel stresses exist in joints with rigid adherends C Peel stresses are caused by misalignment of the joint adherends D In a single-overlap joint, only shear forces exist

55

Identify the true statement from the following list. A Abrasion is sufficient to treat polymers B Plasma treatment increases the surface energy of composite materials C Acetone cleaning and degreasing is sufficient to treat the adherends superficially D An increased contact angle translates into increased wettability

56

Which of the following tests is dynamic in nature. A Creep test B Fatigue test C Diffusion test D Durability test

57

How can the toughness of an adhesive be determined? A Through torsion test without pre-crack B Through the test of a simple overlap joint C Fracture mechanics test such as the double cantilever beam (DCB) D Through a contact angle measurement test

58

Identify, from the following list of techniques, which is the one that ensure a good quality of bulk specimens using a knit supported film adhesive? A Hydrostatic pressure with a silicone rubber sealant frame B Application of low pressure in a hot plate press

4.2 Multi-choice Questions

C Using a metallic mold with metallic spacers to accurately control the thickness, cured inside an oven D Avoid pressure application and only use temperature 59

Which destructive test allows us to determine the behavior of a joint that will work at high temperatures and under a low but constant load? A Contact angle measurement test B Creep test C Water-break test D Crash test

60

Identify the test that better serves to evaluate the effect of strain rate on the strength of an adhesive joint? A Creep test B Hopkinson split pressure bar C Water-break test D Contact angle measurement test

61

The medical sector use bioadhesives because of A their lowest cost B their good compatibility with organic tissues C their slow curing processes D their suitability for automated implementation

62

Which of following factors is not relevant for the safe use of adhesives? A Adhesive toxicity B Chemical incompatibilities C Need for application equipment D Physical form

63

Which information cannot be typically found on adhesive data sheets? A Mechanical properties of the product B Instructions of use C Costs associated with the use D Safety precautions

64

What can we conclude when a contact angle of 0∘ is found for an adhesive applied over a metal surface? A There is perfect spreading of the adhesive B The adhesive does not wet the surface of the adherend C The shape of the adhesive on the adherend is sphere-like D None of the above

177

178

4 Essay and Multi-choice Questions

65

The toughness of the adhesive in mode II (shear) can be measured by A a torsion test using specimens without pre-crack B a test of single-overlap joint C a fracture mechanics test such as End Notched Flexure (ENF) D a water-break test

66

The finite element method allows A to obtain strength predictions with simple calculations B to determine the stress distribution for any geometry C to deal with the problem of singularities D to predict mechanical strength without knowing the mechanical properties of materials

67

Identify the true statement? A Only the thermal properties of the adherend influence the performance of the joint. B The thermal properties of the adherend do not influence the joint stresses C The thermal properties of the adherend will influence the joint stresses D Only the thermal properties of the adherend will influence the joint performance

68

A fracture mechanics-based approach allows to predict A the minimal allowable crack length for the fatigue life of adhesive joint B the minimum number of cycles in fatigue loading to be sustained by adhesive joint C the maximum allowable crack length for the fatigue life of adhesive joint D the maximum number of cycles in fatigue loading to be sustained by adhesive joint

69

Please consider the four joints shown below, assuming the loading indicated in the figure. In this case,

(a)

(b)

(c)

(d)

4.2 Multi-choice Questions

A joint (c) is the less resistant B joint (d) is the most expensive C joint (b) is the cheapest D none of the above 70

Porosities were identified in a joint bonded with a rigid brittle adhesive using ultrasound. However, after experimental testing, the joint strength was similar to that of defect-free joints. How do you explain this behavior? A The porosities are located at the edge of the overlap B The porosities are located in the middle of the overlap C The porosities are located at the interface D The porosities are uniformly distributed

71

Which non-destructive method allows us to identify poor adhesion in an adhesive joint? A Thermal method B Ultrasound C Test with calibrated liquids D None of the above

72

Structural adhesives have tensile strengths in the order of A B C D

73

0.7 MPa 7 MPa 70 MPa 700 MPa

Which test allows us to determine the stress–strain curve of an adhesive? A Double cantilever beam (DCB) toughness test B Tensile test C Wedge peel test D End-notched flexure (ENF) toughness test

