Perspectives of Nanoscience and Nanotechnology [1 ed.] 9783038132011, 9783908451570

This collection highlights the frontiers of current research in nanotechnology. The scope of the work, Perspectives of N

134 79 47MB

English Pages 211 Year 2008

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Perspectives of Nanoscience and Nanotechnology [1 ed.]
 9783038132011, 9783908451570

Citation preview

Perspectives of nanoscience and nanotechnology

Acta Materialia Gold Medal Workshop

Perspectives of nanoscience and nanotechnology Acta Materialia Gold Medal Workshop

Selected, peer reviewed papers from the European Materials Research Society, Fall Meeting, Warsaw University of Technology, 17th – 21st September, 2007

Edited by

Witold Łojkowski and John R. Blizzard

TRANS TECH PUBLICATIONS LTD Switzerland • UK • USA

Copyright  2008 Trans Tech Publications Ltd, Switzerland

All rights reserved. No part of the contents of this book may be reproduced or transmitted in any form or by any means without the written permission of the publisher.

Trans Tech Publications Ltd Laubisrutistr. 24 CH-8712 Stafa-Zurich Switzerland http://www.ttp.net

Volume 140 of Solid State Phenomena ISSN 1012-0394 (Pt. B of Diffusion and Defect Data - Solid State Data (ISSN 0377-6883)) Full text available online at http://www.scientific.net

Distributed worldwide by

and in the Americas by

Trans Tech Publications Ltd Laubisrutistr. 24 CH-8712 Stafa-Zurich Switzerland

Trans Tech Publications Inc. PO Box 699, May Street Enfield, NH 03748 USA

Fax: +41 (44) 922 10 33 e-mail: [email protected]

Phone: +1 (603) 632-7377 Fax: +1 (603) 632-5611 e-mail: [email protected]

THE EUROPEAN MATERIALS RESEARCH SOCIETY The European Materials Research Society (E-MRS) is a non-profit making scientific society founded in 1983 with its Headquarters in Strasbourg, France. The Society focuses its activities on creating synergy from the interaction between interdisciplinary, innovative fields of materials research. Great attention is given to disseminating and exchanging information and promoting technology transfer from public institution to industry. The primary objective of E-MRS is to promote and enhance the efficiency of research in Europe in the field of Advanced Materials. E-MRS seeks to quickly inform researchers about the scientific and technological developments in their fields of interest, taking advantage of the Society’s links with other Materials Research Societies belonging to the International Union (IUMRS). Since 2002, in parallel with the Spring Meeting held in Strasbourg, Fall Meetings have been held in Warsaw, The 2006 Fall Meeting consisted of 10 symposia running concurrently, Plenary Sessions, a joint session with other symposia and included an exhibition of relevant products and services as well as training activities for young researchers and scientists. E-MRS General Secretary

P. Siffert E-MRS BP 20 67037 Strasbourg Cedex 2 France

Telephone: Fax: e-mail:

+33 3 88 10 65 43 +33 3 88 10 62 55 +33 3 88 10 62 93 [email protected]

THE E-MRS FALL MEETING CHAIRPERSONS Krzysztof J. Kurzydłowski, Warsaw University of Technology, Warsaw. Małgorzata Lewandowska, Warsaw University of Technology, Warsaw. Witold Łojkowski,

Institute of High Pressure Physics, Polish Academy of Sciences, Warsaw.

Andrzej Mycielski,

Institute of Physics, Polish Academy of Sciences, Warsaw.

Paul Siffert,

European Materials Research Society, Strasbourg, France.

E-MRS FALL MEETING 2007 CONFERENCE ORGANISERS Polish Materials Science Society ul. Wołoska 141 02-507 Warsaw Poland

Telephone: +48 (0)22 234 84 41 Fax: +48 (0)22 848 4947 e-mail: [email protected]

Institute of High Pressure Physics, Polish Academy of Sciences ul. Sokołowska 29/37 01-142 Warsaw Poland

Telephone: +48 (0)22 632 50 10 Fax: +48 (0)22 632 42 18 e-mail: [email protected]

Institute of Physics, Polish Academy of Sciences Al. Lotników 32/46 02-668 Warsaw Poland

Telephone: +48 (0)22 843 66 01 Fax: +48 (0)22 843 09 26 e-mail: [email protected]

Faculty of Materials Science and Engineering Warsaw University of Technology ul. Wołoska 141 02-507 Warsaw Poland

Telephone: +48 (0)22 849 99 29 Fax: +48 (0)22 234 85 14 e-mail: [email protected]

E-MRS: European Materials Research Society, Telephone: +33 3 88 10 65 43 BP 20 +33 3 88 10 62 55 67037 Strasbourg Cedex 2 Fax: +33 3 88 10 62 93 France e-mail: [email protected]

Local Organising Committee J.R. Blizzard H. Garbacz M. Lewandowska R. Pielaszek A. Rytel A. Szadkowski

E-MRS European Coordination Group A. Slaoui F. Priolo P. Siffert

EMRS President EMRS Vice-President EMRS General Secretary

ACKNOWLEDGEMENTS The organisers of the E-MRS Fall Meeting 2007 acknowledge the support given by:

EUROPEAN MATERIALS RESEARCH SOCIETY

Polish Ministry of Science and Higher Education The Polish Materials Science Society Materials Design Division of the Faculty of Materials Science, Warsaw University of Technology

Institute of Physics, Polish Academy of Sciences Institute of High Pressure Physics “UNIPRESS” Polish Academy of Sciences ESTECO Elsevier

Sponsors of the Acta Materialia Gold Medal Workshop 

The International Network of Nano and Micro Technology “Namic”, Poland



Polish Ministry of Science and Higher Education



Elsevier.



Institute of Nanotechnology, FZK, Karlsruhe, Germany



Institute of High Pressure Physics, PAS, Warsaw, Poland

E-MRS FALL MEETING 2007 Acta Materialia Gold Medal Workshop Introduction The Board of Governors of Acta Materialia took the decision to award the 2007 Acta Materialia Gold Medal to Prof. Herbert Gleiter. Traditionally, the recipient of the medal has the right to decide where the award ceremony, which is combined with an Award Lecture, is to be held. Prof. Herbert Gleiter suggested that the ceremony should take place in Warsaw. When this became known to the organisers of the E-MRS Fall Meeting, they reacted quickly and invited Prof. Gleiter to make his Award Lecture as a Plenary Session Presentation at the Fall Meeting. The Fall Meeting organisers also considered that the occasion provided an opportunity to highlight the frontiers of research in nanotechnology, and therefore this special Acta Materialia Gold Medal Workshop was organised as one of the events at the Fall Meeting. The scope of the workshop, Perspectives of Nanoscience and Nanotechnology, focussed on the presentation of the current status of the science and technology in various fields, examples of industrial and commercial applications and other information that could facilitate decisions and future directions of research. On being awarded the Acta Materialia Gold Medal Professor Gleiter delivered his Plenary Presentation which he entitled ‘Our thoughts are ours, their ends none of our own – Are there ways to synthesise materials beyond the limitations of today?’ John R. Blizzard Witold Lojkowski

Acta Materialia Gold Medal Workshop Organisers Witold Łojkowski, Institute of High Pressure Physics, Polish Academy of Sciences, Warsaw, Poland. [email protected] Krzysztof J. Kurzydłowski, Faculty of Materials Science, Warsaw University of Technology, Warsaw, Poland. [email protected] Ottilia Saxl, Institute of Nanotechnology, Stirling, FK9 4NF, United Kingdom [email protected] Hans-Jörg Fecht, Ulm University, Ulm 89081 Germany and Forschungszentrum Karlsruhe, , Germany. [email protected]

Our thoughts are ours, their ends none of our own – Are there ways to synthesise materials beyond the limitations of today? Professor Herbert Gleiter Acta Materialia Gold Medal 2007 The methods available today to modify the structure and properties of crystalline materials may be divided into the following two groups: modifications arising from the introduction of lattice defects and modifications arising from alloying of two or more components. In materials with grain sizes of 1mm or more, the introduction of lattice defects modifies the microstructure. However, the modifications of the atomic structure are limited to less than 1 volume percent of the material. The way to modify the atomic structure by up to 50 volume percent of a material by introducing defects was opened by reducing the crystal size of polycrystalline materials to a few nanometers. Materials of this kind are called nanocrystalline, or nanostructured materials. The step towards modifying the entire atomic structure of solid materials seems to be possible by means of naonoglasses. Nanoglasses are glasses that are generated by consolidating nanometer-sized glassy spheres at high pressure, of several GPs. The existing structural investigations on metallic nanoglasses as well as studies by means of molecular dynamics suggest that nanoglasses consist – in the as prepared state – of the following two structural components. Glassy regions – resulting from the consolidation of spheres – and interfaces between the glassy regions. In these glassy/glass interfaces, the free volume is enhanced and nearest neighbour co-ordination deviate from the ones in the glassy regions. If these nanoglasses are annealed, the enhanced free volume in the glass/glass interfaces seems to delocalize and, thus, modifies the atomic structure of the entire material. In fact it is found that, after long annealing time, nanoglasses consist of a surface region with an enhanced density (due to a high hydrostatic pressure) and a glassy core region with a significantly – up to 10% - reduced density. In other words, nanoglasses may pave the way to tune the free volume, density, of glasses at constance chemical composition. The modifications of solid materials by alloying may be divided into the following two groups: components that can be alloyed by melting followed by solidification, and alloys of components that are immiscible in the solid state, e.g. alloys of metals and ionics such as Au-NaCl. The preparation of alloys of this type seems attractive because they are likely to exhibit new properties. So far, apparently two approaches have been considered for preparing such alloys, In the first approach, applicable to systems with mobile charge carriers, electronic screening effects at interface boundaries are utilized, If nanocomposites of immiscible components are prepared with a crystal size comparable to the electronic screening length, the electronic structure of the entire specimen is modified due to the screening effects. As has been shown, this modification may result in the formation of solid solutions of conventionally immiscible components, eg. of Ag and Fe. In systems without mobile charge carriers, vapour deposition of ions of one of the components onto an electrically charged substrate may be used to generate solid solutions.

E-MRS Fall Meeting 2007 Participants in the Acta Materialia Gold Medal Workshop

Adamus, Zbyszek

Polish Academy of Sciences, Institute of Physics, al. Lotników 32/46, 02-668 Warsaw, Poland.

Ban, Irena

FKKT, Smetanova 17, Maribor 2000, Slovenia.

Bedis, Hanene

Faculté de Sciences Mathématiques, Physiques et Naturelles de Tunis, Campus Universitaire Tunis 1060, Tunisia.

Bieńkowski, Krzysztof

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warsaw, 01-142, Poland.

Celichowski, Grzegorz.

University of Łódź, Department of Chemical Technology and Environmental Protection, Pomorska 163, 90-236 Łódź, Poland.

Cieplak, Marta Z..

Polish Academy of Sciences, Institute of Physics, al. Lotników 32/46, 02-668 Warsaw, Poland.

Dietl, Tomasz

Polish Academy of Sciences, Institute of Physics, al. Lotników 32/46, 02-668 Warsaw, Poland.

Dommann, Alex

Cente Suisse d’Electronique et de Microtechnique (CSEM), Jaquet-Droz 1, Neuchâtel 2002, Switzerland.

Dzwolak, Wojciech

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warsaw, 01-142, Poland.

Fecht, Hans-Jörg

Ulm University, Albert Einstein Allee 47, Ulm 89081 Germany Forschungszentrum Karlsruhe, Institute of Nanotechnology, POB 3640, Karlsruhe 76021, Germany.

Fichtner, Maximilian

Forschungszentrum Karlsruhe, Institute of Nanotechnology, POB 3640, Karlsruhe 76021, Germany.

Fidelus, Janusz D.

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warsaw, 01-142, Poland.

Fuchs, Harald

Physikalisches Institut, University of Munster, (WWU), Wilhelm Klemm Str. 10, Munster 48149, Germany.. Centre for Nanotechnology (CENTECH), Heisenbergstr. 11, Munster 48149, Germany.

Gburski, Zygmunt

University of Silesia, Institute of Physics, Uniwesytecka 4, Katowice 40-007, Poland.

Gleiter, Herbert D.

Forschungszentrum Karlsruhe, Institute of Nanotechnology, POB 3640, Karlsruhe 76021, Germany

Godlewsi, Marek

Polish Academy of Sciences, Institute of Physics, al. Lotników 32/46, 02-668 Warsaw, Poland. Cardinal Stefan Wyszynski University, College of Science, ul. Dewajtis 5, 01-815 Warsaw, Poland.

Goesele, Ulrich

Max Planck Institute of Microstructure Physics (MPH), Weinberg 2, Halle 06120, Germany.

Grigorjeva, Larisa

Institute of Solid State Physics, University of Latvia, 8 Kengarava, Riga 1063 Latvia.

Grzybowska-Świerkosz, Barbara A

Polish Academy of Sciences, Institute of Catalysis and Surface Chemistry, Niezapominajek 8, 30-239 Kraków, Poland.

Hahn, Horst W.

Darmstadt University of Technology, Institute of Materials Science Petersenstrasse 23, 64287 Darmstadt, Germany.

Jaworek, Anatol

Polish Academy of Sciences, Institute of Fluid Flow Machinery, Fiszera 14, 80-231 Gdańsk, Poland.

Jaworowicz, Jerzy

Laboratoire de Physique de Solides, UniversitéParis-Sud (CNRS UMR8502), Orsay 91405 France. University of Białystok, Institute of Experimental Physics, Lipowa 41, 15-424 Białystok, Poland.

Kapusta, Czesław

AGH University of Science and Technology, Nano-Materials Research Centre, Mickiewicza 30, 30-059 Kraków, Poland.

Kelly, Anthony

Cambridge University, Cambridge CB21EW, United Kingdom

Kimmel, Giora

Institute for Applied Research, Ben Gurion University of the Negev, Beer-Sheva, 84105 Israel.

Kisielewski, Jan

University of Białystok, Laboratory of Magnetism, Institute of Esperimental Physics, Lipowa 41, 15-424 Białystok, Poland.

Kosec, Marija

Jozef Stefan Institute, Electronic Ceramics Department, Jamova 39, 1000 Ljubljana, Slovenia.

Koutzarova, Tatyana I.

Institute of Electronics, Bulgarian Academy of Sciences (IE-BAS), 72, Tzarigradsko Chausee, Sofia 1784, Bulgaria.

Krawczyńska, Agnieszka T

Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Kurzydłowki, Krzysztof J.

Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Le, Minh Q.

Institute of Materials Science, 18 Hoang Quoc Viet, Hanoi 8404, Vietnam.

Lewandowska, Malgorzata

Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Lojkowski, Witold

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warszawa 01-142, Poland.

Manna, Indranil

Indian Institute of Technology, Kharagpur (IIT), Kharagpur, India.

Massalski, Tadeusz B.

Carnegie Mellon University, Pittsburg, PA 15213, United States.

Maziewski, Andrzej

University of Białystok, Laboratory of Magnetism, Institute of Esperimental Physics, Lipowa 41, 15-424 Białystok, Poland.

Millers, Donats

Institute of Solid State Physics, University 0f Latvia, 8 Kengarava, Riga 1063 Latvia..

Mohlala, Sarah M.

National Centre for Nano-Structured Materials, Council for Scientific and Industrial Research, Pretoria, 0001, Republic of South Africa.

Neumann, Peter

Max Planck Institut fur Eisenforschung, Department of Computational Materials Design, Max Planck Str. 1, Düsseldorf 40237, Germany.

Novikov, Vladimir Y.

MISA, Treptower Str. 74d, Hamburg 22147, Germany.

Opalińska, Agnieszka

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warszawa 01-142, Poland. Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Pan, Caofeng

Beijing National Centre for Electron Microscopy, Tsinghua University, Beijing 100084, China.

Pankratov, Vladimir

Institute of Solid State Physics, University 0f Latvia, 8 Kengarava, Riga 1063 Latvia

Paul, Heiko

Ulm University, Albert Einstein Allee 47, Ulm 89081 Germany

Philibert, Jean

UPS, Saint Germain en Laye 78100, France

Pietrzyk, Mieczyslaw A.

Polish Academy of Sciences, Institute of Physics, al. Lotników 32/46, 02-668 Warsaw, Poland.

Przewoźnik, Janusz J.

AGH University of Science and Technology, Faculty of Physics and Applied Computer Science, Mickiewicza 30, 30-059 Kraków, Poland.

Rabkin, Eugen

Technion - Israel Institute of Technology, Technion City, Haifa 32000, Israel.

Raczyński, Przemysław

University of Silesia, Institute of Physics, Uniwesytecka 4, Katowice 40-007, Poland.

Roy, Debdas

Indian Institute of Technology, Kharagpur (IIT), Kharagpur, India.

Salas-Adame, Blanca

Universidad Nacional Autónoma de Mexico, Instituto de Investigaciones en Materiales (UNAM), Circuito Exterior S/N Cd. Universitaria Coyoacan Mexico D.F., Mexico 70-360

Saxl, Ottilia

Institute of Nanotechnology, 6 The Alpha Centre, Stirling, FK9 4NF, United Kingdom.

Schimmel, Thomas

Forschungszentrum Karlsruhe, Institute of Nanotechnology, POB 3640, Karlsruhe 76021, Germany. Universität Karlsruhe (TH) Institut für Angewandte Physik, Wolfgang Gaede Str. 1, Karlsruhe 76131, Germany.

Schneider, Krystyna

AGH University of Science and Technology, Faculty of Physics and Applied Computer Science, Mickiewicza 30, 30-059 Kraków, Poland.

Semenova, Irina

Institute of Physics of Advanced Materials, Ufa State Aviation Technical University, 12 Karl Marx St., Ufa 450000, Russian Federation.

Shvindlerman, Lazar S.

Russian Academy of Sciences, Institute of Solid State Physics, Chernogolovka 142432, Russian Federation.

Siejka-Kulczyk, Joanna

Warsaw University of Technology, Faculty of Materials Science and Engineering , Wołoska141, Warszawa 02-507, Poland.

Sikhwivhilu, Lucky M.

National Centre for Nano-Structured Materials, CSIR, 1-Meiring Naude Road, Brummeria, P.O. Box 395, Pretoria, 0001, Republic of South Africa.

Skierbiszewski, Czeslaw

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warszawa 01-142, Poland.

Smith, George D.

University of Oxford, Department of Materials, Parks Road, Oxford, OX1 3PH, United Kingdom.

Smits, Krisjanis

Institute of Solid State Physics, University 0f Latvia, 8 Kengarava, Riga 1063 Latvia.

Sobczyk, Arkadiusz T

Polish Academy of Sciences, Institute of Fluid Flow Machinery, Fiszera 14, 80-231 Gdańsk, Poland.

Sobczyk, Joanna.

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warszawa 01-142, Poland.

Stobinski, Leszek

Polish Academy of Sciences, Institute of Physical Chemistry, Kasprzaka 44/52, 01-224 Warsaw, Poland.

Strachowski, Tomasz

Polish Academy of Sciences, Institute of High Pressure Physics (UNIPRESS), Sokołowska 29/37, Warszawa 01-142, Poland Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Thomas, Gareth

University California Berkeley and San Diego, 2415 Campus Dr., Irvine Cal., Berkeley, CA 92612, United States.

Tsakalakos, Thomas

Department of Materials Science and Engineering, Rutgers University, 607 Taylor Road, Piscataway, NJ 08854, U.S.A.

Tuliński, Maciej

Poznań Techncial University, pl. Marii Skłodowskiej-Curie 5, 60965 Poznań , Poland.

Van de Voorde, Marcel

Universtity of Technology Delft, Rotterdamseweg 137, Delft 2628 AL, Netherlands.

Vitek, Vaclav

University of Pennsylvania (PENN), 3231 Walnut Street, Philadelphia 19104, United States.

Wawer, Kinga

Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Webster, Thomas J.

Brown University, Division of Engineering, 184 Hope Street, Providence, RI 02917, United States.

Wei, Pai-Chun

Department of Materials Science and Engineering, National tsing Hua University, Hsinchu, Taiwan. Center for Condensed Matter Sciences, National Taiwan University, Taipei 10617, Taiwan.

Wejrzanowski, Tomasz

Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska 141, Warszawa 02-507, Poland.

Werner, Matthias R.

Nano and Micro Technology Consulitn (NMTC), Soorstr. 86, Berlin 14050, Germany.

Widlicki, Pawel

Warsaw University of Technology, Faculty of Materials Science and Engineering, Wołoska141, Warszawa 02-507, Poland

Wojnar, Ryszard

Polish Academy of Sciences, Inst. of Fundamental Technological Research, Świętokrzyska 21, 00-049 Warsaw, Poland.

Yavari, Alain R.

LTPCM-CNRS, Institut National Polytechnique de Grenoble, 1130 rue de la Piscine, BP 75, Grenoble 38402, France.

Zhou, Wuzong

University of St. Andrews, School of Chemistry, St. Andrews, Fife, KY16 9ST, United Kingdom.