74

The finite element method is the most effective method to analyze A a single-lap joint with a ductile adhesive B a single-lap joint with thickness variation and a fillet C a T joint with treated steel adherends D a single-lap joint with treated steel adherends

75

The Hart-Smith analytical model can be used to A to determine the water uptake during the time B to determine the crack path

179

180

4 Essay and Multi-choice Questions

C to design joints with elastic adherends and ductile adhesives D to determine the number of cycles sustained by a joint under fatigue loads 76

The effect of overlap length in single-lap joints was evaluated and it was found that the failure load is proportional to the overlap length and the failure was cohesive in the adhesive. This means that A the adhesive is excessively stiff B the generalized yield criterion can be used to design the joint C the adherend yield criterion can be used D The adhesive was cured at high temperature

77

A double overlap adhesive joint had adhesive failure and obtained a failure load of 20 kN. The surface preparation was changed and the failure was now cohesive. The failure load was A 5 kN B 10 kN C 12.5 kN D 25 kN

78

Adhesive joining technology is the preferential joining technique for composite structures, but because of the low transverse strength (through thickness) of the composite, delamination may occur. From the following five different solutions proposed, identify the configuration that provides the highest joint strength. Steel Composite Reference Steel Composite Outside taper Steel Composite Inside taper Steel Composite Adhesive fillet Steel Composite Inside taper and adhesive fillet

4.2 Multi-choice Questions

A Outside taper B Inside taper C Adhesive fillet D Inside taper and adhesive fillet 79

Experimentally obtained joint strength values are usually much lower than the theoretical joint strength values predicted by adhesion theory because A the bonded area, in practice, is much larger B the stress distribution is not uniform along the length of overlap C there is always a strong layer at the interface D there is always a weak layer at the interface

80

Considering the following four joints and assuming that the load applied is indicated in the figure:

(a)

(c)

(b)

(d)

A joint (a) is the most difficult to manufacture B joint (c) is the strongest C joint (d) is the cheapest D joint (d) is the strongest 81

A single-overlap joint with treated steel adherends was tested, reaching a strength of 30 MPa. The same joint type and geometry were repeated with aluminum adherends instead of treated steel and a joint strength of 15 MPa was obtained. In both cases, failure was cohesive in the adhesive. In order to better understand these results, the adhesive subjected to pure and uniform shear and a strength of 35 MPa was obtained. What is the reason for the lower joint strength obtained with the aluminum joint? A Plastic deformation of aluminum has occurred B The adhesive has exceeded its storage time C Poor surface preparation D None of the above

82

The use of non-destructive tests allows to A evaluate the presence of voids on the bonded area B determine the cure temperature used C determine the joint strength D determine the wettability of the adherends

181

182

4 Essay and Multi-choice Questions

83

An adhesive joint shows a lack of adhesive on the edges and high porosity. What type of action that can be performed to solve this problem? A Assessment of the joint using a non-destructive test B Evaluation of the joint using a destructive test C Application of larger adhesive quantity and automatic adhesive mixing D Increase temperature and curing time

84

An adhesive joint was created with a thickness of 1 mm using an adhesive in liquid form. After testing, the joint strength was lower than predicted, although cohesive in the adhesive. What can be the cause for this result? A Existence of voids in the bonded area B Excessive adhesive at the edges of the overlap C The surface treatment was excessive D Adhesive wets the surface of the adherend too well

85

A thermal method was used to detect voids in an adhesive joint. However, after the joint was tested, an increased joint strength was obtained. Identify the cause for this increase. A Elimination of air bubbles B Post-curing of the adhesive because of the inspection method used C Adhesive degradation D Insufficient adhesive curing

86

How can the thermal degradation of adhesives be evaluated? A By loss of mass of adhesive at high temperatures B By ignition of adhesive at high temperatures C By gain of mass from the adherend at high temperatures D With higher shear strength at high temperatures