Table of Contents Organisation Intro Participants

Nanostructures for Photonics How Can the Intra-Shell Emissions of Rare Earth and Transition Metal Ions in Thin Films and Nanoparticles Be Stimulated? M. Godlewski, S. Yatsunenko, A. Opalińska and W. Łojkowski Optical Properties of Nanocrystalline YAG:Ce H. Paul, D. Kessler and U. Herr Blue Laser Diodes by Low Temperature Plasma Assisted MBE C. Skierbiszewski

3 9 17

Oxide Nanostructures Stability, Instability, Metastability and Grain Size in Nanocrystalline Ceramic Oxide Systems G. Kimmel and J. Zabicky Mesoporous Crystals of Transition Metal Oxides W.Z. Zhou Structural, Magnetic and Electronic Properties of Surface Oxidised Fe Nanoparticles J. Przewoźnik, T. Tyliszczak, D. Rybicki, J. Żukrowski, W. Szczerba, M. Sikora, C. Kapusta, H. Stepankova, R.F. Pacheco, D. Serrate and M.R. Ibarra Nanosized Barium Hexaferrite Powders Obtained by a Single Microemulsion Technique T. Koutzarova, S. Kolev, K. Grigorov, C. Ghelev, I. Nedkov, M. Ausloos, R. Cloots, T. Mydlarz and A. Zaleski Synthesis of Titania Nanostructures and their Application as Catalyst Supports for Hydrogenation and Oxidation Reactions L.M. Sikhwivhilu, S.S. Ray and N.J. Coville Magnetic Anisotropy of Co Films Annealed by Laser Pulses J. Kisielewski, K. Postava, I. Sveklo, A. Nedzved, P. Trzciński, A. Maziewski, B. Szymański, M. Urbaniak and F. Stobiecki

29 37 47 55 61 69

Carbon Nanostructures First Principle Investigation of Structural Properties of Potassium Doped Fullerene Clusters – Kn(C60)2 M. Sokół and Z. Gburski MD Study of the Endohedral Potassium Ion Fullerene Cluster (K+@C60)7 A. Piątek, A. Dawid, K. Górny, R. Nowak and Z. Gburski Molecular Dynamics Simulation Study of the Liquid Crystal Phase in a Small Mesogene Cluster (5CB)22 W. Gwizdała, A. Dawid and Z. Gburski Preparation and Characterization of Polymer/Multi-Walled Carbon Nanotube Nanocomposites M.S. Mohlala and S.S. Ray Formation of Carbon Fibres in High-Voltage Low-Current Electrical Discharges A.T. Sobczyk, A. Jaworek, E. Rajch and M. Sozańska A Titanium-Decorated Fullerene Cluster – A Molecular Dynamics Simulation A. Piątek, R. Nowak and Z. Gburski

77 81 89 97 103 109

b

Perspectives of nanoscience and nanotechnology

Nanostructures for Medicine Nanotechnology for Treating Damaged Organs J. Lu and T.J. Webster Electrospray Nanocoating of Microfibres A. Jaworek, A. Krupa, A.T. Sobczyk, M. Lackowski, T. Czech, S. Ramakrishna, S. Sundarrajan and D. Pliszka Nanomaterials in Dental Applications M. Lewandowska, J. Siejka-Kulczyk, M. Andrzejczuk and K.J. Kurzydłowski The Influence of Graphene Sheet on the Dynamics of Cholesterol Molecules in the Lodgment Located near a Transmembrane Protein – MD Study P. Raczynski, A. Dawid and Z. Gburski Computer Simulation of the Dynamics of Homocysteine Molecules Surrounding a Carbon Nanotube P. Raczynski, A. Dawid, Z. Dendzik and Z. Gburski Dielectric Relaxation of a Cholesterol Domain Near a Graphite Wall - A Computer Simulation P. Raczynski and Z. Gburski

119 127 133 141 147 153

Bulk Metal Nanostructures The Mechanical Properties of Nano-TiO2 Dispersed Al65Cu20Ti15 Amorphous/Nanocrystalline Matrix Bulk Composite Prepared by Mechanical Alloying and High Pressure Sintering D. Roy, R. Mitra, T. Chudoba, Z. Witczak, W. Łojkowski, H.J. Fecht and I. Manna Enhanced Fatigue Properties of Ultrafine-Grained Titanium Rods Produced Using Severe Plastic Deformation I.P. Semenova, G.K. Salimgareeva, V.V. Latysh and R. Valiev Nanostructure Formation in Austenitic Stainless Steel A.T. Krawczynska, M. Lewandowska and K.J. Kurzydłowski Nanoscale Nickel-Free Austenitic Stainless Steel M. Tulinski, K. Jurczyk and M. Jurczyk The Effect of Grain Size Distribution on the Mechanical Properties of Nanometals T.B. Tengen, T. Wejrzanowski, R. Iwankiewicz and K.J. Kurzydłowski The Influence of the Initial State on Microstructure and Mechanical Properties of Hydrostatically Extruded Titanium K. Topolski, H. Garbacz, W. Pachla and K.J. Kurzydłowski Influence of Hydrostatic Extrusion Parameters on the Microstructure and Mechanical Properties of 6082 Aluminium Alloy P. Widlicki, P. Wiecinski, H. Garbacz and K.J. Kurzydłowski

Additional Oral Presentations

161 167 173 179 185 191 197

NANOSTRUCTURES FOR PHOTONICS

Solid State Phenomena Vol. 140 (2008) pp 3-8 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.3

How can the Intra-shell Emissions of Rare Earth and Transition Metal Ions in Thin Films and Nanoparticles be Stimulated? M. Godlewski 1,2, a, S. Yatsunenko 1, b, A. Opalińska 3, c and W. Łojkowski 3, d 1

Institute of Physics, Polish Academy of Sciences, Al. Lotników 32/46, 02-668 Warsaw, Poland 2 Deptartment of Mathematics & Natural Science, Cardinal Stefan Wyszyński University, 01185 Warsaw, Poland 3 Institute of High Pressure, Polish Academy of Sciences, Sokołowska 29/37, 01-142 Warsaw, Poland e-mail:

a d

[email protected]; b [email protected]; c [email protected]; [email protected]

Keywords: nanoparticles, luminophors, wide band gap oxides, rare earth ions, luminescence Abstract. Nanoparticles of the wide band gap oxides doped with rare earth (RE) ions are prospective materials for application in optoelectronics as phosphors in a new generation of light sources. In this paper the mechanisms of the excitation of efficient 4f-4f intra-shell transitions in RE doped nanoparticles are discussed. These mechanisms either enhance the rate of host to impurity energy transfer or stimulate the intra-shell transitions of RE ions. Introduction The widely used incandescent lamps, of which there are about 9 billion items in the world, are very inefficient as they have an efficiency of only about 3-4 percent, i.e., only 3-4% of the energy supplied is converted into visible light. For this reason they will soon be replaced by new generations of more efficient lamps. These will be compact fluorescent (CF) lamps and/or GaNbased white light emitting diodes, w-LEDs. The replacements will result in a huge financial saving, more than 80 billion USD per year, and reduced energy consumption, which will also result in reduced emission of CO2 gas to the atmosphere. These modern light sources both require luminophors to convert UV (in CF lamps) or violet/blue emission (in w-LEDs) to a visible emission. The currently available UV light conversion phosphors are optimized for down-conversion of the emission of mercury vapours, which emit in UV (55 % at 254 nm, 9 % at 185 nm) and only 5 % at the visible and near UV spectral regions. Unfortunately, light conversion efficiency is only very high if we count number of absorbed and emitted photons, i.e., it is 100 % if absorption of one UV photon results in emission of one visible photon. However, conversion efficiency is usually below 50 % if we calculate the photons energy. In the first generation of commercialized w-LEDs 400 nm blue/violet emission from InGaN quantum wells was mixed with a yellow emission of YAG:Ce phosphor to achieve an impression of a white light. This concept (hybrid LED) is still used in the more efficient recent w-LEDs. One can expect an improved efficiency if more efficient phosphors are developed. There are several alternative methods for the improvement of the currently used luminophors. In this paper two of them are briefly described. The first relates to a study of a small size of powder particles and the benefits of introducing such nanopowders will be explained. The second takes the advantage of the intra-shell emission of Praseodymium Pr3+ ions, for which one UV photon can

4

Perspectives of nanoscience and nanotechnology

excite two visible ones. In this case one can expect an improved efficiency, since one absorbed UV photon leads to an emission of two visible photons. This is why Pr doped phosphors are described as phosphors with a 200 % efficiency (1 UV photon induces 2 visible ones). Emission of RE doped luminophors Doping phosphors with RE ions helps to achieve light emission in a given region of the spectra region because of the attractive properties of RE ions. Their 4f-4f transitions result in sharp, atomic-like photoluminescence (PL) bands, which are temperature and host insensitive. It is for this reason that RE doped phosphors are widely used in optoelectronics. Unfortunately, the intra-shell transitions of RE ions are difficult to excite. This is due to the screening of electrons from the 4f shell by electrons in the external and filled 5s and 5p shells. In the consequence the 4f shell remains atomic-like and 4f-4f transitions are parity forbidden and usually insensitive to environment effects [1]. In consequence, host excitation is usually inefficient, which is why efficient 4f-4f transitions are, in most cases, achieved only on 4f-5d or charge transfer (CT) excitation (see [1] and references given there). Why nanoparticles? We found recently that the efficiency of Mn2+ intra-shell transitions in ZnS and CdS nanoparticles are enhanced compared to their efficiency in bulk samples, which we related to enhanced interactions between the excited states of Mn2+ ions and the photo-induced free carriers [2-5]. Such interactions are enhanced in nanoparticles, due to the quantum confinement imposed on the free carriers. A larger overlap of impurity and free carrier wave functions results in increased rates of host-to-impurity energy transfer as well as increased rates of intra-shell recombinations.

PLE Intensity (arb. units)

Regarding the increased rate of host-to-impurity energy transfer, we have demonstrated that the 4T1 to 6A1 PL of Mn2+ ions in ZnS and CdS nanoparticles can be efficiently pumped under band-to-band excitation followed by host-to-impurity energy transfer [5]. For the bulk ZnS sample the PL excitation (PLE) spectrum of this intra-shell emission consists of several peaks corresponding to transitions from 6A1 (Mn2+ ground state) to the excited 4G (4T1, 4T2, 4E and 4A1) and 2I (2T2) states (see Fig. 1 (left)), whereas the PLE spectrum of nanoparticles (Fig. 1 (right)) is dominated by the strong band-to-band excitation of ZnS host [5].

intra-shell

1.0

nano ZnMnS

0.8

0.6

0.4

bulk ZnMnS

0.2

0.0 3

4

5

3

4

5

Photon Energy (eV)

Figure. 1. Room temperature PLE spectra of the 4T1 to 6A1 intra-shell emission for bulk sample (left) and (right) for nanoparticles of ZnMnS with 1 % Mn.

Solid State Phenomena Vol. 140

5

Based on experimental results discussed elsewhere (see references [2-5]) we concluded that the enhanced host to Mn2+ energy pumping is related to confinement enhanced Mn-free carrier interactions. We also demonstrated that these interactions affect the rates of intra-shell transitions as long as the photo-excited free carriers coexist with the excited Mn2+ ions [2-4]. Even though such interactions are less important for RE ions, due to the screening effects referred to earlier, we can also expect that, in this case, impurity-free carrier interactions can be confinement dependent and can enhance the radiative recombination rates of the RE ions. The appearance of a fast PL decay components of 4f-4f intra-shell transitions is thus likely to be of the same origin as the one reported for Mn doped ZnS and CdS nanoparticles. RE stimulation by co-doping with transition metal ions We proposed recently that RE emission can be stimulated by co-doping nanoparticles with transition metal (TM) ions. We observed that, for ZnS nanoparticles co-doped with Mn and Tb ions, the relative intensity of Mn2+ and Tb3+ intra-shell emissions depends on excitation conditions [5]. A new broad PLE band was observed for Tb3+ ions in Mn doped nanoparticles. This band, which is due to the band-to-band host excitation, was not observed for ZnS powders doped only with Tb ions. The following energy transfer scenario based on experimental results presented in the reference [5] has been proposed. Energy transfer from the host-to-RE ions proceeds in nanoparticles co-doped with Mn ions via a Mn-to-RE transfer. Firstly, after band-to-band excitation, host-to-Mn2+ transfer occurs, which is then followed by an energy transfer from one of the excited Mn2+ multiplets to one of the Tb3+ excited states. The sensitizing of Tb3+ PL via Mn2+ co-doping is a very attractive property of ZnS nanoparticles. It enables the use of broad and intensive host PLE bands to excite sharp, atomic-like RE transitions. It is likely that similar situations should be found in other nanopowders co-doped with RE and TM ions. Photon cutting As already mentioned, the luminophors used in modern CF lamps efficiently convert the UV photons emitted by Hg vapours to visible photons with a ratio close to 1–to–1. This efficiency is however lower if we calculate the energy of absorbed and then emitted photons. For example, it is below 50 % for green phosphor, for which 254 nm photon emitted by Hg vapours results in 542 nm emission of the phosphor. It was proposed that the energy efficiency can be improved if a single UV photon emitted by Hg is able to excite two visible photons. A 200 % light conversion efficiency can be achieved in this way, i.e., 1 UV photon induces 2 visible photons with the sum of their energies close to the energy of the exciting UV photon. This process was called photon cutting or a photon cascade. It was proposed that such a situation is most likely for phosphors doped with Pr3+ ions. Pr3+ doped YF3 and LaF3 [6], SrAlF5:Pr3+ [7] showed the expected PLE spectra. In the photon cutting (photon cascade) process observed in reference [6] efficient 4f2 to 4f15d1 excitation is followed by a fast energy transfer to the highest lying 1S0 state of Pr3+ ion (see Fig. 2 showing energy level scheme for Pr3+ ion), and a radiative decay from this level in which two or more photons are emitted (1S0 – 3P1, 1I6; 3P0 – 3H4, 3H5, 3H6, 3F2, 3F3, 3F4).

6

Perspectives of nanoscience and nanotechnology

host 1 1 4f 5d or CT 6

1

S0

Energy (eV)

5

4

3

3

P2

3

1

P1, I6

3

P0 1

D2

2 1

1

G4

3

3

3

F3, F4

3

0

3

H5

H6, F2

3

H4

Figure. 2. Energy structure of Pr3+ ion in wide band gap materials with the indicated observed 4f-4f transitions.

Unfortunately the first transition is observed at too short wavelength to give an impression of white light emission after mixing with the following transitions. This makes it difficult to achieve optimal light emission conditions (two visible photons). One of the possible solutions is to push down the 1S0 level (by changing the RE ligand), or to move down the position of the 5d state hoping that 4f15d1 to 4f2 de-excitation will proceed via the 1I6; 3P0 excited states of Pr3+ resulting in the observation of a broad (due to 4f15d1 to 4f2 transition) and atomic sharp emissions. The latter possibility seems to be simpler. For example, for ZrO2 nanoparticles doped with Pr3+ ions the 4f2 to 4f15d1 excitation is shifted down in energy. In consequence the 1S0 state is resonant and 4f2 –PLE band overlaps with the charge transfer (CT) transition and band-to-band host excitation. Summary Summarizing, we have briefly reviewed the attractive properties of RE doped nanoparticles. The mechanisms of PL enhancement in RE doped phosphors have been elucidated. It has been shown that 4f-4f transitions can be stimulated by selecting the appropriate excitation conditions, by codoping with TM ions and the use of nanoparticles. Then the concept of photon cutting was introduced and explained based on the properties of Pr3+ ions in wide band gap materials. By photon cutting we mean the situation in which one exciting

Solid State Phenomena Vol. 140

7

photon results in a two (or more) photon cascade with the sum of the energy of the emitted photons close to the energy of the primary (exciting) photon. Acknowledgements This work was partly supported by Grant No. 1 P03B 090 30 awarded by Poland’s Ministry of Science and Higher Education for the years 2006-2008. The ZrO2 nanoparticles used in the investigated were prepared within the SWB Witnano network cooperation.

References [1] M. Godlewski and M. Leskelä: CRC Critical Reviews in Solid State and Materials Sciences Vol. 19 (1994), p. 199 [2] M. Godlewski, V.Yu. Ivanov, A. Khachapuridze, and S. Yatsunenko: Phys. Stat. Solidi (b) Vol. 229 (2002), p. 533 [3] M. Godlewski, S. Yatsunenko, A. Khachapuridze, V.Yu. Ivanov, Z. Gołacki, G. Karczewski, P.J. Bergman, P.J. Klar, W. Heimbrodt, and M.R. Phillips: J. Alloys Compd. Vol. 380 (2004), p. 45 [4] M. Godlewski, S. Yatsunenko, and V.Yu. Ivanov: Israeli J. Chem. Vol. 46 (2006), p. 413 [5] M. Godlewski, S. Yatsunenko, M. Zalewska, A. Kłonkowski, T. Strachowski, and W. Łojkowski: Solid State Phenomena Vol. 128 (2007), p. 123 [6] W. W. Piper, J. A. DeLuca, and F. S. Ham: J. Lumin. Vol. 8 (1973), p. 344 [7] A.P. Vink, P. Dorenbos, J.T.M. de Haas, H.Donker, P.A. Rodnyi, A.G. Avanesov, and C.W.E. Eijk: J. Phys.:Condens. Matter Vol. 14 (2002), p. 8889

Solid State Phenomena Vol. 140 (2008) pp 9-16 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.9

Optical Properties of Nanocrystalline YAG:Ce Heiko Paul a, Dominik Kessler b and Ulrich Herr c University of Ulm, Institute of Micro and Nanomaterials, Albert-Einstein-Allee 47, 89081 Ulm, Germany e-mail:

a

[email protected];

b

[email protected];

c

[email protected]

Keywords: luminescence, quenching, nanocrystalline, YAG:Ce, lifetime Abstract. Due to its high quantum efficiency (QE) for luminescence, conventional coarse-grained YAG:Ce (Y3Al5O12:Ce) finds widespread use in light conversion and scintillator applications. Nanocrystalline YAG:Ce may possess modified optical properties which are advantageous for technological applications, but this will depend on highly efficient energy conversion. In this work, the effect of the particle size and Ce concentration on the quantum efficiency and the optical lifetime of the YAG:Ce emission will be characterized and discussed. Nanocrystalline YAG:Ce with an average particle size of 20 to 50 nm was synthesized by the chemical vapour reaction (CVR) method and subsequently analyzed using various techniques. When comparing the nanocrystalline samples to a coarse-grained reference sample, the particle size and doping concentration was found to have a significant influence on quantum efficiency. It was established that the nanocrystalline samples investigated exhibit a lower QE at ambient temperature than the coarse-grained reference. The results of the optical lifetime measurements are discussed in relation to this reduction in QE. Introduction Modern white light-emitting diodes (LEDs) with their high power efficiency, superior brightness levels and long-term stability are about to occupy a dominant position in the lighting applications market. These white semiconductor-based light sources rely on InAlGaN LEDs with a typical electroluminescence between 400 and 480 nm. In order to achieve the desired impression of white colour for the human eye, an additional YAG:Ce conversion layer is placed on top of the LED. The YAG:Ce layer absorbs a certain quantity of the blue photons and re-emits them at a wavelength around 550 nm, a down conversion, which is the complementary colour of the blue LED emission. The YAG:Ce used as the conversion material, in its coarse-grained form, is a well known, intensely studied phosphor. However, since the colour impression of the white LED depends on the amount of transferred photons, (i.e. the ratio of the blue and yellow intensity, light scattering in the conversion layer is one of the major topics for an impression of a homogenous white colour. In the size regime of nano particles (particle sizes D < 100 nm), Rayleigh scattering is the dominating effect for a typical 450 nm emission of a blue LED, and the scattered intensity is proportional to D6. Therefore, nanocrystalline YAG:Ce with particles of size below 50 nm should provide significantly improved light scattering behaviour by the conversion layer. On the other hand, quenching of the luminescence in nano particles may be more prominent than in the conventional fluorescent materials. In this paper we report the effect of particle size of nanoYAG:Ce powder on the optical properties with reference to their application as a conversion material. The effect of the Ce concentration xCe on the QE and the optical lifetime τ of the 4f–5d transition in nano-YAG:Ce is reported and the results compared to those obtained from a coarse grained reference sample.