87

It is most appropriate to use a finite element model to predict the strength in a A double overlap joint with ductile adhesive B simple overlap joint with brittle adhesive C simple overlap joint with constant thickness and without fillet D simple overlap joint with thickness variation and a fillet

88

The importance of surface preparation before the adhesive is applied is important because it A simplifies the adhesive application process B decreases the bonded area

4.2 Multi-choice Questions

C increases the adhesion of surfaces and the mechanical strength of bonding D decreases the surface energy of adherends 89

Comparing the fatigue behavior of an adhesive joint and a riveted joint, one can state that A the joints have a similar behavior B the adhesive bond has worse fatigue behavior C the riveted connection has the best fatigue behavior D adhesive bonding has the best fatigue behavior

90

To manufacture joints with a bondline thickness of 3 mm and that will cure vertically, what type of adhesive should be used? A An adhesive supplied in liquid form B A thixotropic adhesive C An adhesive that cures in 24 hours D An adhesive with a reduced pot life

91

Identify which of the following is a typical effect associated with the presence of moisture in adhesive joints: A increased deformation speed B decrease of the Young’s modulus C increased stiffness D none, as humidity does not change the adhesive properties or behavior

92

An adhesive joint subjected to fatigue loads shows A decreased thickness of bonded area B a degradation of joint performance C a decrease of glass transition temperature D an improvement of mechanical properties over time

93

Which of the following factors is less important for the safe use of adhesives? A Existence of dangerous incompatibilities B Toxicity of adhesives C Flammability of adhesives D Adhesive color

94

Which of the following adhesives is not considered to be a structural adhesive? A Phenolic B Acrylic

183

184

4 Essay and Multi-choice Questions

C Hotmelt D Epoxy 95

Which of the following personal protective equipment (PPE) is essential for the safe handling of adhesives? A Steel-toed boots B Gloves C Leather apron D Helmet with visor

96

At temperatures well below the glass transition temperature, cured adhesives A show large plastic deformations B suffer decomposition C are fragile D become viscous

97

Which type of failure mechanism should occur after successful design of an adhesive joint? A B C D

98

Interfacial failure Adhesive failure Cohesive failure Mixed failure

The thermal method is suitable to detect the presence of defects in adhesive joint such as the A presence of porosity in bondline B presence of voids C bad adhesion D crack in the fillet

99

After visual inspection of the failure surface of an adhesive joint with aluminum adherends revealed that the failure was close to the interface, the failure surfaces were analyzed using a scanning electron microscopy analysis. This identified zones with only carbon. Knowing that the surface treatment consisted of sandblasting, what type of treatment that should be adopted for the production of joints? A Anodizing B Chemical etching C Clean with detergent D Maintain the same treatment

4.2 Multi-choice Questions

100

For some levels of absorbed moisture, an increase of adhesive joint strength is noticeable, and this occurs because of A adhesive plasticization. B an attack only on adherend. C adhesive hydrolysis D an attack on the adhesive–adherend interface

185

187

Solutions Essay Questions – Example Answers 1. Answer: The main five stages are as follows: ● Adhesive selection – The adhesive is selected as a function of design requirements, taking into account the role of the adhesive, its chemical composition, hardening mechanism, physical form, cost, the substrates that will be used, and the method of application. ● Joint design – Adhesive joints could be designed analytically or numerically, allowing to predict their mechanical behavior under load. ● Surface treatment – This stage is critical because the surface treatments must be properly selected in order to ensure a good adhesion between the adherend and the adhesive, ultimately seeking cohesive failure. ● Joint fabrication – This stage groups several steps that must be properly controlled, such as storage, mixing, adhesive application, fixturing of the parts, and hardening. ● Control – This stage can be divided in two types of tests, destructive and non-destructive. In the destructive tests, the joints are tested until failure, and this is a method for comparing materials and processes that are being evaluated. In the non-destructive test, joint strength can be correlated with some physical, chemical, or other parameters that can be measured without causing damage. 2. Answer: Pot life is a period of time, after mixture, where a two-part adhesive doubles its viscosity (typically, this can last between a few minutes and one hour depending on the adhesive family). The shelf life is the period starting from the date of production, where the manufacturer ensures that the adhesive is adequate for use. This requires the storage condition to be well controlled. A shelf life can vary from six months to two years. In sum, the pot life is the time available to use the adhesive after the mixture and the shelf life is the allowable storage time, ensuring that the mechanical properties of the adhesive remain constant within this period.