10

Perspectives of nanoscience and nanotechnology

Sample Preparation and Experimental Procedure The samples were prepared by the chemical vapor reaction technique (CVR) in a hot-wall reactor at temperatures between 1500°C and 1650°C. The precursors used were commercially available Aluminium-sec-butoxide, Yttrium-[TMOD] and Cerium-[TMHD]. Helium was used as the carrier gas for the evaporated precursor materials. The composition of the samples was controlled by the temperatures of the precursors and the Helium flow rates. In addition oxygen was supplied to the precursors to ensure the complete conversion of the precursor materials to the oxide nano particles. The total gas flow rate in the reactor typically varied between 200 and 250 sccm, while the working pressure was adjusted to values between 50 and 100 mbar. The as-prepared powder samples were of perfectly spherical shape with diameters ranging from 20 to 50 nm. To improve the phase formation and the optical properties, the samples were exposed to an ex-situ annealing stage for 15 min at 1100°C. Photoluminescence (PL) was used to investigate the relative quantum efficiency, emission properties and excitation behaviour, while lifetime measurements of the luminescence decay revealed information about the relaxation mechanisms of the excited Ce states. X-ray diffraction (XRD) was used to examine the structure and grain size of the material. Furthermore, analysis of the chemical composition and doping concentration was undertaken by energy dispersive x-ray analysis (EDX) in a scanning electron microscope (SEM). Morphology, agglomeration and sintering of the powder samples were determined using SEM images. The effect of grain size on the Ce3+ luminescence

normalized excitation spectrum

Fig. 1 shows the normalized excitation curves at room temperature for a nanocrystalline YAG:Ce sample (D ≤ 50 nm) and the coarse-grained reference (several microns in diameter), recorded for emission at 530 nm. 1.0

nano-YAG:Ce coarse-grained reference

0.8 0.6 0.4 0.2 0.0 200

250

300

350

400

450

500

550

excitation wavelength [nm]

Fig. 1. Excitation spectra for emission at 530 nm of nano-YAG:Ce with xCe = 0.25 at.% and the coarse-grained reference. It can be clearly seen that the typical Ce3+ related absorption peaks around 450 nm, 340 nm and 240 nm appear in both cases. However, compared to the reference sample, which exhibits nearly the same excitation at 340 nm as for 450 nm, the nanocrystalline sample shows a drastically reduced excitation at 340 nm. This reduction of the emission for excitation into higher electronic states of the crystal-field split 5d orbital might be related to the relative energy gap between the Ce3+ 5d states and the conduction band (CB) of the YAG host lattice [1]. Only the lowest 5d level has a sufficient gap to the conduction-band edge and therefore the highest conversion yield [2]. Even the

Solid State Phenomena Vol. 140

11

second lowest level (fed by the 340 nm absorption) is already located very close to, or even inside, the CB. This means that electrons promoted from a 4f level into this excited state have the possibility i) to leave the Ce ion by migration, or ii) to form a bound exciton with the Ce ion, or iii) to form a free exciton with a hole in the valence band of the host lattice, before they can relax to the emitting 5d level [3]. In some of these cases the excited electron (or exciton) will relax in a nonradiative way, and thereby decrease the luminescence yield. A schematic drawing of the energy levels is given in Fig. 2 [1,4]. Conduction Band

450 nm

178 nm (Eg~ 7eV)

340 nm

230 nm

5d

4f Valence Band

Fig. 2. Energy diagram of the electronic states in the Ce ion (4f: ground states, split by spin-orbit coupling; 5d: excited states, split by crystal-field interaction) and the energy bands of the host lattice [1,4]. Another interesting detail in the excitation spectrum of the nanocrystalline sample is the obviously narrower absorption peak at 450 nm (see Fig. 1). This can be caused by a reduced number of possible transitions between different vibrational wave functions of the ground state and the excited state of the Ce3+ ion (configuration coordinate model [5]).

normalized emission spectrum

Fig. 3 shows the room temperature emission spectra of the two samples. 1,0

nano-YAG:Ce coarse-grained reference

0,8 0,6 0,4 0,2 0,0 500

550

600

650

700

750

emission wavelength [nm]

Fig. 3. Emission spectra for excitation at 450 nm of nano-YAG:Ce with xCe = 0.25 at.% and the coarse-grained reference While the coarse-grained reference has its maximum at a wavelength of 545 nm, the nanocrystalline sample peaks at 530 nm, corresponding to a difference of 64 meV. There are several possible reasons for a shift in the emission. Firstly, the crystal field of the Oxygen anions O2- surrounding the Ce3+ can differ in the case of nanocrystalline samples from those in bulk material and hence, provide a different splitting of the 5d levels. This would lead to a modified energy gap between the 4f and 5d states, and therefore to a shift of the emission [6]. Another reason for an emission shift

12

Perspectives of nanoscience and nanotechnology

could be a different Ce3+ density in the YAG host lattice. It is known that with increasing doping concentration the luminescence experiences a red shift [2]. However, a strong dependence of the emission peak on the Ce concentration for the nano-YAG:Ce samples examined in this study was not observed. Effect of the Cerium concentration on the optical properties To investigate the consequences of the Ce doping level xCe on the optical properties of nanocrystalline YAG, a series of samples with 0.01 at.% ≤ xCe ≤ 0.65 at.% was prepared and analyzed by photoluminescence. Fig. 4 shows the peak intensities (λem = 530 nm) of the emission of the different nano-YAG:Ce samples normalized to the luminescence of the coarse-grained reference. As expected, the intensity starts to increase when the Ce concentration is increased. The number of active luminescence centres in the host lattice increases, thus, more photons of the 450 nm exciting radiation can be absorbed and converted. However, for large Ce concentration, (see the data point at 0.65 at.% Ce in Fig. 4), the emission intensity is decreasing again, indicating nonradiative quenching.

normalized emission intensity

1,0

coarse grained reference

0,9 0,8 0,7 0,6 0,5 0,4 0,3 0,2 0,1 0,0 0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

Ce concentration [at.%]

Fig. 4. Normalized emission intensity (λem = 530 nm, λex = 450 nm) as a function of Ce concentration in various nano-YAG:Ce samples. Additionally the luminescence of the coarse reference is shown (+). (T = 295 K) For a similar Ce concentration, the coarse-grained reference sample shows strong luminescence with a QE (as independently determined) close to unity. The effect of concentration quenching in nano-YAG:Ce may arise from the resonance energy transfer (dipole-dipole interaction) between Ce atoms, which leads to the migration of excited states in a random-walk manner through the lattice, until the excited state returns either at a Ce3+ ion to the ground state (under the emission of a photon), or more likely, by non-radiative relaxation at a quenching site [7-11]. In nano-sized materials, where the average distance from any lattice site to the particle’s surface is only on the order of a few nanometers, the density of quenching sites per unit volume may be considerably higher than in coarse-grained or bulk material. The discontinuity in lattice periodicity and broken chemical bonds at the surface generate quenching centers in terms of additional electronic states located in the band gap of the YAG (see Fig. 2), which may drain the excited electrons from the luminescence centers [3].

Solid State Phenomena Vol. 140

13

0,28 0,26

relative absorption

0,24 0,22

remission intensity [a.u.]

To further investigate the dissipation of excitation energy at quenching sites by non-radiative decay, remission measurements were performed. Fig. 5 illustrates the dependence of the relative absorption (calculated from the remission spectra) on the doping concentration.

I0

I' 350

400

450

500

550

λ [nm]

0,20 0,18 0,16 0,14 0,12 0,10 0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Ce concentration [at.%]

Fig. 5. Relative absorption at 450 nm of the nano-YAG: Ce samples as a function of Ce concentration (T = 295 K). As shown in the inset, the relative absorption was calculated from the remission intensities as I0 − I' . The coarse grained reference I0

sample showed a relative absorption of 0.79 From the results in Fig. 5, the quenching of the luminescence for higher values of xCe becomes evident. While the relative absorption exhibits a behavior similar to the luminescence at low Ce concentrations in Fig. 4, a completely different behavior is observed for large concentrations of Ce. Dexter and Schulman calculated the transition probability wtr for a dipole-dipole interaction between ions to be proportional to R-6, where R is the distance between the two interacting ions [7]. The rapid decrease of wtr with increasing distance explains the strong concentration quenching when the doping level reaches a critical limit, while the efficiency of energy transfer between neighboring Ce ions vanishes for low Ce concentration, because of the large average inter-ion spacing. The quenching by dipole-dipole transfer is presumably the dominating mechanism for an excitation wavelength of 450 nm, whereas for excitation with 340 nm the previously mentioned electron transfer through the host lattice can also occur. Low temperature photoluminescence measurements at 77K were made to investigate the quenching process active in the nano-YAG:Ce samples. Fig. 6 shows the ratio of the luminescence intensities measured at 77K and room temperature (295K): I77K/I295K, as a function of the Ce concentration for different nano-YAG samples. As the large values of I77K/I295K in Fig. 6 indicate, the quenching process is largely suppressed at 77K. In fact, the Ce3+ emission increases by more than a factor of 60 for the sample with the highest Ce concentration.

14

Perspectives of nanoscience and nanotechnology

intensity ratio I77K / I295K

70 60 50 40 30 20 10

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Ce concentration [at.%]

Fig. 6. Ratio of the luminescence intensities at 77K and 295K vs. the Ce concentration. This increase may, in part, be attributed to the reduced transfer probability wtr as a result of the temperature dependence of the spectral overlap of the absorption and emission curves of the Ce3+ [7]. A comparison of the absorption and emission peaks revealed, that the spectral overlap at 77K is reduced to approximately 1/3 compared to the one at 295K. On the other hand, samples with low Ce concentration exhibit an increase of the emission at 77K by a factor of 10 to 20. For these samples, the energy transfer between Ce ions is less important and the observed increase reflects mainly the reduction of the thermal quenching processes. The larger increase by a factor of approximately 60, observed for the largest Ce concentration in Fig. 6, indicates a reduction of both, thermal and concentration quenching. Furthermore, at a temperature of 77K the nano-YAG:Ce sample with 0.65 at.% Ce reaches a conversion efficiency almost equivalent to that of the coarse reference sample. Effect of the Cerium concentration on the optical lifetime Fig. 7 depicts the dependence of the optical lifetime τ on the Ce concentration. It is apparent, that with increasing values of xCe the measured lifetime drops drastically. This behavior supports the assumption of a strong concentration related quenching process involving a non-radiative relaxation channel. Dexter and Schulman derived the following relationship between the luminescence yield η and doping concentration x for the effect of concentration quenching by resonance energy transfer:

η∝

1 , 1 + A⋅ x 3

(1)

where A is a general factor [7]. Since

τ = τ rad ⋅η ,

(2)

where τrad is the intrinsic radiative lifetime, which is assumed to be independent of the doping concentration, the model in Eq. 1 can directly be applied to the data in Fig. 7.

Solid State Phenomena Vol. 140

15

100

optical lifetime [ns]

90 80 70

data fitted to Eq. 1

60 50 40 30 0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Ce concentration [at.%]

Fig. 7. Optical lifetime measurements for different Ce concentrations. The samples were excited with a pulsed laser at 338 nm.

The dashed line in Fig. 7 represents the fit of Eq. 1 to the measured data. The excellent agreement of the model and the measured data proves that concentration quenching is an important effect for the optical lifetime (since thermal quenching should be almost independent of xCe). Nevertheless, at lower values of xCe (where τ is expected to be the intrinsic lifetime of the Ce3+ transition), the measured lifetimes are significantly larger (≈ 95 ns) than typical values found in the literature of approximately 70 ns. This deviation probably originates from the fact that the excitation for the optical lifetime measurements was carried out at 338 nm, which excites a level close, or even inside, the YAG conduction band (see above). A similar effect was also observed by Zych et al. for high energy excitation [1]. Under high energy excitation, electrons can be transferred to the conduction band (CB) (Ce3+ → Ce4++e-), while electrons from the valence band (VB) may be promoted to the Ce ion (Ce4++e-→ Ce3++hole) and a free electron/hole pair is generated (Born-Haber cycle) [9]. Alternatively, the free electron in the CB can be bound to the Ce4+ ion. If these two possible configurations decay (according to their specific lifetime) and energy is transferred back to the Ce3+ ion, radiative relaxation with an increased optical lifetime τ can occur [1]. Conclusions The photoluminescence and lifetime results of nanocrystalline YAG:Ce samples at room temperature and at 77K show, that besides a strong thermal quenching of the luminescence in nanocrystalline phosphors, an additional decay of the emission intensity and lifetime occurs for high doping concentrations. It was possible to attribute the concentration quenching to the effect of resonance energy transfer between different luminescence centres. Furthermore the obviously increased lifetime at lower doping levels, by the interaction between photo-induced free electrons and holes and the 5d level of the Ce3+ ion, are explained.

16

Perspectives of nanoscience and nanotechnology

Acknowledgement This work was supported by the German Federal Ministry of Education and Research (BMBF). References [1] E. Zych, J. Phys.: Condens. Matter 12, 1947 (2000) [2] D. Haranath, Appl. Phys. Lett. 89, 173118 (2006) [3] E. Zych, J. Luminescence 75, 193 (1997) [4] D.S. Hamilton, Phys. Rev. B 39, 8807 (1989) [5] G. Blasse and B.C. Grabmaier, Luminescent Materials, Springer-Verlag (1994) [6] D.J. Robbins, J. Electrochem. Soc. 126, 1550 (1979) [7] D.L. Dexter, J. Chem. Phys. 22, 1063 (1954) [8] M.J. Weber, Phys. Rev. B 4, 2932 (1971) [9] M. Raukas, Appl. Phys. Lett. 69, 3300 (1996) [10] D.L. Dexter, J. Chem. Phys. 21, 836 (1953) [11] T. Förster, Annalen der Physik 6, 55 (1948)

Solid State Phenomena Vol. 140 (2008) pp 17-26 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.17

Blue Laser Diodes by Low Temperature Plasma Assisted MBE Czeslaw Skierbiszewski Institute of High Pressure Physics, Polish Academy of Sciences, ul Sokolowska 29/37, 01-142 Warsaw, Poland e-mail: [email protected] Keywords: laser diodes, molecular beam epitaxy

Abstract. Recent progress in the growth of nitride based semiconductor structures made by plasma assisted MBE (PAMBE) is reported. The technology is ammonia free and the nitrogen for growth is activated by an RF plasma source from nitrogen molecules. A new approach for the growth of nitrides by PAMBE at temperature range 500 - 600°C is described. The key for this technique is to use a thin, dynamically stable metal (In or Ga) layer on the (0001) GaN surface, which enables a high quality 2D step-flow growth mode to be achieved at temperatures much lower than those determined by thermodynamic considerations. A new perspective for PAMBE in optoelectronics has been opened recently by a demonstration of continuous wave operation of InGaN blue–violet laser diodes. These laser diodes were fabricated on bulk GaN substrates with a low threading dislocation density. Introduction The potential associated with nitride-based Light Emitting Diodes (LEDs) and Laser Diodes (LDs) for solid-state lighting has generated a continuing interest in group III-N materials and their alloys. Until very recently, the major achievements and developments in the field of InGaN laser diodes (LDs) have been made by the Metal-Organic Vapour-Phase Epitaxy (MOVPE) technique [1]. The key problem facing molecular-beam epitaxy (MBE) of gallium nitride is that GaN begings to decompose at about 800ºC in vacuum. To achieve temperatures required for high quality growth, ~1100ºC (which is about half of the GaN melting point temperature), requires matching overpressures of active nitrogen species in the range of 0.1 – 1 bar to stabilize the GaN surface [2]. It is clear that for MBE, which relies on a negligible interaction of the atomic beams, such conditions for growth at high temperatures are very difficult to achieve. However, recent progress in understanding the growth mechanisms of nitrides by plasma-assisted MBE (PAMBE) has led to the ability to obtain atomically flat GaN surfaces at temperatures below 800ºC, where the GaN surface is stable under vacuum conditions [3,4]. In PAMBE the active nitrogen is generated from N2 molecules by an RF plasma cell. It was proved experimentally that it is possible to achieve a step-flow growth mode at temperatures below 800ºC, when a thin dynamic Ga (or In) layer is formed on the surface, i. e. for metal rich conditions [3,4]. This finding was confirmed by first principle ab-initio calculations, in which a substantial reduction of the energy barrier for N adatom diffusion on GaN surface was predicted [5]. It was important for fundamental research, as well as for industrial applications, to prove that low temperature PAMBE can produce high quality structures and subsequently to demonstrate nitride-based continuous wave (cw) LDs [6]. This paper reports the progress made by PAMBE for optoelectronic devices which shows the potential of the technology. Experimental The growth of all nitride structures presented in this paper was performed in a customized VG90 Oxford MBE reactor equipped with a Veeco RF plasma source (operating at 240 W for 0.8 sccm N2 flow). The pressure during growth was 1.5⋅10-5 Torr. The substrates used were either high pressure grown bulk GaN [7] or GaN/Al2O3 templates made by MOCVD. The epi-ready bulk substrates were prepared in a three-step process of mechanical polishing, dry etching, and deposition of a 2

18

Perspectives of nanoscience and nanotechnology

µm GaN:Si buffer layer in the MOVPE reactor. The back surfaces of the substrates were coated with a 0.7 µm molybdenum layer to improve the thermal coupling for radiative heating. Special holders capable of accommodating small, irregularly shaped substrates and designed to minimize edge effects ensured high temperature uniformity across the entire substrate area. The typical size of GaN bulk high pressure substrates was 4 mm x 5 mm, while MOCVD GaN/Al2O3 templates were 10mm x 10 mm. All layers discussed in this paper were grown on the (0001) Ga-polarity surface. Growth of III-N by PAMBE During growth, atoms which reach the growth surface can be bound on the terraces (atom C on Fig. 1), atomic ledges (atom B) or atomic kinks (atom A).

B

A

C

2D step-flow Figure 1. The two-dimensional step flow epitaxy model. The binding energies for the atoms on the terrace (ETERRACE), on the ledge (ELEDGE) and on the kink position (EKINK) increases as the number of dangling bonds of atoms C, B and A (ETERRACE < ELEDGE < EKINK) decreases. Therefore atom A has the highest binding energy among atoms A, B and C. The two-dimensional (2D) step flow growth mode takes place when all atoms are finally bound on the atomic kinks (like atom A), while those which did not reach the atomic kinks desorb from the growth surface (Fig. 1). The atoms at the kink positions have a half-filled number of dangling bonds while atoms inside the crystal have no dangling bonds. Therefore the binding energy of the atom at the atomic kink is equal to half the binding energy of an atom inside the crystal. From that point of view, the growth temperature TG for 2D step flow growth is equal to half the melting point TM of the grown material [2]. As an example, for GaAs and Si, TM is 1511 K and 1685 K respectively, while for GaN it is in the range between 2540 K (experimentally determined for 6 GPa [8]) and 2800 K (theoretically calculated [9]). It is demonstrated that the highest quality GaAs layers are achieved at growth temperatures in excess of about 520oC (570oC for Si) [2]. These temperatures are equal to 0.5⋅TM. This rule explains also why for GaN growth, the temperature of 1050oC - 1100oC is used in MOVPE. However, for such high temperature, a high surface concentration of active nitrogen is needed to prevent GaN decomposition. Thus a very high flow of ammonia, with an overpressure in the range of 0.1 - 1 bar, must be used. Such overpressures are not compatible with the MBE technology which relies on high vacuum conditions for the delivery of atoms from the effusion cells to the growing layer. In fact, for typical MBE growth conditions GaN starts to decompose rapidly at temperatures in excess of 800°C, restricting the epitaxy to temperatures much below the optimum point. Indeed, due to the arguments given above and the low diffusivity of N adatoms, the early efforts to grow GaN in MBE at temperatures below 800oC, using group V-rich conditions typical for III-V epitaxy, gave unsatisfactory results. These early difficulties, seemingly well grounded in

Solid State Phenomena Vol. 140

19

the simple thermodynamics of the processes involved, led many to believe that the only path for successful growth of nitrides in MBE reactors is to push the growth conditions as close as possible to those present in MOVPE reactors. The breakthrough in the study of growth kinetics in PAMBE came with the finding, that in Ga rich conditions, it is possible to grow relatively smooth layers at low growth temperatures. However, such growth conditions were prone to the formation of Ga droplets on the GaN surface and high quality material was mainly formed in the regions between the droplets. Further study of the Ga auto-surfactant effect led to the conclusion that this problem could be avoided, provided the Ga to N flux ratio was maintained in a very narrow range of values: just below the formation of the droplets, but high enough to ensure the formation of a metallic Ga bilayer on the Ga polarity surface [3,4]. Under such conditions, in many respects resembling liquid phase epitaxy (LPE), impinging N atoms easily penetrate this metallic layer and a new diffusion channel for N atoms is formed just below Ga (or In) surface cover (see Figure 2).

(a) N

GaN T=710oC

(b) Ga bi-layer

N

GaN

Figure 2. The N adatom kinetics on GaN (Ga polarity) surface at 710°C. (a) Without the Ga layer, the high barrier for lateral N diffusion leads to a 3D growth mode (b) With bi-layer Ga coverage – the effective lateral diffusion channel below Ga is opened for the step - flow 2D growth mode It was noticed almost 40 years ago, that for given temperature, the presence of metal layer (Ga or In) on a GaN surface accelerates etching of GaN in comparison to clean surface [10]. The LPE-like conditions reduce the formation energies of the kink atoms, thus reducing the optimal growth temperature. On the other hand, ab-inito calculations revealed that metallic coverage of the GaN (0001) surface also increases dramatically the lateral mobility of N atoms by reducing the activation energy for the process from 1.3 eV to about 0.3 - 0.5 eV. This enhanced surface mobility of nitrogen adatoms, coupled with the reduction of the kink formation energy, promotes 2D growth nucleation and step-flow growth at far below the temperatures needed for “classical” group V-rich conditions. Figure 3 gives insight how surface morphology depends on the choice of Ga and N fluxes for a constant growth temperature. For the nitrogen rich conditions (when the Ga flux ,ΦGa, is smaller than the N flux, ΦN), a 3D growth mode is observed leading to the GaN morphology shown in Fig. 3a (picture from Scanning Electron Microscopy – SEM). For ΦGa higher than ΦN three regions can be distinguished. In the first one a mixed 2D and 3D growth is observed (Fig. 3b), indicating that the bilayer of Ga is probably formed only locally. Fig. 3c shows the surface morphology for a 2D growth mode achieved when the thin (bilayer) coating of Ga is formed over the entire substrate. For

20

Perspectives of nanoscience and nanotechnology

higher Ga fluxes, a large number of Ga droplets were observed (see Fig. 3d). An interesting feature is visible in Fig. 3c where spiral atomic steps are forming well defined hillocks on the surface.

(a)

(b)

(c)

(d)

Figure 3. SEM images of GaN grown on GaN/sapphire templates by PAMBE at a temperature of 710°C and constant nitrogen flux with different Ga fluxes. (a) Ga/N=1, 3D growth mode. SEM magnification x5 0,000 (b) Ga/N =1.8, 3D and 2D mixed growth mode. SEM magnification x 100,000 (c) Ga/N=2.4, 2D step-flow growth mode. SEM magnification x 100,000 (d) Ga/N=2.8, 2D growth mode and Ga droplets on the surface. SEM magnification x 500 This surface was also examined by Atomic Force Microscopy (AFM) – see Fig. 4a, where the spirals are more clearly observed.

Solid State Phenomena Vol. 140

21

Height of hillocks, H (nm)

8

(a)

GaN on GaN/Al2O3 InGaN on GaN/Al2O3

6

4

(b)

2

H

QWs

0 580

600

620

640

660

680

700

720

740

o

Growth temperature ( C)

Figure 4. (a) Atomic Force Microscope images of the InGaN layer on GaN/sapphire template; (b)height of the hillocks as a function of growth temperature in PAMBE Such morphological features are centered at the intersection of a threading dislocation (TD) with the crystal surface, and are formed in all crystal growth processes proceeding through 2D nucleation. For PAMBE, in the group III metal rich regime the spiral growth is always present on high threading dislocation density (TDDs) substrates [11]. It is important to realize that such spiral growth leads to the formation of hillocks whose height increases monotonically with increasing tightness of the spiral. From our experience, PAMBE growth on high dislocation density GaN/sapphire substrates results in almost flat surfaces (hillocks height below 2 nm) at growth temperatures higher than 710oC. However, as the growth temperature decreases, the height of hillocks hincreases strongly and at temperatures around 600oC (where InGaN layers are typically grown) can reach heights of 6 - 7 nm. In contrast the growth on dislocation-free bulk GaN substrates does not suffer from such morphological instability In Figure 5 we present a comparison of the surface morphology of InGaN layers grown by PAMBE on GaN bulk crystals grown by the High Nitrogen Pressure Solution Growth technique (TDDs < 100cm-2) and GaN/sapphire templates.

(a)

3 µm

3 µm

Figure 5. (a) Atomic Force Microscope images of In0.02Ga0.98N layers on a bulk crystal, (b) on a GaN/sapphire template

(b)

22

Perspectives of nanoscience and nanotechnology

Nearly straight atomic steps were present on InGaN layers grown on bulk material, while a dislocation mediated step-flow growth mode was observed on the GaN/sapphire substrate. As a consequence of the non-uniform dislocation distribution in the substrate, the hillock diameters and sidewall inclinations vary substantially across the surface. Recent cathodoluminescence studies of PAMBE-grown InGaN quantum wells (QWs) [12] showed that surface faceting may have a profound influence on In incorporation, QW width and its lateral distribution. Thus the presence of threading dislocations may degrade the optical properties of PAMBE-grown InGaN QWs - not only through the very localized non-radiative recombination centers in the immediate vicinity of the dislocations - but also through much more extended fluctuations in the QW composition and thickness. The incorporation of In in InGaN layers strongly depends on the growth temperature, with higher compositions requiring lower growth temperatures. This is illustrated in Fig. 6, where the In composition is plotted as function of the substrate temperature for a set of InGaN layers on GaN/sapphire substrate. 32

In content (%)

28 24 20

0.3µm InxGa1-xN MBE GaN MOVPE

16 Al2O3

12 500

520

540

560

580

600

o

Growth Temperature ( C)

Figure 6. Indium content of the InGaN layers as a function of the growth temperature in PAMBE. The amount of In was inferred from measurements of the energy gap, Eg, of InxGa1-xN layers (Eg(x) after Wu et. al. [13]) and confirmed by XRD space map experiments where the degree of InGaN relaxation was taken into account. As explained earlier, a lower growth temperature on high TDDs substrates results in increased hillock heights and subsequent degradation of the Quantum Wells (QWs). Although the above observations do not signal any particular advantage of the bulk substrates over GaN/sapphire MOVPE templates for achieving high In incorporation, one would expect the InGaN QWs grown on bulk substrates to be superior to those grown on templates for any chosen growth conditions. Furthermore, when moving towards optoelectronic devices working in the green spectral region (see photoluminescence (PL) spectra of InGaN MQWs grown on low TDDs bulk substrates in Fig. 7a), where considerably more In has to be incorporated, growth on bulk substrates maybe the only option. Indeed, the local disorder on the surface introduced by the presence of dislocations and hillocks acts as a seed for “catastrophic” degradation of epitaxial layers grown in the “difficult” regimes, that is the growth of layers which exceed the critical thickness. Thus, low dislocation substrates should enable the growth of metastable layers, allowing the system to “supercool” considerably before generating various phase transitions (e.g., inversion domains, InN precipitates, Mg precipitates, dislocation clusters, etc.). Importantly, such layers can be extremely robust once

Solid State Phenomena Vol. 140

23

grown due to the very high melting temperatures for GaN and related compounds. Figure 7b presents the temperature dependence of the PL data on low TDDs bulk GaN substrates. InxGa1-xN/In0.02Ga0.98N MQW (10x) x=0.04

1.0 0.8

1.0

x=0.24 x=0.27

x=0.12

PL Intensity (a.u.)

PL Intensity (a.u.)

1.2

T=300K

0.6 0.4 0.2

λ=420 nm

0.8

IQE=50%

0.6 0.4

λ=500nm

IQE=12%

0.2 IQE=Intensity(300K)/Intensity(4K)

0.0

0.0 350

400

450

500

550

600

0

50

100

150

200

250

300

Temperature (K)

λ (nm)

Figure 7. (a) Photoluminescence of the MQWs as a function of the In content. (b) Temperature dependence of the photoluminescence for blue (420 nm) and green (500 nm) MQWs. It was found that the PL intensity ratio taken at room temperature and at 4 K is equal to 0.5 for 420 nm (blue-violet) and 0.12 for 500 nm (green), respectively. This translates to an internal quantum efficiency (IQE) of 50% (for blue-violet) and 12% (for green) of PAMBE-grown QWs (assuming that at low temperatures the nonradiative recombination channels are suppressed). Laser Diodes Fabricated by PAMBE Laser diode structures were deposited on (0001) Ga- polarity conductive, low dislocation density high pressure grown GaN bulk substrates. The 40 nm GaN:Si buffer layer and 450 nm Al0.08Ga0.92N:Si cladding were grown under Ga- rich conditions at 720oC. The bottom waveguide, multiquantum wells (MQWs), electron blocking layer, top waveguide, top cladding, and contact layer were grown under In- rich conditions at 600oC. The active region consisted of five 3-nm In0.1Ga0.9N wells with 7-nm In0.02Ga0.98N barriers (see Fig. 8).

60

PAMBE LD

7

In0.1GaN:Mg 3 nm In0.02GaN:Mg 14 nm

6

In0.02GaN:Mg 2.5 nm/ In0.02Al0.18GaN:Mg 2.5 nm

50

5

40

4

30

3

o

Tc=20 C

20 10 0

0.0

60x SL

Voltage (V)

Optical Power (mW)

70

In0.02GaN:Mg 70 nm EBL In0.02Al0.18GaN:Mg 14 nm In0.02GaN 1nm MQWs In0.02GaN:Si 7 nm In0.1GaN 3 nm 5x (or 3x)

In0.02GaN 100 nm

2

(a) 0.1

0.2

0.3

0.4

0.5

0.6

Current (A)

0.7

0.8

0.9

GaN 40 nm

1

AlGaN:Si 450 nm

0

GaN:Si 40 nm

(b)

Bulk GaN

Figure 8. (a) L-I-V characteristics of PAMBE-grown cw laser diode; (b) laser diode structure.

24

Perspectives of nanoscience and nanotechnology

The devices were processed as ridge-waveguide, oxide-isolated lasers. The mesa structure was etched to a depth of 0.3 µm. The 20 µm- wide and 500 µm-long stripes were used as laser resonators as shown in Fig. 9.

Figure 9. Scanning Electron Microscope image of a cleaved mirror of the laser diode. A 20 µm width mesa etched down to 0.3 µm is visible. The oxidized Ni/Au ohmic contacts were deposited on the top surface of the device, and Ti/Au contacts were deposited on the backside of the highly conducting n-GaN substrate crystal. The cleaved laser mirror facets were coated with symmetrically reflecting mirrors. Figure 8a shows the Light-Current-Voltage (L-I-V) characteristics of the cw LDs with lasing threshold current density and voltage of 5.5 kA/cm2 and 5.7 V, respectively. The laser action was observed up to 60 mW of optical output power (30 mW per facet) at a wavelength of 411 nm [6]. As stated previously, for the group III metal rich regime one can create conditions for a two-dimensional step-flow growth mode, which in principle should give device quality structures by PAMBE. However, until very recently the efficiency of InGaN QWs grown by PAMBE was far lower those obtained by MOVPE. The poor quality of InGaN QWs from PAMBE was mainly related with the growth on high threading dislocations (TDs) density substrates. When low TDs density GaN bulk substrates were used, the two-dimensional step-flow growth mode with parallel atomic steps was observed in PAMBE layers even at low growth temperatures (e.g. for In0.02Ga0.98N layer grown at 600oC – see Figure 5). Therefore, the smooth interfaces required for LDs can be obtained as shown in Figure 10 [14].

Sl

s

G In

: aN

In

M

g

N Ga l A

:M

In

g

G

Q 3x

W

}

i :S N a

Figure 10. Transmission Electron Microscope image of the active region of a PAMBE laser diode with 3 QWs.

Solid State Phenomena Vol. 140

25

The emergence of low temperature PAMBE technology as a viable alternative for the fabrication of blue-violet leaser diodes discussed in this paper may signal the end of the exclusive domination of MOVPE in this field of technology. This is highly desirable, since the state of the art in this sector still lags considerably behind GaAs or InP based laser devices. The resulting competition is bound to accelerate the refinement of the technology, much as it did for the better established material systems. Even more importantly, PAMBE, in combination with high quality GaN substrates, due to their unique capability for 2D growth at low temperatures, may open the way to the highperformance GaN-based green lasers, the target which appears to be incompatible with MOVPE growth Summary The principles of the low temperature growth of the GaN and InGaN layers by plasma assisted MBE have been described. The role of the dislocation mediated growth on the morphology of the PAMBE grown layers has been presented. The continuous wave blue-violet laser diodes grown by PAMBE on low dislocation density bulk GaN substrates has been demonstrated. These diodes operate at 411 nm wavelength with a cw optical power of 60 mW, with lasing threshold current density and voltage of 5.5 kA/cm2 and 5.7V, respectively. Acknowledgements This work was partially supported by the Polish Ministry of Science and Higher Education Grant Number 3T11B04729. References [1] S. Nakamura, G. Fasol, and S. J. Pearton, The Blue Laser Diode: The Complete Story, 2 ed. (Springer-Verlag, 2000). [2] A. Ishizaka, Y. Murata, J. Phys.: Condens. Matter 6, L693 (1994) [3] B. Heying, I. Smorchkova, C. Poblenz, C. Elsass, P. Fini, S. DenBaars, U. Mishra, J.S. Speck, Appl. Phys. Lett. 77, 2885 (2000) [4] C. Adelmann, J. Brault, D. Jalabert, P. Gentile, H. Mariette, G. Mula, and B. Daudin, J. Appl. Phys. 91 (2002) 9638. [5] J. Neugebauer, T. K. Zywietz, M. Scheffler, J. E. Northrup, H. Chen, and R. M. Feenstra, Phys. Rev. Lett. 90 (2003) 056101 [6] C. Skierbiszewski, P. Wisniewski, M. Siekacz, P. Perlin, A. Feduniewicz-Zmuda, G. Nowak, I. Grzegory, M. Leszczynski, and S. Porowski, Appl. Phys. Lett. 88 (2006) 221108 [7] I. Grzegory and S. Porowski, Thin Solid Films 367 (2000) 281 [8] W. Utsumi, H. Saitoh, H. Kaneko, T. Watanuki, K. Aoki, and O. Shimomura, Nature Materials 2 (2003) 735 [9] J. A. Van Vechten, Phys. Rev. B 7 (1973) 1479 [10] R. C. Shoonmaker, A. Buhl, J. Lemley, Journal of Physical Chemistry 69, 3455 (1965). [11] B. Heying, E. J. Tarsa, C. R. Elsass, P. Fini, S. P. DenBaars, and J. S. Speck, J. Appl. Phys. 85 (1999) 6470 [12] S. Haffouz, H. Tang, J. A. Bardwell, P. Lefebvre, T. Bretagnon, T. Riemann, and J. Christen, J. Appl. Phys. 100 (2006) 013528 [13] J. Wu, W. Walukiewicz, K. M. Yu, J. W. Ager, E. E. Haller, H. Lu, and W. J. Schaff, Appl. Phys. Lett. 80 (2002) 4741 [14] C. Skierbiszewski, P. Perlin, I. Grzegory, Z.R. Wasilewski, M. Siekacz, A. Feduniewicz, P. Wisniewski, J. Borysiuk, P. Prystawko, G. Kamler, T. Suski,S. Porowski, Semicond. Sci. Technol. 20, 809 (2005)

OXIDE NANOSTRUCTURES

Solid State Phenomena Vol. 140 (2008) pp 29-36 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.29

Stability, Instability, Metastability and Grain Size in Nanocrystalline Ceramic Oxide Systems Giora Kimmel a and Jacob Zabicky b Institutes for Applied Research, Ben Gurion University of the Negev, POB 653, Beer Sheva 84105, Israel e-mail: a [email protected]; b [email protected] Keywords: Nanocrystalline oxides; amorphous oxides; metastable ceramic oxides; stability of ceramic oxides; thermodynamics of nano structural materials; magnesium titanates; Titania; zirconia-alumina; synthesis of ceramic oxides by sol-gel technique; X-ray powder diffraction; HTXRD.

Abstract. The following nanocrystalline binary oxide systems were studied: Mg-Ti, Ni-Ti, Zr-Al, as well as some pure and doped unary oxides. The xerogels were heated at a constant T (200 to 1600°C) for 3 to 6 hours. There was a threshold tempearture for oxide formation and in many cases the products were metastable nanocrystalline phases, depending on the grain size and composition, including doping. The oxide phases of Ni-Ti, Mg-Ti, and Zr-Al, formed at 900 °C are different from those formed at higher temperature. New ranges of solid solutions and the formations of higher temperature structures were found. A transition phase can be defined as a structure formed at relative low tempearture, irreversibly transforming at higher temperature into an equilibrium phase of the same elemental composition. Some low temperature transition phases have a structure similar to that of a high temperature equilibrium phase, e.g., (the equilibrium phase is given in parentheses) tetragonal ZrO2 (monoclinic) and low-T qandilite-like solid solutions (qandilite + geikielite). Others are unique with no representation in the equilibrium phase diagram, e.g., gamma-like alumina (corundum) and anatase (rutile), which are formed as nanocrystalline oxides due to a low growth rate caused either by a low temperature of calcination or due to additives. To asses the importance of crystal size in the stabilization of transition phases, the following studies were undertaken: (a) XRPD analysis of all unary, doped and binary compositions; (b) the evolution of transition phases in HT XRPD of the Mg titanates; (c) the phase evolution was studied with time at temperatures were mixtures of transition and equilibrium phases were found; (d) the retention of pure tetragonal ZrO2 on quenching Al-Zr oxides after calcinations at high tempetature; (e) additional evidence from HRTEM, SEM and DTA experiments was also collected. A model, correlating the size effect with the unusual phases and structures is proposed. Introduction The retention of the high temperature phase after cooling to room temperature is very common in metal oxides. However, it does not occur in every system. As an example, qandilite (Mg2TiO4) formed above 1100 °C as a high-temperature phase which should decompose at lower temperature to MgO and MgTiO3, is easily retained at room temperatures even after slow cooling [1,2]. The qandilite decomposes to the low-temperature state after prolong thermal annealing below 1000 °C [3]. On the other hand, ZrO2 which is formed above 1200 °C is not retained in a tetragonal structure at room temperature without rapid cooling or adding other elements that are known as tetragonalzirconia stabilizers [4,5,6]. Some metal oxides which are obtained only at high temperatures are stable at low temperatures and cannot transform to any low-temperature phase which exists in the same system. For examples, TiO2 is formed as anatase at low temperatures and transforms irreversibly to rutile above 900 °C. This transformation is accompanied with particle coarsening

30

Perspectives of nanoscience and nanotechnology

[7,8,9]. Alumina is formed as the gamma-alumina phase at low temperatures and transforms irreversibly to corundum above 1000 °C [10]. To date, it is possible to identify two different cases of phases found outside the range of stability: a high-temperature phase retained at low temperatures and a low-temperature phase that transform irreversibly to another stable phase. However, both cases are included in the category of metastable phases. In the present study a new approach is presented in order to define the conditions for the stability for ceramic oxide phases, formed by the sol-gel technique and formed as non-equilibrium nano structured materials. Experimental Materials. The following nanocrystalline binary oxide systems were studied: Mg-Ti, Ni-Ti, , Zr-Al, as well pure or doped unary oxides were also studied. The xerogels were heated at constant temperature between 200 and 1600 °C) for 3 to 6 hours. Characterization. The main characterization tool was X-ray powder diffractometry. The equipment used included: (a) Conventional Kα-beam Bragg-Brentano diffractometer; (b) Guinier Kα1-beam digitized (imaging plate) camera. Both scanned from Cu characteristic radiation. (c) Hot X-ray powder diffraction. The powder was loaded on platinum ribbon which was the sample holder and the heat source. The data collected in linear position sensitive detector. The diffraction scans were performed in 22 cycles of 300 s runs during isothermal treatments from 500 to 1200 °C. All diffraction data was analyzed by advanced data processing methods including Rietveld refinement [11] by using the pogrammes DBWS [12] and FullProf [13] for strucural and quantitative analysis and Ritquan [14] for size-strain analysis. Some samples were observed by SEM, TEM and DTA. Results The HT qandilite phase which was formed at high temperature can be retained at room temperature even after slow cooling. According to the phase diagram of the MgO-TiO2 system, qandilite (Mg2TiO4) exists only above 1100 °C. At lower temperatures the equilibrium state is a mixture of MgO and geikielite (MgTiO3). A sol-gel co-precipitated at room temperature of MgO+TiO2 with Mg:Ti in a ratio of ~1.7, thermally annealed above 1150 °C and cooled to room temperature resulted in a mixture of 0.7Mg2TiO4+0.3MgTiO3, which is the high temperature state. The qandilite is a metastable phase (frozen from high temperature). However, at above 700 °C it gradually decomposes to MgO and geikielite, and after annealing at 1100 °C a mixture 0.4MgO+0.6MgTiO3 was obtained, without qandilite as predicted by the equilibrium phase diagram. These experiments readily confirm the equilibrium phase diagram, but also demonstrate the possibility of obtaining qandilite at room temperature. It seems that the qandilite will be absent in all samples calcined below the eutectoid point ~1150 °C. However, at temperatures below 800 °C, the qandilite was formed even in non-stoichiometric samples. For samples calcined at 600 °C for 3 hours, a single qandilite phase formed at the range of Mg:Ti between 1.2 to 2.0. This qandilite which was obtained directly during calcinations below the eutectoid point appears to be in a non-equilibrium state, a classic case of the second type of nonequilibrium phase. The differences between the two types are summarized in Table 1.

Solid State Phenomena Vol. 140

31

HT qandilite

LT qandilite

T of formation

above 1150 °C

below 800 °C

grain size

~100 nm

~10 nm

composition

stoichiometric

range of solubility

decomposition at 1100 °C

slow

fast

Table 1: Comparison between HT and LT qandilite The two non equilibrium phases are shown in Fig. 1. More details about these experiment had been published elsewhere [15,16,17].

c

b

a

Fig. 1: XRPD diffractogram of sol-gel product of Mg,Ti oxides, where Mg:Ti ~1.7, calcined 3 h at 600 °C. Diffractograms: (a) annealed 3 h at 700 °C. Nanocrystalline (~15 nm grain size) non-stoichiometric qandilite-like Mg2-2xTixO4 single phase; (b) annealed 3 h at 1100 °C. A mixture of MgTiO3 (geikielite) and MgO (c) annealed 3 h at 1300 °C. A mixture of qandilite and geikielite. At (b) and (c) the diffractograms are sharp and the X-ray quantitative analysis confirm the equilibrium states. Using HTXRD it was found that with compositon close to the qandilite composition the qandilite formed at an early stage of heating at all temperatures. The non-stoichiometric solid solution in the qandilite-like structure was found only in diffractograms obtained at 700 °C and lower temperatures. These diffractograms were always broadened, due to the crystal size being less than 15 nm. (See Fig. 2). At higher temperatures the non stoichiometric qandilite was mixed with geikielite.

32

Perspectives of nanoscience and nanotechnology

1300 °C

600 °C

Fig: 2: Diffractograms from HTXRD of sol-gel product of Mg,Ti oxides, where Mg:Ti ~1.7 at temperatures from 600 °C up to 1300 °C with step of 100 °C, during the first run (~300 s annealing time) . The peak at the range of 39-40 ° is from the Pt ribbon used as a heater and sample holder. The non-stoichiometric qandilite-like single phase appears as a single phase up to 800 °C. However, the qandilite is still a major phase at all temperatures below the eutectoid point, exhibiting non-equilibrium state. Another example of the second type of low-temperature metastable phase of zirconia was found in non-equilibrium quasi amorphous, tetragonal and cubic states after calcinations of zirconia-rich zirconia-alumina xerogels below 1300 °C [17] . In this system it was found that the HT zirconia states were stabilized by both, the formation at low temperature and the amount of alumina. The effect of the temperature of formation is shown in Fig. 3, and the effect of the amount of alumina is shown in Fig. 4.

1500 °C

1100 °C

700 °C

300 °C

Fig. 3. XRPD diffractograms of sol-gel product of Zr, Al oxides, calcined 3 h at temperatures from 300 °C (bottom) up to 1500 °C (top) with step of 200 °C. At 1300 and 1500 °C a mixture of monoclinic zirconia + 0.8 % mol corundum was obtain, confirming equilibrium state. At 300 and 500 °C, the structure was quasi-amorphous zirconia. At 700 and 900 °C nano crystalline tetragonal zirconia was found. At 1100 °C a mixture of tetragonal zirconia and monoclinic was found.

Solid State Phenomena Vol. 140

33

The non-equilibrium structures of zirconia gradually return to the equilibrium state by raising the temperature of formation (Fig. 3) or reducing the amount of alumina (Fig. 4).