Adhesive Bonding Technology and Testing, First Edition. Ricardo João Camilo Carbas, Eduardo André Sousa Marques, Alireza Akhavan-Safar, Ana Sofia Queirós Ferreira Barbosa, and Lucas Filipe Martins da Silva. © 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.

188

Solutions

3. Answer: The three families are as follows: Epoxies – Good strength and durability. Cure: – A one-part epoxy adhesive cures at high temperatures, typically at 150∘ C; – two-part epoxy adhesives can cure at room temperature or this process can be accelerated with temperature. Applications: These adhesives can be used in aircraft, helicopters, cars, trains, sport equipment, among many other applications. Polyurethanes – Show good strength at low temperatures, good toughness, and wetting ability. Cure: – A one-part polyurethane adhesive cures at room temperature and with moisture; – the two-part adhesives cure at room temperature, or this process can be accelerated with temperature. Application: Cryogenic applications and automotive industry (for example bonding glass to vehicle body. Phenolics – Show excellent resistance to environment and temperature. Cure: – With temperature and pressure. Application: Wood products and metallic structures (hybrid phenolics). 4. Answer: The three types of joints are as follows: Single-lap joints – This is the type of joint that is the most commonly used and can be found in cars to join different components. T joints – These joints are highly used in the aerospace industry and can be found to reinforce the external panels and wing sections of airplanes. Tubular joints – These joints can be used in refrigeration circuits to circulate refrigeration liquids inside the tubes. 5. Answer: The stress concentration of the overlap length at the edges can be reduced using adherend shaping (which controls the local stiffness of the adherend), mixed adhesive joints (which introduce a more flexible adhesive at the edges of the overlap length), and functionally graded joints (which gradually vary the stiffness of the adhesive along the overlap). 6. Answer: The two passive treatments are sanding and grit blasting. These treatments can be used in metals and the main purpose is to remove contaminants and weak layers and increase the roughness of materials. The two active treatments are as follows: chemical treatment (can be used in non-ferrous metals and their main purpose is to create a strong and stable oxide in the surfaces) and physical treatments (can be used in polymers and composites and its main purpose is to change the chemical structure of substrate surface by exposing it to highly energetic charges. 7. Answer: An adhesive thickness increase can lead to an increase of voids within the adhesive, and this reduces the joints strength. For certain joints, such as single-lap joint, the increase of adhesive thickness further increases the

Solutions

misalignment between adherends and that will lead to an increase of bending moment at the ends of overlap length. 8. Answer: Four of the most important properties are as follows: Young’s modulus of the adhesive – The stiffness of the adhesive layer can be determined through a tensile test on bulk specimens. Yield strength of the adhesive – The stress that the adhesive can reach before plastic deformation starts can be determined using a tensile test on bulk specimens. Fracture toughness of the adhesive – The resistance an adhesive offers to the propagation of a crack and damage can be determined using fracture mechanics tests such as the DCB test. Glass transition temperature of the adhesive – Is the temperature at which the transition between the glass-like rigid solid to a more flexible, rubbery compound occurs. This can be determined using a DSC or DMA devices. 9. Answer: The two main groups are as follows: Destructive tests: ⚬ Creep test – Measurement of the behavior of material or joints under constant load/stress during long time periods. ⚬ Fatigue test – Measurement of the behavior of material or joints under cyclic loads. ⚬ Environmental test – Determination of the rate at which an adhesive joint will lose strength because of environmental factors equivalent to those found in service. Non-destructive tests: ⚬ Tap test – Gently hitting the adhesive joint with a light hammer and, as a function of tone of the generated sound, identify good/poor adhesion and voids or unattached area. ⚬ Ultrasonic method – Applications of an ultrasonic pulse through the joint, allowing us to identify defects within the adhesive layer. ⚬ Thermal method – Heating one surface of a bonded structure and observing the temperature rise of the opposite face. Areas where disbonds exist, which resist the transfer of heat, will show as cool areas. 10. Answer: Two post-fracture tests are as follows: Optical microscopy – This analysis allows us to observe, with high resolution, the failure mechanisms that have occurred, as well as the presence of microscopic defects in bonded areas. Scanning electron microscopy – This analysis allows us to determine the failure surfaces with the highest resolution possible and can also allow us to identify the chemical composition of the materials present in the fracture surface – This process is the most powerful at determining if the failure mechanisms are cohesive or adhesive.