Fig 4. XRPD diffractograms of sol-gel product of Zr, Al oxides, calcined 3 h at 900 °C. The % molar amount of the alumina is marked on the diagram from 1% at the bottom to 100% at the top. At 1% mol alumina the structure is monoclinic zirconia. At 2% mol alumina a tetragonal zirconia appears. From 3% alumina the monoclinic phase disappears and the tetragonal zirconia becomes a single non-equilibrium phase. The tetragonal grain size decreases with increased amounts of alumina. Above 25 % mol alumina the alumina solubility was at saturation and with higher alumina amounts gamma alumina was found. (The diffractogram at the top is for pure gamma alumina). Recently, a stabilized tetragonal pure zirconia was produced by calcination of alumina-zirconia xerogel at high temperatures (1300, 1500 °C) for 3 hours. At these temperatures there is no solubility effect and the phase analysis showed a mixture of pure alumina (corundum) and zirconia. A Rietveld diagram of the phase analysis is shown in Fig. 5.

Fig 5. Xerogel of with Zr and Al cations was calcined 3 h at 1300 °C. Rietveld (FullProf) diagram of zirconia alumina mixture is shown. The analysis showed complete separation of alumina (corundum) and zirconia. The amount of alumina was 64 % mol 24 % mol tetragonal zirconia and 12 % mol monoclinic zirconia.

34

Perspectives of nanoscience and nanotechnology

At high alumina amounts (above 50%) the tetragonal zirconia was partially stabilized at room temperatures. The stabilization of the tetragonal zirconia was correlated with the amount of alumina. Since alumina is not dissolved in the zirconia at this temperatures, the phenomena of retained tetragonal called for additional investigations. It was found by SEM and XRD that the zirconia grain size decreases when the amount of alumina was increased. A gradual transition from anatase to rutile was found in titania sol-gel products calcined between 800 to 900 °C, see Fig. 6.

1300 °C

300 °C

Fig. 6. XRPD diffractograms of sol-gel product of Ti oxide, calcined for 3 hours at temperatures from 300 °C (bottom) up to 1300 °C (top) at steps of 100 °C. In the lower temperature range, up to 800 °C, the structure is anatase with increased grain size with increasing temperature (deduced from a line broadening analysis). At 1000 °C and above the titania becomes rutile. At 800 and 900 °C, both polymorphs, anatase and rutile were detected. A correlation between the anatase grain size and the amount of rutile, as obtained from XRD is illustrated in Fig. 7. DTA scans failed to find an endothermic peak in titantia started as anatase and transformed to rutile.

% Rutile

100 80 60 40 20 0 0

20

40

60

80

100

Anatase grain size [nm]

Fig. 7: Amount of rutile versus anatase grain size Discussion The gradual transition from the metastable to stable structures as a function of temperature, as shown for qandilite and anatase, and concentration, as shown for zirconia, led to rejection of the idea that the temperature and concentration are thermodynamic variables which may be used as coordinates in the equilibrium phase diagram, because the transition of a single component should be sharp. Moreover, DTA scans failed to find endothermic peaks in the transition from anatase to

Solid State Phenomena Vol. 140

35

rutile in increasing temperature. It is considered that due to the contribution of surface energy, an variable which directly controls the structure should be introduced. A similar approach is referred to in literature [18]. It is suggested that particle size should be used as an additional thermodynamic variable. This requires splitting the free energy term for a particle (Eq. 1) into volume and surface energies (Eq. 2), and then the free energy term for a particle can be written as Eq. 3. (1)

G = E + PV − TS

Where: G = free energy; E = internal energy; P = pressure; V = volume; T = temperature; S = entropy a (2) E = ρ (eb v + es a ) = m(eb + es ) v Where for a single particle: α = particle area; ν = particle volume; m = particle mass; ρ = particle density; eb = bulk energy per unit of mass a es = surface energy per unit of m v

G = Eb + Es Λ + PV − TS Where: Λ =

(3)

a ; Es = esm; Eb = ebm v

With large grain Λ →0 the matter has bulk properties. When Λ is very large as for nano structural materials, phases with low surface energies like amorphous/cubic/tetragonal zirconia and anatase, as published by Navtotsky [19] and Navrotsky et al [20], are favoured, they should be regarded as stable phases! All crystal properties may be functions of Λ including the unit cell parameters, which are functions of conventional thermodynamic variables (temperature, pressure, composition). Some publications report on variations of cell parameters for nanocrystalline oxides [21,22,23]. In ionic crystals, the unit cell volume tends to increase when particle size decreases (Λ increases). Usually, the cell parameters start to increase with a grain size below 30 nm [22,23]. Initial calculation of the attractive potential for a one dimension ion arrangement, as a final segment, shows that there is a sharp decrease in the attractive potential, when the number of pairs is below 100 (See Fig. 8). 0.70

Factor

0.65 0.60 0.55 0.50 1

10

100

1000

10000

2N (N is number of pairs)

Fig. 8: Half of Madelung constant as a function of number of cation/anion pairs for a line of alternating cations and ions. For an infinite number of pairs this number is ln2. For a typical bond length of 0.3 nm, it means length of the ions segment of 30 nm.

36

Perspectives of nanoscience and nanotechnology

Summary Grain size is a crucial factor in stabilizing metastable phases. It is suggested that grain size should be used as a thermodynamic factor via the new coordinate Λ. Then, phases with lower grain size which are regarded as metastable will be considered stable, confined by a range of Λ. The gradual transitions versus grain size can be attributed to non-uniform grain size. The morphology and crystal structure of nano structural materials is dictated by a minimum surface/volume ratio. References [1] E.M. Levin, C.R. Robbins and H.F. McMurdie: Phase Diagrams for Ceramists (American Ceramic Society, Westerville, OH 1964). [2] B.A. Wechsler and A. Navrotsky: J. Solid State Chem. Vol. 55 (1964), p. 165 [3] R.L. Millard, R.C. Peterson and B.K. Hunter: Am. Mineral. Vol. 80 (1995), p. 885 [4] U. Martin, H. Boysen and F. Frey: Acta Cryst. B Vol. B49 (1998), p. 403 [5] M. H. Bocanegra-Bernal and S. D´iaz de la Torre: J. Mater. Sci. Vol. 37 (2002), p. 4947 [6] L.H. Schoenlein, L.W. Hobbs and A.H. Heuer: J. Appl. Cryst. Vol. 13 (1980), p. 375 [7] A. A. Gribb and J. F. Banfield: Am. Mineral. Vol. 82 (1997), p. 717 [8] H. Zhang and J. F. Banfield: Am. Mineral. Vol. 84 (1999), p. 528 [9] H. Zhang and J. F. Banfield: J. Mater. Res. Vol. 15 (2000), p. 437 [10] R.S. Zhou and R.L. Snyder: Acta Cryst. Vol. B47 (1991), p. 617 [11] H. M. Rietveld: J. Appl. Crystallogr. Vol. 2 (1967), p. 65 [12] R.A. Young, A. Sakthivel, T.S. Moss, and C.O. Paiva-Santos: J. Appl. Crystallogr. Vol. 28 (1995), p. 366. [13] J. Rodrigez-Carvajal: Fullprof, Program for Rietveld refinement, Laboratories Léon Brillouin (CEA-CNRS), Saclay, France, 1997. [14] L. Lutterotti, and P. Scardi: J. Appl. Cryst. Vol. 25 (1992), p. 459 [15] G. Kimmel and J. Zabicky: Mater. Sci. Forum Vols. 278-281 (1998), p. 624 [16] G. Kimmel and J. Zabicky: Adv. X-ray Anal. Vol. 42 (1998), p. 238 [17] G. Kimmel, J. Zabicky, E. Goncharov and P. Ari-Gur: J. Metast. Nanocryst. Mater. Vols. 2021 (2004), p. 576 [18] K.S. Pitzer and L. Brewer: Thermodynamics (MvGraw-Hill Book Company, USA 1961). [19] A. Navrotsky: J. Mater. Chem. Vol. 15 (2005), p. 1883 [20] A. A. Levchenko, G. Li, J. Boerio-Goates, B. F. Woodfield and A. Navrotsky: Chem. Mater. Vol. 18 (2006), p. 6324 [21] S. Tsunekawa, R. Sivamohan, S.I. Ito, A. Kasuya and T. Fukuda: Nanostr. Mater. Vol. 11 (1999), p. 141 [22] F. Zhang, S.W. Chan, J.E. Spanier, E. Apak, Q. Jin, R.D. Robinson, and I.P. Herman: Appl. Phys. Let. Vol. 80 (2002), p. 127 [23] G, Kimmel, J. Zabicky, E. Goncharov, D. Mogilyanski, A. Venkert, Y. Bruckental, and Y. Yeshurun: J. of Alloys and Compounds Vol. 423 (2006), p. 102

Solid State Phenomena Vol. 140 (2008) pp 37-46 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.37

Mesoporous Crystals of Transition Metal Oxides Wuzong Zhou School of Chemistry, University of St Andrews, St Andrews, Fife KY16 9ST, UK. e-mail: [email protected] Keywords: Transition metal oxides, porous materials, electron microscopy.

Abstract. Using mesoporous silicas as hard templates and facilitating crystal growth of transition metal oxides inside the pores, some mesoporous crystals can be produced after removing the templates. This paper gives a brief review of the research of mesoporous crystals of transition metal oxides in the last five years, including the technical development and potential applications of the new form of oxides. Introduction Considering the dimensions of nanomaterials, nanoparticles can be regarded as 0 dimensional (0D) nanomaterials. When nanoparticles extend in only one dimension, they become nanowires. If they extend two dimensionally, they may form thin films. Both nanowires and thin films maintain more or less nanomaterial properties. When nanoparticles grow three dimensionally, normal bulk crystals are developed and their nanomaterial properties are lost. However, there is one way to keep nanomaterial properties when the crystals extend three dimensionally, that is by forming mesoporous crystals templated by mesoporous silica. Ordered mesoporous silica was first reported in early 1990s [1,2]. Since then, many types of mesoporous silicas with various pore networks have been synthesized. The most popular phases include MCM-41 [1,2], SBA-15 [3,4], KIT-6 [5], SBA-16 [4,6,7] and FDU-12 [8,9]. These materials have been used as hard templates for growing crystals of oxides in the last five years, because they contain regular pores with uniform and tunable pore sizes. Indeed, mesoporous transition metal oxides can also be prepared using organic soft surfactants as templates, similar to the synthetic methods for mesoporous silicas. However, the structures of these oxides are either amorphous, or have an amorphous framework with some nanocrystallites since crystallization of most metal oxides requires high temperature and the soft templates would decompose before the crystallization is completed [10]. Both SBA-15 and MCM-41 contain arrays of cylindrical pores (Fig. 1a).

Fig. 1.

Schematic drawing of the pore systems in four mesoporous silicas commonly used as templates, (a) SBA-15, (b) KIT-6, (c) FDU-12 and (d) SBA-16.

hard

38

Perspectives of nanoscience and nanotechnology

The difference is that the main pores are connected by small channels in SBA-15 and no such bridging channels in MCM-41. In KIT-6, there are two sets of branch-like pores, the so-called bicontinuous pore system (Fig. 1b). Both FDU-12 and SBA-16 have spherical cages in face-centred cubic and body-centred cubic arrangements, respectively. The cages are connected by small windows as shown in Fig 1c and d. When a metal oxide replicates the pore networks, the morphologies of the products are expected to be hexgonally arranged nanorods linked to each other by small bridges (SBA-15 templated), nanowires in an Ia3 d symmetry (KIT-6 templated), or nanospheres connected by small bridges (FDU-12 or SBA-16 templated). These porous oxides may have a large surface area and some physical and chemical properties in between those of nanoparticles and bulk materials. Therefore, it can be expected that these new forms of materials will have high potential applications in industry as catalysts, gas separation materials and gas sensor materials. The method of replication of the pore systems in the mesoporous silicas was first reported by Korean scientists 10 years ago [11]. However, the crystal growth of metals and carbon inside the nanoscale pores normally can not extend for a long distance. Consequently, mesoporous single crystals in a reasonable size (e.g. 100 nm) could not be fabricated. However, crystal growth of metal oxides is much more controllable and many transition metal oxides have been produced in this porous form. In this paper, the syntheses of porous crystals of oxides templated by different mesoporous silicas are reviewed and the potential application of the materials and future development are discussed. Impregnation of precursors In order to grow crystals of a metal oxide in mesopores, the selection of a suitable precursor and a method to introduce it into the pores are the two most important steps. In the first report about mesoporous crystals of transition metal oxides, a surface functionalisation method was used for producing porous Cr2O3 using SBA-15 as a template [12]. This method is relatively more complicated compared to the methods developed subsequently. It included functionalising the inner surface of the mesoporous silica template by aminosilylation of the surface silanols and anchoring a selected heteropolyacidic precursor, such as H2Cr2O7 and H3PW12O40 [12-14]. The functionalised surface is positively charged and the metal-containing ions are negatively charged (Cr2O72- and PW12O403-). Therefore the driving force of migration of the precursor is mainly ionic attraction. The disadvantage of this method is that suitable heteropolyacids for some metals are not available. The advantage is that the loading level is high. For example, porous Cr2O3 templated by SBA-15 using this method has a similar particle size and morphology to those of the template SBA-15 particles, implying that the original pore of SBA-15 can be almost fully filled by the precursor and Cr2O3 crystals replicate the pores and the template’s particle size and morphology as shown in Fig. 2 [12].

Fig. 2. TEM images of porous Cr2O3 templated by SBA-15. The image on left shows typical particle size and morphology. The insets are corresponding SAED patterns from Cr2O3 crystal structure (left) and ordering of the nanorods (right). The right image with a higher magnification shows the nanorods.

Solid State Phenomena Vol. 140

39

One of the possible reasons is that the anionic precursors have high densities comparing with more commonly used metal nitrates. Two ordered structures can be observed from these crystals as seen by the selected area electron diffraction (SAED) patterns in Fig. 2, one is from the crystal structure of the oxide and the other from the ordering of the pores. The second method for the impregnation of precursors into the templates is the so-called dual solvent method, first reported by Imperor-Clerc, et al [15]. They mixed a suspension of SBA-15 in dry hexane with an aqueous solution of manganese nitrate, and stirred it overnight. A very high loading level of the nitrate precursor was believed to be achieved, e.g. a 97% filling of the porosity of SBA-15 according to the results of N2 adsorption/desorption. However, gas sorption can not directly prove high loading. In fact, TEM images seem to show a low filling of metal oxide in the pores. The real loading level is probably not so high because some pores may be blocked by the precursor molecules and are not detectable by the gas adsorption/desorption method. Zhao’s group added indium nitrate into the synthetic system for making mesoporous silica and a hybrid monolith was obtained after thermal treatment, which was then calcined at high temperature to remove the organic surfactants and simultaneously allowed crystal growth of In2O3 in the pores [16]. This one-step nanocasting method has not been used for syntheses of other porous metal oxides. A more popular method is called the evaporation method. A mesoporous silica template is mixed with a metal nitrate in ethanol. It is expected that during the evaporation of ethanol by capillary action the nitrate precursor would enter the pores [17-19]. However, this understanding is not correct. According to this mechanism, all metal nitrates soluble in ethanol could be used as precursors for the preparation of porous crystals. In fact, some metal nitrates do not work using this method. For example, using Pb(NO3)2 as a precursor, no porous crystals were obtained [20]. XRD and TEM examinations of the specimens at different stages of synthesis stages revealed that the migration of the nitrate precursors into the pores did not take place during the evaporation of ethanol. Instead, the nitrate was deposited on the surface of mesoporous silica particles. When the specimen was heated to above the melting point of the nitrate, the precursor, in a liquid phase, would move into the mesopores by capillary action. As a consequence, if a metal nitrate decomposes before melting, it is not a suitable precursor for making porous crystals of metal oxide by the evaporation method. Pb(NO3)2 decomposes at about 200 ºC before melting and this is the reason why it can not be introduced into the pores of any mesoporous silica by the evaporation method [20]. Based on the above argument and a previously reported solid state grinding route [21], a solventfree solid-liquid method has been developed for nitrate precursors with low melting points, in which the precursor is mixed directly with the silica template in the solid state. When the specimen is heated to the melting point of the precursor, it turns to a liquid phase and moves into the pores by capillary action. The limitation of this method is that a suitable precursor must have a melting point lower than its decomposition temperature [20, 22]. The disadvantage of this method is that the distribution of the precursor might be poor in comparison with the evaporation method. Therefore, temperature control in the solid-liquid method is important during the thermal treatment. After the impregnation of precursors, the chemical processes in all the methods are more or less the same. With increasing temperature, the precursors will decompose inside the pores and, via some possible intermediate phases, metal oxide crystals will finally grow to fill the pores. The silica templates are finally dissolved using 10% aqueous HF solution at room temperature or a 2M NaOH solution at 70 – 90ºC. In the following sections, the formation of the mesoporous crystals of transition metal oxides in different templates will be discussed.

40

Perspectives of nanoscience and nanotechnology

Porous metal oxides templated by SBA-15 SBA-15 has a similar pore structure to the first mesoporous silica, MCM-41, as both contain an array of hexgonally arranged parallel cylindrical pores as seen in Fig. 1a. However, the pores in MCM-41 templated by a cationic quaternary ammonium surfactant are small, typically being 1 to 3 nm; the walls are thin and the pores are not connected to form a three dimensional network [1,2]. Although the pore size of MCM-41 can be tuned by varying the reaction temperature and time [23], and enlarged by using a swelling agent, e.g. 1,3,5 trimethylbenzene (TMB), MCM-41 has no three dimensionally connected pore system and therefore is not a suitable template for porous crystals. Nanorods are normally produced from the MCM-41 template. The pores in SBA-15, templated by using triblock copolymers such as poly (ethylene oxide)-poly (propylene oxide)-poly(ethylene oxide) chains, (abbreviated to PEO-PPO-PEO), are much larger, 5 – 10 nm in diameter, the walls are thicker and all the pores are normally connected by small channels [3,4]. When these small channels are blocked, nanorods of metal oxides will be produced [24]. When the three dimensional pore system is available in SBA-15, the porous crystals of metal oxides can be fabricated. The morphology of the oxides is the same as the original pore structure. Figure 3 shows TEM and high resolution TEM (HRTEM) images of mesoporous Cr2O3 templated by SBA15, showing randomly located small bridges in between the main nanorods and the single crystal property, i.e. any nanorod or small bridge is a part of the same crystal [12]. To date porous crystals of many other transition metal oxides have also been made using SBA-15 as a template, e.g. Co3O4 [19], WO3 [13], In2O3 [16] etc.

Fig. 3 (a) TEM image of porous Cr2O3 templated by SBA-15, showing small bridges connecting the nanorods. (b) HRTEM image of the same sample showing the single crystal property.

It is interesting that the crystal growth phenomena in silica mesopores differs from that in a large container. Dickinson, et al., found that when the precursor Cr(NO3)3·9H2O was loaded into the SBA-15 pores, the crystallization of Cr2O3 started at 350°C and no other crystalline phases appeared during the whole process of thermal treatment. However, heating the precursor Cr(NO3)3·9H2O at 350°C without the presence of the silica template led to a Cr2O5 crystalline intermediate phase. To obtain Cr2O3 in the latter case, the temperature must be increased to 400ºC or higher. This confinement effect of nanoreactors to crystal growth, in this work the role of the silica mesopores, seems to be universal and is also found in the growth of other oxides. Another example is Co3O4. Using Co(NO3)2·6H2O as the precursor, crystal growth of Co3O4 inside the mesopores can take place at 150 ºC following the formation of some intermediates phases, such as crystalline Co(NO3)2·2H2O and Co(OH)NO3·H2O. However, heat treatment of cobalt nitrate at 150 ºC without the presence of mesoporous silica leads only to Co(NO3)2·4H2O crystals although

Solid State Phenomena Vol. 140

41

the formation of Co3O4 has been observed when the temperature was increased to 225 ºC. The crystalline intermediates during the latter process are Co(NO3)2·4H2O and Co(NO3)2 [19]. This decomposition process was confirmed by a thermal gravimetric analysis performed by Rumplecker et al [25]. Porous metal oxides templated by KIT-6 KIT-6 is an equally good template for growing metal oxide crystals. As shown in Fig. 1b, the pore structure consists of two sets of pores, depicted in white and black, respectively. Direct experimental evidence of any possible small channels connecting these two pores has never been obtained. When Cr2O3 crystals were grown in KIT-6 after introducing Cr(NO3)3·9H2O using the dual solvent method, we found, in the majority of particles, only one set of pores had been replicated (Fig. 4c, 4d) and a small number of particles replicated two sets of pores in the central areas (Fig. 4a). In both cases, all particles gave a single diffraction pattern as shown in Fig. 4b [26].

Fig. 4. TEM images of KIT-6 templated porous Cr2O3 on two principal zone axes, (a) the [111] and (c) [100], of the KIT-6 related unit cell. (b) Corresponding SAED pattern to (a). (d) Corresponding HRTEM image of (c), showing an atomic image on the [2 21] zone axis of the Cr2O3 structure.

The most extraordinary phenomenon revealed by the images of Fig. 4 is that the crystal orientation seems to have a close relationship with the symmetry of the mesostructure. For example, when looking in the [111] direction of the KIT-6 like mesostructure, a hexagonal pattern of the pores can be seen (Fig. 4a). This direction of view is exactly parallel to the [001] zone axis of the rhombohedral Cr2O3, which also shows a hexagonal pattern (Fig. 4b), although the scales of the above two patterns are significantly different. When the [100] project of the mesostructure was found, showing a square pattern (Fig. 4c), an image down the [2 2 1] zone axis of the Cr2O3 crystal structure was obtained with the contrast pattern being the closest to a square symmetry. These observations seem to indicate a novel confinement effect of a nanoreactor to the crystal orientation. However, such a phenomenon has not been observed with other porous oxides. Among other porous metal oxides templated by KIT-6, such as In2O3 [17,18], iron oxides [27,28] and Co3O4 [19], porous Co3O4 is the second material which has been subjected to in-depth investigation. Using Co(NO3)2·6H2O as the precursor and KIT-6 as the template, porous crystals of the oxides can be regarded as a negative replica of the whole bicontinuous pore system, as clearly

42

Perspectives of nanoscience and nanotechnology

shown by the image contrast [19]. The fact that Cr2O3 replicates one set and Co3O4 can replicate two sets of the KIT-6 pores has also been confirmed by the surface area measurements, i.e. the specific surface area of Co3O4, 277 m2/cm3, is more than double the value of that for Cr2O3, 114 m2/cm3. It is certain that there are some small channels connecting the two neighbouring channels in KIT-6. What is not fully understood is the reason why it is difficult to get the crystal growth of Cr2O3 extending from one pore to the other through the bridging channels. In the report by Dickinson, et al. [19], it was believed that the higher temperature for the crystal growth of Cr2O3 may block the small channels. In later work by Yue et al., it was found that the low symmetry of the oxide crystals of Cr2O3 might result in an anisotropic interaction with the wall and cause it to distort. The small channels were probably blocked by the distortion. More experimental data are needed before the conclusion can be confirmed. Porous metal oxides templated by SBA-16 and FDU-12 The selection of SBA-16 and FDU-12 as templates for making porous oxides is based on the special morphologies of their pores, i.e. spherical cages connected to each other three dimensionally by small windows. The corresponding morphologies of the oxide products are therefore nanospheres connected by small nanorods as shown in Fig. 1. Porous crystals of Co3O4 have been grown in the pores in FDU-12 and SBA-16 [22]. Since SBA-16 has a body-centred cubic structure (space group Im 3 m ) and FDU-12 is face-centred cubic (space group Fm 3 m ), each spherical nanocage is 8 coordinated by the neighbouring nanocages in SBA16, but 12 coordinated in FDU-12 (Fig. 1c and 1d). The small bridges are hardly revealed in the TEM images. However, it is possible at the edge of a particle where two adjacent nanospheres can be identified. Low magnification TEM images show that both the FDU-12 and SBA-16 pore networks have been replicated by Co3O4 (Fig. 5). In the case of FDU-12 templated oxide, the common defects in the fcc structure, irregular intergrowth of cubic close packed (ccp) and hexagonal close packed (hcp) stacking, are also replicated (Fig. 5a).

Fig. 5. TEM images of porous Co3O4 templated by (a) FDU-12 and (c) SBA-16. The insets in (a) and (c) are SAED patterns of crystalline Co3O4 from these two samples. HRTEM images of Co3O4 templated by (b) FDU-12 and (c) SBA-16. The arrow in (d) points to a bridge connecting two nanospheres.

Solid State Phenomena Vol. 140

43

The SAED patterns again indicate single crystal properties. Since the crystal orientation has no relation to the mesostructure, it is difficult to find a view direction parallel to the principal axes of both the crystal structure and the mesopore structure. It has been found that replication of the cages in FDU-12 and SBA-16 is much more difficult than replication of the pores in SBA-15 and KIT-6. Although many other transition metal oxides were prepared in the pores of FDU-12 and SBA-16, e.g. NiO, In2O3, CeO2, etc., the yields of these products were low [29]. The high yield of porous Co3O4 was actually confirmed by measurements of surface areas. Nitrogen adsorption/desorption of Co3O4 templated by FDU-12 and SBA-16, respectively show both type IV isotherms with a hysteresis loop, indicating the mesoporous properties. The specific surface areas of Co3O4 crystals templated by FDU-12 and SBA-16 are 151(1) and 122.4(6) m2/g, respectively, which are much larger than the surface area of Co3O4 templated by KIT-6 (92 m2/g) [22]. Another interesting phenomenon observed was that the crystal symmetry plays an important role during the crystal growth in the nanoscale cages. To date, all the successful fabrications of mesoporous oxides were found to have a cubic structure, such as Co3O4, NiO, In2O3 CeO2 and Mn2O3, although the yields of some were low. All non-cubic oxides could not be replicated in the three dimensionally connected cages in FDU-12 and KIT-6 using the evaporation method or the solid-liquid method. The tested oxides were Cr2O3 (rhombohedral), Fe2O3 (rhombohedral) and MnO2 (tetragonal). Obviously, the impregnation of the precursors for the non-cubic oxides causes no special problems or difficulties. It has been proposed, that during crystal growth inside the cages, there is a significant interaction between the crystals and the wall of the nanocages. Cubic oxides can grow isotropically, filling the spherical cages completely. The interaction between the crystals and the silica framework is on the whole surface area of the nanocages. However, noncubic metal oxide crystals grow and interact with the nanocage wall anisotropically, resulting in larger distortion of the silica framework. Consequently, some windows connecting the nanocages become blocked. Although many metal nitrates can be used as precursors for growing crystals inside the silica mesopores, some other nitrates are not suitable since they react with silica. In this case, mesoporous carbon [30,31], which is fabricated using mesoporous silica as a template, can be used as a hard template for preparing the porous metal oxides [32,33]. In this case, the final products are positive replicas of the mesoporous silicas. Potential applications A porous crystal is a new form for materials, which may have a size similar to that of a normal powder. However, the solid parts of the mesoporous crystals are in the nanoscale. Therefore, it can be expected that the materials will have physical and chemical properties in between those of nanoparticles and bulk specimens. So far the properties and potential applications of these new porous materials have not been extensively investigated. Nevertheless, the reported results of very limited studies are quite interesting. The magnetic behaviour of porous crystals of Cr2O3 was investigated and compared with the behaviour of the bulk material [34]. A significant difference in the magnetic behaviour could be detected. Unlike the bulk material, which exhibits antiferromagnetic behaviour below 308 K, the porous crystal behaved like a nanoparticulate material [35]. With cooling in a zero magnet field before increasing the heat and then cooling with a magnetic field of 0.01 T, the magnetic behaviour was similar above 100 K. It is only under 100 K that the magnetic behaviour changes. Applying porous metal oxides in fuel cells and Li-batteries as electrode materials is another direction of development. In a recent work, it was found that using porous crystals of Cr2O3 as a negative electrode in a Li-battery gave a better electrochemical performance [36]. In addition to the

44

Perspectives of nanoscience and nanotechnology

large surface area in the porous crystals, which is essential for a better reactivity, the novel morphology of the materials also had a major advantage in the specific application. Cr2O3 reacts towards Li through a conversion reaction mechanism leading, upon discharge, to the formation of large metallic chromium nanoparticles (10nm); which are embedded into the Li2O matrix. During the reactions, the volume of the framework changes and the high porosity of the porous crystals gives flexibility to the volume of the framework. It was also found that a copious amount of polymeric materials from electrolyte degradation could fill the pores giving further enhancement of the reactions. Many transition metal oxides have already been used as heterogeneous catalysts in industry or been identified, at the laboratory level, as having high potential for application as catalysts [37-39]. The most remarkable feature of the porous crystals is that the materials have nanomaterials properties, including large specific surface areas, but do not need support materials. Therefore, they can be regarded as self-supported catalysts. The methods to prepare thin films of mesoporous silica have been well established and their use as hard templates enables thin film of mesoporous metal oxides to be prepared. It is possible that these thin films may find their way to applications in gas sensors and batteries. Summary Several methods for the impregnation of metal-containing precursors into mesoporous silicas have been developed during the last few years. It seems to be certain that most transition metal oxides can be produced in porous crystals with different morphologies using various mesoporous silicas as templates for the production. It is expected that these materials have potential applications in catalysis, fuel cell, gas sensors, and Li-batteries, etc. Their physical properties lie between those of nanoparticles and bulk specimens, although currently the knowledge and understanding of these properties is very limited. Acknowledgements The author thanks the EPSRC for financial support; Professor H.Y. He for his long term collaboration; Dr. C. Dickinson and Mr. W.B. Yue for their excellent experimental work. References [1] C.T. Kresge, M.E. Leonowicz, W.J. Roth, J.C. Vartuli and J.S. Beck: Nature Vol. 359 (1992), p. 710. [2] J.S. Beck, J.C. Vartuli, W.J. Roth, M.E. Leonowicz, C.T. Kresge, K.D. Schmitt, C.T.-W. Chu, D.H. Olson, E.W. Sheppard, S.B. McCullen, J.B. Higgins and J.L. Schlenker: J. Am. Chem. Soc. Vol. 114 (1992), p. 10834. [3] D.Y. Zhao, J.L. Feng, Q.S. Huo, N. Melosh, G.H. Fredrickson, B.F. Chmelka and G. D. Stucky: Science Vol. 279 (1998), p. 548. [4] D.Y. Zhao, Q.S. Huo, J.L. Feng, B.F. Chmelka and G. D. Stucky: J. Am. Chem. Soc. Vol. 120 (1998), p. 6024. [5] F. Kleitz, S.H. Choi and R. Ryoo: Chem. Commun. (2003) p. 2136. [6] P.I. Ravikovitch and A. V. Neimark: Langmuir Vol. 18 (2002), p. 1550. [7] Y. Sakamoto, M. Kaneda, O. Terasaki, D.Y. Zhao, J.M. Kim, G.D. Stucky, H.J. Shin and R. Ryoo; Nature Vol. 408 (2000), p. 449. [8] J. Fan, C.Z. Yu, F. Gao, J. Lei, B.Z. Tian, L.M. Wang, Q. Luo, B. Tu, W.Z. Zhou and D.Y. Zhao: Angew. Chem. Int. Ed. Vol. 42 (2003), p. 3146. [9] J. Fan, C.Z. Yu, J. Lei, Q. Zhang, T.C. Li, B. Tu, W.Z. Zhou and D. Y. Zhao: J. Am. Chem. Soc. Vol. 127 (2005), p. 10794.

Solid State Phenomena Vol. 140

45

[10] P.D. Yang, D.Y. Zhao, D.I. Margolese, B.F. Chmelka and G.D. Stucky: Nature, Vol. 396 (1998), p. 152. [11] C.H. Ko and R. Ryoo: Chem. Commun. (1996), p. 2467. [12] K.K. Zhu, B. Yue, W.Z. Zhou and H.Y. He: Chem. Commun. (2003) p. 98. [13] B. Yue, H.L. Tang, Z.P. Kong, K.K. Zhu, C. Dickinson, W.Z. Zhou and H. Y. He: Chem. Phys. Lett. Vol. 407 (2005), p. 83. [14] W. Kaleta and K. Nowinska: Chem. Commun. (2001), p. 535. [15] M. Imperor-Clerc, D. Bazin, M. Appay, P. Beaunier and A. Davidson: Chem. Mater.: Vol. 16 (2004), p. 1813. [16] H.F. Yang, Q.H. Sui, B.Z. Tian, Q.Y. Liu, F. Gao, S.H. Xie, J. Fan, C.Z. Yu, B. Tu and D.Y. Zhao: J. Am. Chem. Soc. Vol. 125 (2003), p. 4724. [17] B.Z. Tian, X.Y. Liu, H.F. Yang, S.H. Xie, C.Z. Yu, B. Tu and D.Y. Zhao: Adv. Mater. Vol. 15 (2003), p. 1370. [18] B.Z. Tian, X.Y. Liu, L.A. Solovyov, Z. Liu, H.F. Yang, Z.D. Zhang, S.H. Xie, F.Q. Zhang, B. Tu, C.Z. Yu, O. Terasaki and D.Y. Zhao: J. Am. Chem. Soc. Vol. 126 (2004), p. 865. [19] C. Dickinson, W.Z. Zhou, R.P. Hodgkins, Y.F. Shi, D.Y. Zhao and H.Y. He: Chem. Mater. Vol. 18 (2006), p. 3088. [20] W.B. Yue, W.Z. Zhou : Chem. Mater. Vol. 19 (2007), p. 2359. [21] Y.M. Wang, Z.Y. Wu, H.J. Wang and J.H. Zhu: Adv. Func. Mater. Vol. 16 (2006), p. 2374. [22] W.B. Yue, A.H. Hill, A. Harrison and W.Z. Zhou: Chem. Commun. (2007), p. 2518. [23] C. Cheng, W.Z. Zhou and J. Klinowski: Chem. Phys. Lett. Vol. 263 (1996), p. 247. [24] K.K. Zhu, H.Y. He, S.H. Xie, X. Zhang, W. Z. Zhou, S. Jin and B. Yue: Chem. Phys. Lett. Vol. 377 (2003), p. 317. [25] A. Rumplecker, F. Kleitz, E.L. Salabas and F. Schuth: Chem. Mater. Vol. 19 (2007), p. 485. [26] K. Jiao, B. Zhang, B. Yue, Y. Ren, X.Y. Liu, S.R. Yan, C. Dickinson, W.Z. Zhou and H.Y. He: Chem. Commun. (2005), p. 5618. [27] F. Jiao, A. Harrison, J.C. Jumas, A.V. Chadwick, W. Kockelmann and P.G. Bruce: J. Am. Chem. Soc. Vol. 128 (2006), p. 5468. [28] F. Jiao, J.C. Jumas, M. Womes, A.V. Chadwick, A. Harrison and P.G. Bruce: J. Am. Chem. Soc. Vol. 128 (2006), p. 12905. [29] W.B. Yue and W.Z. Zhou: J. Mater. Chem. Vol. 17, (2007), p. 4947 . [30] R. Ryoo, S.H. Joo, M. Kruk and M. Jaroniec: Adv. Mater. Vol. 13 (2001), p. 677. [31] H.J. Shin, R. Ryoo, M. Kruk and M. Jaroniec: Chem. Commun. (2001), p. 349. [32] J. Roggenbuck, G. Koch and M. Tiemann: Chem. Mater. Vol. 18 (2006), p. 4151. [33] H.F. Li, S.M. Zhu, H.A. Xi and R.D. Wang: Microp. Mesop. Mater. Vol. 89 (2006), p. 196. [34] C. Dickinson, A. Harrison, J.A. Anderson and W.Z. Zhou, Stud. Surf. Sci. Catal. Vol. 165 (2006), p. 335. [35] S.A. Makhlouf: J. Magn. Magn. Mater. Vol. 272 (2004), p. 1530. [36] L. Dupont, S. Laruelle, S. Grugeon, C. Dickinson, W.Z. Zhou and J-M. Tarascon: J. Power Sources, Vol. 175 (2007), p. 502. [37] T. Yokoyama and N. Fujita: Appl. Catal., A Vol. 276 (2004), p. 179. [38] D. Mehandjiev and E. Nikolovazhecheva: J. Catal. Vol. 65 (1980), p. 475. [39] L. Yan, X.M. Zhang, T. Ren, H.P. Zhang, X.L. Wang and J.S. Suo: Chem. Commun. (2002), p.860.

Solid State Phenomena Vol. 140 (2008) pp 47-54 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.47

Structural, Magnetic and Electronic Properties of Surface Oxidised Fe Nanoparticles J. Przewoźnik1,a, T. Tyliszczak2, b, D. Rybicki1, c, J. śukrowski1, d,W. Szczerba1, e, M. Sikora1, f, Cz. Kapusta1, g, H. Stepankova 3, h, R.F. Pacheco 4,i , D. Serrate 4, j and M.R. Ibarra 4, 5, k 1

AGH University of Science and Technology, Faculty of Physics and Applied Computer Science, (AGH), Mickiewicza 30, Kraków 30-059, Poland 2

Lawrence Berkeley National Laboratory (LBNL), 1 Cyclotron Road, Berkeley, CA 94720, U.S.A. 3

Charles University, Faculty of Mathematics and Physics, Ke Karlovu 3, Prague 12116, Czech Republic

4

Universidad de Zaragoza-CSIC, Facultad de Ciencias, Pedro Cerbuna 12, Zaragoza 50009, Spain 5

Instituto de Nanociencia de Aragon, Universidad de Zaragoza (INA), Pedro Cerbuna 12, Zaragoza 50009, Spain

e-mail: a [email protected], b [email protected], c [email protected], d [email protected], e [email protected], f [email protected], g [email protected] Keywords: Surface oxidised Fe powders; ball milling; structural properties; magnetic and electronic properties;

Abstract. A combined XRD, Mössbauer, SEM, STXM and NMR study of naturally oxidised, ball milled iron powders is presented. The XRD patterns show the peaks of the bcc-Fe phase with the line widths increasing with the milling time. This corresponds to a flattening of the crystallites, as confirmed by SEM, and increased strain due to the accumulation of defects. The effect is consistent with the variation of the Mössbauer line-widths with the milling time. Scanning Transmission Xray Microscopy (STXM) measurements provided oxygen maps of the particles and revealed that the dominant oxide in the nanometric oxide layer is magnetite. The 57Fe spin echo NMR study reveals a dominant signal corresponding to a bcc-Fe core and a much weaker resonance corresponding to a magnetite amount of less than 1%. Introduction The surface of iron metal exposed to air undergoes immediate oxidation [1]. In a humid and/or chemically aggressive environment a variety of hydroxides are formed, whereas in a dry atmosphere at ambient temperature, the main oxidation product is magnetite or its cation deficient form, maghemite. At elevated temperatures, the oxidation process results in the formation of haematite [2]. This paper reports the findings of a study undertaken in order to characterise the morphology, structural, magnetic and electronic properties of these materials. Samples of a commercial iron powder, Alfa Aesar of particle size < 10 µm, milled at vacuum for various times, 112 hrs (denoted as Fe 112h), 224 hrs (Fe 224h) and 336 hrs (Fe 336h), as well as the unmilled material, (denoted Fe 0h) were used. Powder X-ray diffraction (XRD) and Mössbauer spectroscopy were used for the structural characterisation and the morphology was studied by SEM. The scanning transmission X-ray microscopy (STXM) at the oxygen K absorption edge and nuclear magnetic resonance (NMR) were used to determine the type of oxide and its electronic and magnetic properties.

48

Perspectives of nanoscience and nanotechnology

Results and Discussion

Fe 0h Fe 336h

0.8

(2 2 0)

0.1

(3 1 0)

(2 1 1)

0.9

(2 0 0)

Intensity (a.u.)

1.0

(1 1 0)

X-ray diffraction measurements. XRD patterns were collected using a D5000 Siemens diffractometer using a Cu Kα radiation and a graphite secondary monochromator, from 20º to 124º in steps of 0.03º in 2θ and about 20 s per step. Examples of two XRD patterns are shown in Fig.1.

0.0 40

60

80

100

120

2Θ (deg.)

Fig. 1. The normalised X-ray diffraction patterns for the Fe 0h and Fe 336h samples. The refinement of the XRD patterns was carried out by the Fullprof program [3] based on the Rietveld method. In the refinement the Thompson-Cox-Hastings pseudo-Voigt function [4] for peak shape and the polynomial function for the background was assumed. In the refinements of the profile once subtracted, the instrumental broadening from the LaB6 reference, the purely Lorentzian shape of the diffraction peaks was found with no Gaussian contribution in the Thompson-CoxHastings pseudo-Voigt function. The FWHM versus θ dependence was modelled by FWHM Lorentzian = X tan θ + Y / cosθ , where X- Lorentzian isotropic strain parameter and YLorentzian isotropic size parameters were used. The strain related FWHM contribution can be ∆d  180ε  expressed as: ∆X =  is a dimensionless "microstrain" parameter.  , where ε = d  π  Usually, the strain broadening is constrained according to Wilson: ∆( 2θ ) = εN coeff tan θ , and the Scherrer equation was used to model the broadening due to the small crystallite size: N coeff Kλ ∆ (2θ ) = , where ε- strain parameter, K denotes the Scherrer crystal shape constant, DD cosθ π represent the volume averaged diameter of the crystallites in all directions, N coeff = . 180

Surprisingly, zero values for the Y- Lorentzian isotropic size parameters were obtained. To check if the findings were correct the Williamson-Hall plots were drawn [5]. Such plots give a valuable insight into the nature of any structural imperfections present in the sample. One should remember that Lorentzian line profile is implicitly assumed in such plots and the Williamson-Hall plot should give reliable results for our samples. The Williamson-Hall plot is based on the fact that the broadening due to lattice strain (has tanθ dependence) and that due to the size effects (a 1/cosθ dependence) have a different angular dependence.

Solid State Phenomena Vol. 140

49

(3 1 0) (2 1 1)

(2 0 0)

0.3

0.2

(2 2 0)

Y=A+B*X A = –0.03 ± 0.07 B = 0.37 ± 0.10

(1 1 0)

∆FWHM*cos(θ) (deg.)

The negative intercept with the vertical axis (zero intercept to within experimental error) indicate that there is no apparent “size” effect but the clear positive slope of the fitted straight line (B≥0) indicates that microstrain effects are appreciable in the Fe 0h sample; see Fig.2.

0.1

0.0 0.0

0.2

0.4

0.6

0.8

sin(θ)

Fig.2. The Williamson-Hall plot for Fe 0h sample. ∆FWHM is the fitted FWHM minus instrumental FWHMinstr contribution for the given diffraction peak. This plot shows that to within experimental error the grain “size” cannot be estimated using this method (the zero value of the intercept means that the grain size is of the order of a micrometer). As the experimental points in the plot display some scatter around the fitted line some anisotropy of the microstrains is evident.

1.0

(3 1 0) (2 1 1)

1.5

(2 2 0)

Y=A+B*X A = –0.1 ± 0.5 B = 2.0 ± 0.8 (2 0 0)

2.0

(1 1 0)

∆FWHM*cos(θ) (deg.)

For the Fe 336h sample, showing maximumline broadening, see Fig.3, there is a clear positive slope of the fitted straight line (B≥0), which indicates that microstrain effects are large in the Fe 336h sample.

0.5

0.0 0.0

0.2

0.4

0.6

0.8

sin(θ) Fig.3. The Williamson-Hall plot for Fe 336h sample (showing maximum line broadening). The fitted value of the B parameter is 5.5 times larger than for Fe 0h sample indicating much larger microstrain effects. The experimental points in the plot display a greater degree of scatter around fitted line indicating larger anisotropy ((h k l) dependence) of the microstrains.