189

190

Solutions

Multi-choice Questions – Solutions 1.

(c)

21.

(c)

41.

(b)

61.

(b)

81.

(a)

2.

(b)

22.

(a)

42.

(a)

62.

(d)

82.

(a)

3.

(b)

23.

(b)

43.

(d)

63.

(c)

83.

(c)

4.

(a)

24.

(c)

44.

(c)

64.

(a)

84.

(a) (b)

5.

(a)

25.

(c)

45.

(c)

65.

(c)

85.

6.

(a)

26.

(d)

46.

(c)

66.

(b)

86.

(a)

7.

(c)

27.

(b)

47.

(b)

67.

(c)

87.

(d)

8.

(b)

28.

(a)

48.

(a)

68.

(c)

88.

(c)

9.

(a)

29.

(d)

49.

(a)

69.

(d)

89.

(d)

10.

(b)

30.

(c)

50.

(b)

70.

(b)

90.

(b)

11.

(c)

31.

(b)

51.

(b)

71.

(d)

91.

(b)

12.

(c)

32.

(b)

52.

(a)

72.

(b)

92.

(b)

13.

(c)

33.

(b)

53.

(d)

73.

(b)

93.

(d)

14.

(d)

34.

(a)

54.

(c)

74.

(b)

94.

(c)

15.

(b)

35.

(a)

55.

(b)

75.

(c)

95.

(b)

16.

(b)

36.

(b)

56.

(b)

76.

(b)

96.

(c)

17.

(b)

37.

(a)

57.

(c)

77.

(d)

97.

(a)

18.

(a)

38.

(a)

58.

(a)

78.

(d)

98.

(c)

19.

(a)

39.

(a)

59.

(b)

79.

(b)

99.

(d)

20.

(c)

40.

(a)

60.

(b)

80.

(d)

100.

(a)

191

Index a ABAQUS/CAE 118–120, 122–123, 125, 132, 139 adhesive joints 172 failure analysis 19–24 loading mode importance 1 stress distribution 12 surface treatments and methods 6 adhesively bonded joint 79 adhesive thickness 97–104 experimental details 100–101 experimental work 99 materials 99 report 99–104 work description 98 experimental result fracture surfaces 91–94 tensile tests 90–91 experimental results and discussion 72–79 adhesive joint tests 76–77 contact angle measurement 73–74 dyne pens 74–75 failure mechanism 77–78 experimental work 67–68, 84 failure loads prediction 86–97 adherend yield 72, 89 designer software 89 Volkersen model 87–89 yield 72, 87 joint assembly 71 manufacture 67

materials 67, 71, 83–84, 86 report 68–79, 84–97 strength and failure mechanism 104–117 characterization 107–112 experimental results comparison 115–116 experimental work 105–106 failure load theoretical prediction 112–115 manufacture and determine 105 materials 105 report 106–117 strength prediction 66–67 surface treatment 65, 71 tensile tests 71 work description 83 adhesive thickness 97–104 experimental details adherend 100 adhesive 101 failure surfaces 103 geometry 101 predictions 101–102 adhesive viscosity 170–171 aeronautical industry 167–168 aluminium adherends 171 automotive industry 139–163 adherend dimensions 142 adherend material selection 145 adhesive selection 142–145 advantages 141

Adhesive Bonding Technology and Testing, First Edition. Ricardo João Camilo Carbas, Eduardo André Sousa Marques, Alireza Akhavan-Safar, Ana Sofia Queirós Ferreira Barbosa, and Lucas Filipe Martins da Silva. © 2023 WILEY-VCH GmbH. Published 2023 by WILEY-VCH GmbH.