50

Perspectives of nanoscience and nanotechnology

Mössbauer measurements The 57Fe Mössbauer effect measurements were performed at room temperature in the transmission geometry mode and the conversion electron (CEMS) mode using a constant acceleration type spectrometer and 57Co/Rh as the source. Two examples of the spectra measured in transmission are shown in Fig.4a and Fig.4b for the Fe 0h and Fe 336h samples, respectively. 100

Transmission (%)

Transmission (%)

100

95

90

85

95

90 0

0

x2

x2 -10

-5

0

Velocity (mm/s)

5

10

-10

-5

0

5

10

Velocity (mm/s)

a. Fe 0h b. Fe 336h Fig.4. The Mössbauer spectrum (circles) of the samples at room temperature. The fitted spectrum (solid line) and the misfit (bottom insert) are also shown. The spectra were fitted using the transmission integral in order to take the absorber thickness effects on the line width into account. A single Zeeman sextet of Lorentzian lines was fitted. In all the fits a source line half width Γs = 0.07 mm/s was assumed but the absorber Γa values were fitted and total Γ (=Γs + Γa) was derived. An increase of the linewidth with milling time was observed, as indicated in Fig.4, which is attributed to the increasing number of defects and amount of strain. As the gamma rays of the Mössbauer line energies penetrate micrometric grains easily, it was expected the spectrum consisting of contributions from the metallic core and oxide layers would be obtained. The outermost lines of magnetite, maghemite and hematite should appear beyond -8 mm/s and +8 mm/s. No signal in this range was detected for any of the samples studied meaning that the amount of iron oxides is less than 1%. The Mössbauer measurements were also performed in the conversion electron mode of detection [6]. The CEMS spectrum of the Fe 224h sample at room temperature is shown in Fig.5. As the conversion electrons escaping from the material have the penetration depth of 300-500 nanometers, it was anticipated that a detectable contributions from the oxide layers would be obtained. However, no signal in the range beyond -8 mm/s and +8 mm/s was detected, which is where the outermost lines of magnetite, maghemite and hematite should appear. This indicates that the oxide content is much smaller than 1% or the oxide layer exhibits a lower probability of recoilless absorption of gamma quanta from that of the bulk oxide.

Relative Intensity (%)

Solid State Phenomena Vol. 140

51

105

102

-12

-8

-4

0

4

8

12

8

12

Diff. (%)

Velocity (mm/s) 0.2 0.0 -0.2 -12

-8

-4

0

4

Velocity (mm/s)

Fig.5. The Conversion Electron Mössbauer spectrum of the Fe 224h sample at room temperature. The fitted spectrum (solid line) and the misfit (lower part) are also shown. A comparison of the ε- strain parameter (corrected for the resolution function of the diffractometer) and the Mössbauer linewidth exhibits a very similar tendency for the samples studied, Fig.6. 0.165 0.016

Fe 336h

0.160

ε

Fe 224h

0.155

0.008

0.004

Fe 112h

0.150

Fe 0h

Fe 0h

0.000

Fe 224h

1

Γ (mm/s)

Fe 336h

0.012

Fe 112h

2

3

4

0.145

Sample no. Fig.6. The values of the strain ε parameter and the Mössbauer linewidth Γ for various samples. This indicates that the broadening of the diffraction lines is predominantly strain contributed. The diffraction peaks from oxides were not observed in these samples, which is mainly due to the very small content, below 1% and possible the nanocrystallinity of the oxide layer. Scanning electron microscopy measurements. The measurements with scanning electron microscopy were carried out on the unmilled Fe 0h material and the milled samples. The HITACHI S-3400N SEM equipped with a tungsten cathode was used with an applied voltage of 30 kV. The images for the Fe 0h and Fe 336h samples are shown in Fig.7. Milling causes flattening of the grains to form flakes of sub-micrometric thickness which agglomerates in the form of a "composite".

52

Perspectives of nanoscience and nanotechnology

Fig.7. SEM images of the Fe 0h sample( left) and the Fe 336h sample (right). Scanning Transmission X-ray microscopy measurements. The measurements with soft X-ray transmission microscopy have been carried out for the Fe 224h sample at the synchrotron laboratory ALS, Berkeley. Measurements of differential images of selected grains at the energies corresponding to the oxygen K edge provided the oxygen maps, Fig. 8a.

5

hematite

Absorption [a.u.] a a

4

3 Fe224h 2

magnetite

1

0

-1 520

530

540

550

560

570

Photon Energy [eV]

Fig. 8. (a) Oxygen map of a Fe 224h grain. (b) Image of the grain taken at 539.78 eV. The oxygen K-edge spectrum (bottom) was integrated over the marked area. (c) Comparison of Fe 224h average O absorption spectrum with those of hematite and magnetite.

Solid State Phenomena Vol. 140

53

The oxide layer thickness calculated from the x-ray absorption is of order of 15-25 nm. The spectra could be measured for the individual parts of the grain (Fig 8b). The pre-edge peak is related to the transition to the unoccupied O 2p states hybridised with the Fe 3d states. Comparing the shape of the pre-edge peak with that of magnetite and hematite, Fig 8c, the type of the oxide can be determined. The oxide is predominantly magnetite but there is a small contribution from hematite. The oxide layer is also relatively homogeneous, spectra from different part of this particle are similar and they are identical to the average spectrum measured using macro x-ray beam in total electron yield mode. NMR measurements.

Normalized signal intensity (arb. units)

The 57Fe spin-echo NMR measurements were carried out for the Fe 0h sample at 4.2 K. The spinecho spectra have been obtained and the resonant lines of the metallic core (not shown) and the oxide layer, Fig. 9, could be resolved. The probing depth of the NMR measurement is of several micrometers, so the information is obtained from the bulk of the sample.

maghemite c)

magnetite b)

a)

68

Fe 0h

70 72 74 Frequency (MHz)

76

Fig.9. The 57Fe spin-echo NMR spectra at 4.2K of the iron oxides in the Fe 0h sample, a), magnetite, b) and maghemite, c). The resonance of the metallic iron core is centred at 46.65 MHz. A much weaker signal observed at 70 MHz, Fig. 9, corresponds to the 57Fe resonance in magnetite. The resonance is broadened compared with the reference magnetite sample of sub-micrometric particle size, which can be attributed to the large degree of disorder. It is worth noting that the resonant frequency is slightly lower than that of the reference magnetite, which can be attributed to the influence of exchange interaction of the bcc-Fe core of the particle and/or to strain in the nanocrystalline oxide layer. The integrated intensities of this resonance and the resonant line of the metallic core have been corrected for the frequency response and the NMR enhancement factor of both signals in order to obtain the relative amount of iron in the oxide layer and in the metallic core of particles. This value amounts to 0.4(1)% for the Fe 0h material. Taking the average particle size as a few micrometers, the average magnetite layer thickness of tens of nanometers could be derived. As the resonance lines of hematite and maghemite are located at 71 MHz and 72-74 MHz, respectively, Fig. 9 b,c) and Refs. [7,8], the lack of a pronounced signal in this frequency range shows that the possible amounts of hematite and maghemite oxides are very small compared to the magnetite.

54

Perspectives of nanoscience and nanotechnology

Summary The results of a combined XRD, Moessbauer, SEM, STXM and NMR study of surface oxidised, ball milled iron powders show that the linewidths of the bcc-Fe phase in the particle cores increased with the milling time. This corresponds to the increase of strain due to the accumulation of defects and a flattening of the particles, as confirmed by SEM. The effect is consistent with the variation of the Mössbauer line-widths with the milling time. Scanning Transmission X-ray Microscopy (STXM) measurements provided the oxygen maps of the particles and revealed that the dominant oxide in the nanometric oxide layer is magnetite. The 57Fe spin echo NMR study reveals a dominant signal corresponding to the bcc-Fe cores of the particles and a much weaker resonance corresponding to magnetite in their outer oxide layers. The amount of magnetite is below 1% and its 57Fe resonant frequency is slightly lower than that of the bulk magnetite, which can be attributed to the influence of the bcc-Fe core of the particle and/or to the possibility of strain in the nanocrystalline oxide layer. Acknowledgements The work was supported by the European Commission, grant No. 027827, STREP - MUNDIS and by the Polish Ministry of Science and Higher Education, statutory funds for the Faculty of Physics and Applied Computer Sciences at AGH University of Science and Technology in Cracow. References [1] J. Oh Sei, D.C Cook and H.E Townsend: Hyperfine Interactions 112 (1998) p. 59. [2] R.M. Cornell and U. Schwertmann: The Iron Oxides, Structure, Properties, Reaction Occurrence and Uses (VCH Verlagsgesellschaft mbH, Weinheim, 1996). [3] J. Rodriguez-Carvajal: Physica B 192 (1993) p. 55. [4] P. Thompson, D.E. Cox and J.B. Hastings: J. Appl. Cryst. 20 (1987) p. 79. [5] J.I. Langford, in: Accuracy in Powder Diffraction II, NIST Special Publication 846, edited by E. Prince and J.K. Stalick, Gaithersburg 1992. [6] K. Nomura, Y. Ujihira and A. Vertes: J. Radioanal. Nucl. Chem., Art. 202 (1996) p. 103. [7] M. Matsuura, H. Yasuoka, A. Hirai and T. Hashi: J. Phys. Soc. Jpn. 17 (1962) p. 1147. [8] S.-Joo Lee and S. Lee: New J. Phys. 8 (2006) p. 98.

Solid State Phenomena Vol. 140 (2008) pp 55-60 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.55

Nanosized Barium Hexaferrite Powders Obtained by a Single Microemulsion Technique Tatyana Koutzarova 1, a, Svetoslav Kolev 1, b, Kornely Grigorov 1, 4, c, Chavdar Ghelev 1, d, Ivan Nedkov 1, e, Marcel Ausloos 2, f, Rudi Cloots 3, g, Tadeusz Mydlarz 4, h and Andrzej Zaleski 5, i 1

Institute of Electronics, Bulgarian Academy of Sciences, 72 Tzarigradsko Chaussee, 1784 Sofia, Bulgaria 2

3

SUPRATECS, Sart Tilman, B-4000 Liege, Belgium

LSIC, Chemistry Department B6, University of Liege, Sart Tilman, B-4000 Liege, Belgium 4

5

e-mail:

International Laboratory of High Magnetic Fields and Low Temperatures, 53-421 Wroclaw, Poland

Institute of Low Temperature and Structure Research, Polish Academy of Sciences, 50-422 Wroclaw, Poland a

[email protected]; b [email protected]; c [email protected]; d [email protected]; [email protected]; f [email protected]; g [email protected]; h [email protected]; I [email protected] e

Keywords: barium hexaferrite, magnetic properties, microemulsion process

Abstract. Barium hexaferrite (BaFe12O19) powders of particle size of 130 and 180 nm were synthesized by a single microemulsion technique. The influence of the concentration of Ba2+ and Fe3+ metallic ions in the aqueous phase in the microemulsion system on the particle size distribution, crystallinity and magnetic properties of BaFe12O19 was studied. The coercive force and saturation magnetization of the sample obtained at a lower concentration of metallic cations in the aqueous phase were higher than those of the sample obtained at higher concentration. Introduction The physical properties of an inorganic microstructure are fundamentally related to its chemical composition, size, crystal structure and morphology which, depending on the preparation route, can vary [1]. The most common method of synthesis of nano-sized powders is by the use of a “wet chemistry” process. The advantage of this technique is that the mixing of reagents at a molecular level occurs in a solution. The resulting oxide powders have high specific surface area and, consequently, a high reactivity, which reduces the final treatment temperature and the time of synthesis. Some recent investigations showed the possibility of preparing homogeneous nanosized magnetic oxide powders using a microemulsion process [2-5]. One of the advantages of this technique is the ability to prepare of very uniform precursor particles with less than 10% variability [6]. A microemulsion system consists of an oil phase, a surfactant phase and an aqueous phase. The reverse microemulsion system exhibits a dynamic structure of nanosized aqueous droplets which are in constant formation, breakdown, and coalescence. Each aqueous droplet can act as a nanosized reactor for forming nanosized precipitate particles [1]. Various types of microemulsion systems have been developed in a view to produce nanosized magnetic particles [7-9]. In this work a microemulsion system was used consisting of a cetyltrimethylammonium bromide (CTAB) as a cationic surfactant; n-butanol as a co-surfactant; n-hexanol as a continuous oil phase and an aqueous phase. One of the advantages of using CTAB as a surfactant has to do with the free passage of the OH− ions through the walls of water droplets in both directions. This allows a single microemulsion technique to be employed to produce nanoparticles when the precipitating agent contains OH− ions.

56

Perspectives of nanoscience and nanotechnology

In previous papers we reported the development of a technique based on a single microemulsion with the view of preparing nanosized magnetite [10, 11]. In the single microemulsion method there is only one microemulsion system whose aqueous phase contains metal ions only. One of the advantages of the single microemulsion technique is that it is much less expensive than the classical double microemulsion method. The paper reports on work aimed at preparing barium hexaferrite (BaFe12O19) powder of nanometer particle size with high magneto-crystalline anisotropy (K1 = 3.3 x 105 Jm-3) [12] by the single microemulsion method. Since the conditions of synthesis considerably influence the chemical, structural and physical properties, attention was focussed on studying the influence of the metallic ion concentration in the microemulsion aqueous phase on the structural and magnetic properties of the BaFe12O19 nanoparticles produced. Experimental A water-in-oil reverse microemulsion system was used with cetyltrimethylammonium bromide (CTAB), (24 wt.%) as a cationic surfactant, n-butanol (16 wt.%) as a co-surfactant, n-hexanol (20 wt.%) as a continuous oil phase, and an aqueous solution of metallic ions (40 wt.%). The metallic ions concentrations in the aqueous phase were 0.44 and 0.22 and the molar ratio of Ba to Fe was fixed at 1:10 for all syntheses. In the first stage of the synthesis procedure, co-precipitation occurred when the precipitating solution of NaOH was added to the microemulsion containing an aqueous solution of Ba(NO3)2 and FeCl3. The reaction mixture was kept at an optimal pH of 12. The use of CTAB as a surfactant allows the OH− radicals to pass freely through the walls of the water droplets in both directions, so that co-precipitation and the formation of a precursor for the synthesis of BaFe12O19 can take place. The precipitate obtained was separated in a centrifuge and washed with water and a solution of chloroform and methanol (50 v.% and 50 v.%) to remove the excess surfactant. The hydroxide precursor was then dried and milled. The second stage involved the powder obtained being heated at 580°C for 4 h, then ground and calcined at 900°C for 5 h to ensure complete conversion of the precursor into BaFe12O19. Depending on the metal cations concentration in the aqueous phase of the microemulsion system, the samples were denoted: 0.44 M – mE-I and 0.22 M – mE-II. The barium hexaferrite powder was characterized using X-ray diffraction (XRD) analysis with Cu-Kα radiation and a Philips ESEM XL30 FEG scanning electron microscope. The magnetic measurements were carried out at room temperature using a vibration sample magnetometer (VSM) at a maximum magnetic field of 2.3 x 106 A/m. The high magnetic field measurements (up to 1 x 107 A/m) were performed on a homemade pulsed magnetometer [13]. The magnetic measurements were carried out on an unoriented random assembly of particles.

Results and Discussion The XRD spectra of the BaFe12O19 powders synthesized are presented in Fig. 1. The diffractograms of the two samples correspond to the barium hexaferrite structure, except for the additional minor peaks corresponding to haematite in the case of mE-I. The lattice constants obtained from the XRD spectra are a = 0.5807 nm and c = 2.3391 nm for mE-I and a = 0.5895 nm and c = 2.3221 nm for

Solid State Phenomena Vol. 140

57

mE-II. The lattice constant values for sample mE-II are very close to those of BaFe12O19 [14]. In mE-I a decrease of the a parameter and an increase of the c parameter can be noticed. 350

500

300

400

Intensity (imp/s)

Intensity (imp/s)

400

250 200 150 100

200 100

50 0 10

300

20

30

40



a)

50

60

70

0 10

20

30

40

50

60

70



b) Fig. 1. XRD spectrums of mE-I (a) and mE-II (b).

Figure 2 presents the SEM images of the BaFe12O19 powders of sample mE-I and mE-II synthesized at 900°C (Fig.2a and 2b).

a) b) Fig. 2. SEM images mE-I (a) and mE-II (b) sinthesized at 900°C. For both powder samples it is not possible to observe fully formed particles with the hexagonal shape typical for barium hexaferrite. The particles in sample mE-I were predominantly of a platelet-like structure with an irregular shape and an average size of 180 nm. It should be emphasized here that the particles, the platelets, are very thin. The presence of particles with a size in the range 125 – 160 nm was also seen. The BaFe12O19 powder of sample mE-II exhibits a narrower grain size distribution, with the average particle size being 130 nm (Fig.2b). The particles have an irregular shape between spherical and hexagonal. The irregular shape observed in both samples is because the process of formation of the platelet shape typical for BaFe12O19 hexahedral has not been completed due to the small particle size. The critical diameter for single-domain barium hexaferrite particles is about 460 nm [15], so that the particles are single domain in both powder samples. The results of the microstructural investigations indicated that a lower concentration of Fe3+ and Ba2+ ions in the aqueous phase of the single microemulsion system leads to the formation of BaFe12O19 powders with a higher degree of uniformity of particles size.

58

Perspectives of nanoscience and nanotechnology

Further microstructural investigations were carried out on the powder samples obtained by sintering the precursor at 580°C (Fig. 3).

a)

b) Fig. 3. SEM images mE-I (a) and mE-II (b) synthesized at 580°C.

In the case of the higher concentration of metallic ions (mE-I), needle-like Fe2O3 crystallites were observed (Fig. 3a), but this was not seen for the lower concentration samples (Fig. 3b). It was assumed that it is the transition through Fe2O3 with needle-like shape which is the reason for the incomplete crystallite formation and for the non uniformity of the particles size in the samples. The hysteresis loops of the two powders at room temperature and at 4.2 K at a maximum applied field of 2.3 x 106 A/m are shown in Fig. 4a and b.

4.2 K

40

mE-I

20

-30

-20

-10

0 -20

M, emu/g

300K

M, emu/g

mE-II 60

100

mE-II

75 50

mE-I

25 0

10

20

30

5

-30

-20

-10

H, x 10 A/m

0 -25

-40

-50

-60

-75

0

10

20

30

5

H, 10 A/m

-100

a) b) Fig. 4. Hysteresis loops of samples mE-I and mE-II at 300K (a) and 4.2K (b). The magnetic parameters, namely, the remanent magnetization (Mr) and coercivity field (Hc) obtained from the hysteretic curves, are given in Table 1. T [K] mE-I 4.2 mE-I 300 mE-II 4.2

Ms [emu/g] 61 39.43 90,61

Mr [emu/g] 28.38 19.33 44.39

Mr/Ms 0.49 0.49 0.48

Hc × 105 [A/m] 3.1 3.6 3.5

mE-II 300 62 29.26 0.48 3.9 T - temperature, Ms - saturation magnetization, Mr - remanent magnetization, Mr/Ms - squareness ratio, Hc - coercivity field Table 1

Magnetic properties of barium hexaferrite powders mE-I and mE-IІ .