192

Index

design brief 140–142 disadvantages 141 environmental conditions 142 health and safety concerns 163 intended application 143–144 loading conditions 141–142 report 139 surface treatment adherend yielding 151–152 adherend yielding criterion 151 adhesive generalized yielding criterion 150 cohesive zone modelling 152–153 design validation 156–159 durability prediction 147–148 identification 145–146 joint design 148–149 manufacturing design 160–162 mechanical characterization 146–147 numerical models 152–156 quality control techniques 162 surface treatment selection 145

b bulk specimens 25–35 adhesive application 30–31 adhesive pouring technique 25–26 closed mould 31–33 machining procedure 33–34 metallic mould 26–30 testing procedure 35

c cohesive zone modelling 152–153 analysis step 129–131 calculate 134–136 job manager dialog box 136 load module 131–132 meshing 132–133 module/part 122–124 property 124–128 visualization 137–139

d Double Cantilever Beam (DCB) 44–45, 47–48, 50, 52, 54–55, 60, 147, 189 dyne pens 7, 9–12, 71, 74–75, 170

e End Notched Flexure (ENF) 44–45, 48, 50, 52–55 epoxies 98, 143–144, 163, 188 epoxy adhesives 83, 128, 159, 169, 175, 188

f failure load prediction 68, 72, 84, 94, 96, 100, 106, 113, 115 failure surface 103, 109–110, 184, 189 finite element method 176, 178, 179 flame treatments 65, 171 force-displacement curves 108–110 fracture mechanics specimens 44–55 adhesive spacers 50 data reduction schemes 53–55 final preparation 51–52 manufacture 50–51 metallic mould 48–49 surface treatment 49–50 testing procedure 52–53

g generalized yield criterion 94, 97, 112, 114–116 geometry 33, 36, 56, 61–62, 86–87, 89, 101, 122, 124, 139, 141–142, 149, 169–170, 178, 181 Goland model 82, 102, 104–106, 112–113, 116–117

h hardening 104, 161–162, 187 Hart–Smith analytical model 179–180 hybrid adhesives 143

j joint design 97, 139–163, 166, 187

Index

l loading mode

1–6, 36

m maximum flexural stress 82 model configuration 153–154

n non-destructive method 179 numerical models 98, 139, 149, 152, 155–156

p polyethylene 67, 72, 174 polypropylene 174–175 polyurethanes 57, 143, 188 primer 39, 57, 70, 143, 170

q Quads damage initiation criterion 126

r Reissner model 104, 112, 113, 116

stress-strain curve 145, 147, 168, 179 structural adhesive joints 165, 166 structural adhesives 98, 104, 142, 146, 149, 161, 165–168, 183 surface energy 6–12, 65, 69, 70, 73–79, 145, 168, 170, 173, 174, 176, 183 surface treatments 6–12, 49, 56–58, 65, 69–71, 73, 145, 156, 165, 171, 187

t tensile loads 12, 35, 167 tensile separation laws 124–125 tensile tests 35, 71, 77, 90 thick adherend shear test (TAST) 35 final specimen preparation 42 geometrical control 40–41 metallic mould 36–37 specimen manufacture 41–42 surface treatment of adherends 37–40 testing procedure 42–44 thixotropic adhesives systems 160 traction separation laws 124–129 two-part acrylics 143

s single lap joint (SLJ) 56, 66, 169 final preparation 60 manufacture 57–60 surface treatment 56–57 testing procedure 60–63 steel adherends 83, 93, 105, 117, 172, 179, 181 stress distribution 12–14, 44, 79, 85, 87, 91, 98, 106, 112, 114, 141, 144, 173, 178, 181

u ultrasound analysis

162, 173

v Volkersen analytical model 79, 81, 84, 87–89, 94–96, 98, 99, 112, 175

w wettability 8, 11, 168, 173, 176, 181 wind turbine blades 172

193

WILEY END USER LICENSE AGREEMENT Go to www.wiley.com/go/eula to access Wiley’s ebook EULA.