Solid State Phenomena Vol. 140

59

The saturation magnetization values (Ms) were obtained from the magnetization curves in high magnetic fields up to 1 x 107 A/m and are presented in Fig. 5a and b. 100

mE-II

M, emu/g

M, emu/g

70 60 50

mE-II

80 70 60

mE-I

40

90

mE-I

50 40

30

300 K

30

20

4.2 K

20 10 0

10 0 0

2

4

6 6

a)

H, 10 A/m

8

10

12

0

2

4

6

8

10

12

6

H, 10 A/m

b)

Fig. 5. Magnetization curves for mE-I and mE-II at 300K (a) and 4.2K (b). The magnetic measurements results show a saturation magnetization Ms of 39.43 emu/g and 62 emu/g at room temperature for the mE-I and mE-II powders respectively; these values are lower than the theoretical values calculated for single crystal particles of barium hexaferrite, i.e. 72 emu/g, as reported by Shirk and Buessem [16]. The low saturation magnetization can be explained by the fact that the particles are single domain. The squareness ratio (Mr/Ms) is found to be around 0.5, which is close to the value expected for randomly packed single domain particles [17]. The magnetic characteristics are worse than those of single crystals of barium hexaferrite. This is most probably the result of the presence of magnetic and structural defects on the surfaces of the particles due to the particles not having achieved the perfect hexagonal shape characteristic of barium hexaferrite. In general, the magnetic characteristics of sample mE-I are worse than those of sample mE-IІ, both at room and at very low temperature. This is due both to the lower homogeneity of particles shape and size and to the presence of a second (hematite) phase in sample mE-I. It should be pointed out that the magnetic parameters of sample mE-IІ are comparable to those of powders produced by other “wet chemistry” techniques. Conclusions A single microemulsion technique was employed to synthesize barium hexaferrite powders with a particle size below 200 nm. It was found that the size and shape of the barium hexaferrite particles produced was dependent on the concentration of metal cations in the aqueous phase of the microemulsion system, with the effect being related to the shape and size of the precursor particles used in the synthesis process. The barium hexaferrite powders synthesized at low concentration of metallic cations consists of particles with high uniformity with respect to their size (130 nm) but with irregular shape (between spherical and hexagonal). The saturation magnetization and coercivity values of the powders obtained may be attributed to the small particles size and their shape, which gave rise to the presence of structural and magnetic defects on the surfaces of the particles. The magnetic properties of the sample obtained at the lower concentration of metallic cations in the aqueous phase are better than those of the sample obtained at the higher concentration. One of the main reasons for this is the greater uniformity of the particles size of the powder synthesized at the low concentration of metallic cations. It was demonstrated that the single microemulsion method may be used to prepare powders of monodomain barium hexaferrite

60

Perspectives of nanoscience and nanotechnology

nanoparticles with a high size-homogeneity and good magnetic properties which gives prospects for possible applications. Acknowledgments Dr. Koutzarova was supported by the NATO Reintegration Grant (EAP.RIG.981472). The work was supported in part by research agreements between CGRI, Belgium and the Bulgarian Academy of Sciences, FNRS-CREDIT CHERCHEURS (CREDIT 1.5.276.07), a Joint research project between the Polish Academy of Sciences and the Bulgarian Academy of Sciences and the Bulgarian Scientific Fund under grant HT-1/01. The authors wish to thank CATµ laboratory at the University of Liege for providing the SEM micrographs. References [1] J.Wang, P.F. Chong, S.C. Ng and L.M. Gan: Mater. Lett. Vol. 30 (1997), p. 217 [2] M. Drofenik, D. Lisjak and D. Makovec: Mater. Sci. Forum Vol. 494 (2005), p. 129 [3] V. Pillai, P. Kumar, M.J. Hou, P. Ayyub and D.O. Shah: Adv. Coll. Inerf. Sci. Vol. 55 (1995), p. 241 [4] D. Vollath, D.V. Szabó, R.D. Taylor and J.O. Willis: J. Mater. Res. Vol. 12 (1997), p. 2175 [5] V. Pillai and D. Shah: J. Magn Magn. Mater. Vol. 163 (1996), p. 243 [6] L. LaConte, N. Nitin and G. Bao: Mater. Today Vol. 8 (suppl.) (2005), p. 32 [7] V. Chhabra, M. Lal, A. Maitra and P. Ayyub: J. Mater. Res. Vol. 10 (1995), p. 2689 [8] X. Liu, J. Wang, L. Gan, S. Ng and J. Ding: J. Magn Magn. Mater. Vol. 184 (1998), p. 344 [9] M. D. Shultz, M. J. Allsbrook and E.E. Carpenter: J. Appl. Phys. Vol. 101 (2007), p. 09M518 [10] T. Koutzarova, S. Kolev, Ch. Ghelev, D. Paneva and I. Nedkov: phys. stat. sol.(c) Vol. 3 (2006), p. 1302 [11] T. Koutzarova, S. Kolev, Ch. Ghelev, D. Paneva and I. Nedkov, in: Proceeding of the 7th Workshop Nanostrumaterials Application and Innovation Transfer, edited by E. Balabanova and I. Dragieva, Heron Press Ltd., Sofia, (2006) p. 42 [12] J. J. Went, G. W. Rathenau, E. W. Gorter and G. W. van Oosterhout: Philips Tech. Rev. Vol. 13 (1952), p. 194 [13] S. Trojanowski, A. Gilewski and J. Warchulska: Metrology and Measurement Systems Vol. 11 (2004), p. 159 [14] JCPDS 39-1433 [15] L. Rezlescu, E. Rezlescu, P. D. Popa and N. Rezlescu: J. Magn. Magn. Mater. Vol. 193 (1997), p. 288 [16] B. Shhirk and W. Buessem: J. Appl. Phys. Vol. 40 (1969), p. 1294 [17] E. C. Stoner and E. P. Wohlfarth: Phil. Trans. Roy. Soc. London A Vol. 240 (1948), p. 599

Solid State Phenomena Vol. 140 (2008) pp 61-68 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.61

Synthesis of Titania Nanostructures and their Application as Catalyst Supports for Hydrogenation and Oxidation Reactions Lucky Mashudu Sikhwivhilu 1, a, Suprakas Sinha Ray 1, b and Neil John Coville 2,c 1

National Centre for Nanostructured Materials, CSIR, 1-Meiring Naude Road, Brummeria, P.O. Box 395, Pretoria 0001, Republic of South Africa 2

Molecular Sciences Institute and School of Chemistry, University of the Witwatersrand, Johannesburg WITS 2050, South Africa

e-mail: a [email protected]; b [email protected]; c [email protected] Keywords: Titania nanotubes, titanate nanotubes, titania, hydrothermal synthesis

Abstract. Nanomaterials are of great importance for their versatile applications in gas sensors, solar cells and photocatalysis due to their unique optical, electrical and catalytic properties. Titania derived nanotubular and nanospherical particles with a titanate structure were synthesized using a hydrothermal procedure in the presence of very concentrated solutions of KOH and NaOH respectively. Both nanostructures were found to exhibit relatively large specific surface areas, i.e. 280 and 303 m2/g for materials treated in NaOH and KOH respectively. The morphological and structural properties were characterised by TEM, SEM, Raman spectroscopy and XRD. Introduction Titania is widely studied because of its availability and reasonable cost. It has numerous applications including gas sensors, dielectric ceramics, catalysts for thermal- or photo-induced reactions, photovoltaic solar cells and pigments [1-4]. Studies have shown that the physical and chemical properties of TiO2 can be controlled or altered by its particle size, morphology and crystallographic structure. For example, TiO2 particles at nanoscale levels (< 100 nm) show a dramatic change in specific surface area and have displayed unusual optical properties, electrical and catalytic properties. These phenomena have attracted greater research interest in the study of TiO2 nanostructures [5-7]. The synthesis of materials with specific size, microstructure and properties is particularly important in preparative chemistry and material sciences [8-10]. In this paper the synthesis of materials with small dimensions and tubular morphology are reported. One-dimensional (1-D) nanostructures such as nanotubes, nanorods, nanowires and nanofibres are particularly important because of their superior physicochemical properties such as electronic, magnetic, optical, catalytic and mechanical properties. These properties widen the range of potential applications in environmental purification, nanodevices, gas sensors and high effect solar cells [11-14]. Numerous procedures have been used to prepare nanostructured materials with unique morphologies including sol-gel, hydrolysis, anodisation and hydrothermal synthesis [15-19]. In this study hydrothermal procedure was the preferred method to prepare TiO2 derived nanotubes due to the low energy consumption, reduced air pollution, easy solution control and high reactivity of the reactants. The synthesis of TiO2 derived nanotubes using both hydrothermal procedures in the presence of KOH and NaOH is reported. Tubular structures with a diameter of 8-11 nm were obtained. The deposition of Pd and Au nanoparticles was achieved by impregnation and precipitation procedures respectively.

62

Perspectives of nanoscience and nanotechnology

Experimental Methods Preparation of TiO2 derived nanostructures. TiO2 derived nanotubes were prepared using a procedure similar to that described by Kasuga et. al. [13]. Commercial TiO2 powder, P25 Degussa, was used as the starting material. This material has a crystalline structure and is composed of 80 % anatase and 20 % rutile phases. The particle grain size ranged between 40 and 60 nm. In a typical experiment 20g of TiO2 powder was mixed with 300 ml of 20 M of either NaOH or KOH solution. The mixture was heated at 120 oC in an autoclave for 24 hours. After the reaction the product was separated by centrifugation and then washed repeatedly with deionised water until a pH of ~8-9 was attained. The product was washed only with deionised water and was never treated with an acid. The washed product was then dried in a vacuum oven at 120oC for 16 hours. Preparation of Pd/TiO2 derived nanotubes (TNT) composite. The Pd/TNT composite was prepared by a conventional wet impregnation method using palladium acetate, Pd(CH3COO)2 as the precursor. This method involved the dissolution of the precursor in acetonitrile and then adding the mixture to the TNT material. The mixture was dried in an oven and then calcined at 300 oC for 5 hours to give a Pd content of 1 wt. %. Preparation of Au/TNT composite. TNTs were suspended in deionised water and stirred vigorously. A dilute solution of HAuCl4 was slowly added to the TNT suspension with constant stirring. The resulting precipitates were then aged for 2 hours followed by the addition of NaBH4 already prepared in iced water. The resulting solid product was separated by filtration and washed with copious amounts of water. The Au/TNT composite material containing 1 wt. % of Au was then dried in an oven for 12 hours. Results and discussion The synthesis of the TiO2 derived nanostructures was achieved by hydrothermal treatment of commercial TiO2 (P25 Degussa) powder in the presence of KOH and NaOH and the samples are designated TNT, Titania nanotubes, and TNS, Titania nanospheres, respectively. TEM and SEM analysis. The TEM image of TNT shown in Fig. 1a reveals the presence of nanotubes with a narrow size distribution of diameter. The tubes have a diameter ranging from of 811 nm and a length of several hundreds of nanometers. Similar observation was made by Kasuga and co-workers [13]. The TEM image reveals that the tubes agglomerate to form bundles which are randomly distributed. All the tubes were found to be open ended. The TEM image of the TNS prepared using NaOH is shown in Fig. 1b. The particles were very small and they formed agglomerated clusters. A better illustration of the structure of the TNS material is revealed by the SEM, Fig. 1c. The SEM image also confirms that the material is made up of clusters of particles. It was noted that the particles in the TNS have a morphology similar to that of commercial TiO2 (P25 Degussa) (not shown) as both materials possess a morphology consisting of regular spheres.

Solid State Phenomena Vol. 140

a

63

b

c

Fig. 1. TEM images: (a) TNT (treated with KOH), (b) TEM image of TNS (NaOH treated sample) (c) SEM image of TNS The surface properties of the TNT, TNS and TiO2 (P25 Degussa materials are shown in Table 1). Catalyst TiO2 (P25 Degussa) TNT non-sintered TNT sintered at 400 oC TNS non-sintered TNS sintered at 400 oC

BET Surface Area (m2/g) 49.5 303 287 280 96.4

Pore size (nm) 5.9 6.1 6.7 4.4 5.5

Table 1: The physical properties of the nanocystalline materials The relatively large surface area obtained with TNT is attributed to the nanotubular structure of the material. However, TNS has a similar morphology to that of TiO2 (P25 Degussa) (not shown) but a larger surface area. The pore diameters of TNS and TiO2 are comparable. The similar particulate morphology suggests that the larger surface area of the TNS sample is probably due to both intraand inter-particle porosity. TNS has a larger specific surface area of 280 m2/g as compared to that of TiO2 of ~50 m2/g. The large surface area of the material is due to the treatment with NaOH in an autoclave which resulted in the formation of nanosized crystals. The large surface area of TNS material is a result of the small crystal sizes. However, it displayed a lack of thermal stability as shown by the sharp decrease in surface area with an increase in the sintering temperature which decreased from 280 to 96 m2/g,

64

Perspectives of nanoscience and nanotechnology

as the temperature was systematically increased to 400 oC. This change in surface area is indicative of crystal growth occurring at the higher sintering temperatures. It could be observed that the grain size increased with the increase in sintering temperature mainly because the crystals coalesce. The TNT sample has a specific surface area (303 m2/g) larger than that of TNS. The surface area of TNT remained at a high level after sintering at 400 oC, suffering on a slight decrease to 287 m2/g suggesting that the titania nanotubes are thermally stable.

Counts (a.u.)

XRD analysis. The XRD patterns of TNT and TNS are shown in Figs. 2 and 3 respectively. The XRD pattern of the precursor nano-sized powder (P25 Degussa) (not shown) revealed the presence of the crystalline rutile and anatase phases. 80 60 40 20 0 80 60 40 20 0 60

o

400 C

o

350 C

o

300 C

40 20 0

non-sintered

40 20 0 0

10

20

30

40

50

60

70

80

2 Theta (degrees)

Fig. 2. XRD patterns of TNS at cold pressed at room temperature and after sintering at various temperatures The XRD pattern of the non-sintered material shows that the material contains neither the anatase nor the rutile phase. The presence of ill-defined peaks is characteristic of an amorphous material. After sintering at elevated temperature broad anatase peaks appeared. As the sintering temperature was increased the intensity of the anatase peaks became stronger and well resolved, indicating that particles of larger size are being formed. At the high temperature of 400 oC only anatase peaks appeared, clearly with relatively high intensity, showing that the amorphous to anatase transformation was almost complete. There have been a number of reports in literature suggesting that the synthesis of nanosized TiO2 results in the formation of anatase and/or brookite, which, on coarsening, undergoes a transformation to rutile when a certain particle size has been reached [20]. It was observed that the anatase was composed of very fine white particles. The particles became harder during sintering due to particle growth.

Solid State Phenomena Vol. 140

65

The XRD pattern of TNT is shown in Fig. 3. 650 600 550

Counts (a.u.)

500 450 400 350 300 250 200 0

20

40

60

80

100

2 Theta (degree)

Fig. 3. XRD pattern of TNT sample The figure shows that the diffraction peaks for the TNT sample are much sharper than those of the non-sintered TNS but were broader than those obtained from the P25 Degussa TiO2 crystals (not shown). Exhaustive literature search has shown that the profile cannot be attributed to any of the known crystal structures [21]. Structural determination of the TNT, based on its profile only, is not possible because there are too few peaks due to the nanometer size of the tubes and the bending of some atom planes in the tubes [22]. It was observed that sintering the TNT material at 400 oC did not yield any notable crystallographic changes. This clearly shows that TNT material is more thermally stable than TNS and this is consistent with the BET results. This difference in thermal stability is attributed to the different metal ions (K+ and Na+) present in the material, the morphologies and the particle dimensions. Raman spectroscopy. The structural properties were investigated by Raman spectroscopy. The Raman spectra of TNT and TNS samples are shown in Figs. 4(a) and (b) respectively.

90

(a)

1400

TNT

(b)

TNS

80

1200

Intensity (a.u.)

Intensity (a.u.)

70 60 50 40 30

1000 800 600 400

20 0

200

400

600

800 -1

Wavenumber (cm )

1000

0

100

200

300

400

500

600

700

Wavenumber (a.u.)

Fig. 4. Raman spectra of (a) TNT and (b) TNS materials The TNT spectrum revealed the presence of broad and sharp vibration peaks. This effect could be due to the presence of tube bundles and isolated tubes in the material. The peaks at 188, 271, 441

66

Perspectives of nanoscience and nanotechnology

and 652 cm-1 are all attributed to a titanate structure. The 188 and 271 cm-1 peaks are assigned to a pure framework Ti-O-Ti vibration whereas the 441 and 652 cm-1 are assigned to Ti-O-K vibration [23,24]. The nanotubes are made up from potassium titanate structure, KTiO2(OH) [25,26]. The spectrum of TNS shows the presence of broad peaks that are not well resolved rendering it impossible to assign them to known crystal structures. The broadening of the peaks is due to the small particle size. Hydrothermal treatment of TiO2 with a concentrated solution of NaOH is known to yield nanotubular sodium titanate materials [27,28]. The preparation technique is based on the use of an aqueous solution of strength 10 M. In this work it was found that, by using a 20 M solution of NaOH, non-tubular nanoparticles can be synthesized. However, nanotubular structures can be obtained by using concentrated KOH solution (20 M) under similar conditions. This clearly shows that the concentration range within which nanotubes can be formed using NaOH is smaller than that of KOH. This implies that the morphology of the product is influenced by the nature, strength and concentration of the alkali used. The difference in reactivity of KOH and NaOH towards the formation of nanostructures could be related to the size of the alkali metal ion (i.e. K+ and Na+). Further studies to improve the understand this behaviour are already underway. Precious metals dispersed in the TNT material. Palladium and gold nanoparticles were dispersed on titania derived nanotubes (TNT) using wet impregnation and deposition precipitation procedures respectively. Pd loading of 5 wt% was used to prepare Au/TNT nanocomposite material in Fig. 5a whereas Au loading of 1 wt% was used (Fig. 5b).

a

b b

Pd Au

Fig. 5. HRTEM images of (a) Pd/TNT and (b) Au/TNT composites The average diameter of the Pd particles of 2.3 nm was calculated from the TEM images. Similarly the average diameter of the Au particle diameters was calculated as 4.7 nm. The size and shape of the nanotubes (TNT) were not affected by the addition of either Pd or Au. Sintering of the TNT material both in the presence and absence of either Pd or Au did not affect the morphology or dimensions of the tubular structures. The two nanocomposites materials (Pd/TNT and Au/TNT) hold good prospect for use as a catalyst support for hydrogenation and oxidation processes [29].

Solid State Phenomena Vol. 140

67

Conclusions 1. The type of nanostructure formed in the hydrothermal treatment of TiO2 with a base is dependent on the nature and concentration of the alkali. Treatment with 20 M of NaOH leads to the formation of regular spherical TiO2 nanoparticles whereas treatment with 20 M of KOH solution leads to the formation of nanotubular structures. 2. The type of base used and the nanostructure formed determines the thermal stability of the material. The nanospherical material synthesized from NaOH (TNS) was found to be thermally unstable whereas the material synthesized from KOH (TNT) was found to be thermally stable. 3. The synthesis of nanosized TiO2 (TNS) resulted in the formation of an amorphous material, due to the small particle size, which then transformed to anatase on sintering. 4. Treatment of TiO2 with a highly concentrated solution of NaOH (20 M) in an autoclave does not yield nanotubular product but nanoparticles with a large specific surface area. However, treatment of TiO2 under similar conditions yields a nanotubular product. 5.

Pd and Au nanoparticles can be easily dispersed on nanotubular material.

References 1. M.R. Hoffmann, S.T. Martin, W. Choi, D.W. Bahnemann, Chem. Rev. 95 (1995) 69. 2. M. Ferroni, V. Guidi, G. Martinelli, G. Faglia, P. Nelli, G. Sberveglieri, Nanostruct. Mater. 7 (1996) 709. 3. H.Y. Ha, S.W. Nam,T.H. Lim,I.-H. Oh,S.-A. Hong, J. Membr. Sci. 111 (1996) 81. 4. B.E. Handy, I. Gorzkowska, J. Nickl, A. Baiker, M. Schraml-Marth, A. Wokaun, Ber. BunsenGes. Phys. Chem. Chem. 96 (1992) 1832. 5. L.J. Tuller, J. Electroceram. 1 (1997) 211. 6. J.Y. Ying, T.J. Sun, J. Electroceram. 1 (1997) 219. 7. Y.-M. Chiang, J. Electroceram. 1 (1997) 205. 8. P.K. Dutta, M. Jakupca, K.S.N. Reddy, L. Salvati, Nature 374 (1995) 44. 9. M.T. Reetz, M. Maase, Adv. Mater. 11 (1999) 773. 10. X.-G. Peng, L. Manna, W.-D. Yang, J. Wickham, E. Scher, A. Kadavanich, A.P. Allvisatos, Nature 404 (2000) 59. 11. T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihra, Adv. Mater. 11 (1999) 1307. 12. Q. Chen, W.Z. Zhou, G.H. Du, L.M. Peng, Adva. Mater. 14 (2002) 1208. 13. T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Langmuir 14 (1998) 3160. 14. G.H. Du, Q. Chen, R.C. Che, Z.Y. Yuan, L.P. Peng, Appl. Phys. Lett. 79 (2001) 3702. 15. A. Scolan, C. Sanchez, Chem. Mater. 10 (1998) 3217. 16. L.H. Edelson, A.M. Glaeser, J. Am. Ceram. Soc. 71 (1988) 225. 17. C.-C. Wang, J.Y. Ying, Chem. Mater. 11 (1999) 3113. 18. H.-M. Cheng, J.-M. Ma, Z.-G. Zhao, L.-M. Qi, Chem. Mater. 7 (1995) 663. 19. I.C. Flores, J.N. de Freitas, C. Longo, M.-A. De Paoli, H. Winnischofer, A.F. Nogueira, Journal of Photochemistry and Photobiology A: Chemistry 189 (2007) 153. 20. Y. Hwu, Y.D. Yao, N.F. Cheng, C.Y. Tung, H.M. Lin, Nanostructured Materials 9 (1997) 355. 21. Q. Chen, G. H. Du, S. Zhang and L. M. Peng, Acta Cryst. B: 58 (2002) 587. 22. V. Idakiev, Z. Y. Yuan, T. Tabakova and B. L. Su, Appl. Cat. A: Gen. 281 (2005) 149. 23. X-Y. Liu and N.J. Coville, S. Afr. J. Chem. 58 (2005) 110.

68

Perspectives of nanoscience and nanotechnology

24. S. Uchida, R. Chiba, M. Tomiha, N. Masaki and M. Shirai, Studies in Surface Science and Catalysis, 146 (2003) 791. 25. N. Masaki, S. Uchida, H. Yamane, and T. Sato, Chem. Mater. 14 (2002) 419. 26. X. Sun, X. Chen, and Y. Li, Ino Q. Chen, W. Zhou, G. Du, and L.-M. Peng, Adv. Mater. 14 (2002) 1208. 27. H. Peng, G. Li, and Z. Zhang, Mater. Lett. 59 (2005) 1142. 28. L. Qian, Z.-L. Du, S.-Y. Yang, Z.-S. Jin, Journal of Molecular Structure 749 (2005) 103. 29. L.M. Sikhwivhilu, N.J. Colville, T.Niho, M.S. Scurrell, Catal. Lett, DOI 10.1007/s10562-0089439-z

Solid State Phenomena Vol. 140 (2008) pp 69-74 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.69

Magnetic Anisotropy of Co Films Annealed by Laser Pulses J. Kisielewski1, a, K. Postava1, 2, b, I. Sveklo1, c, A. Nedzved1, d, P. Trzciński1 e, A. Maziewski1, f , B. Szymański3, g, M. Urbaniak 3, h and F. Stobiecki3, i 1

Laboratory of Magnetism, Institute of Experimental Physics, University of Bialystok, Lipowa 41, 15-424 Bialystok, Poland. 2

3

Department of Physics, Technical University of Ostrava, 17. listopadu 15, 708 33 Ostrava-Poruba, Czech Republic.

Institute of Molecular Physics, Polish Academy of Sciences, Smoluchowskiego 17, 60-179 Poznań, Poland

e-mail: a [email protected]; b [email protected]; c [email protected]; d [email protected]; e [email protected]; f [email protected]; g h i [email protected]; [email protected]; [email protected]. Keywords: Ultrathin film, Magnetic anisotropy, Perpendicular anisotropy, Laser annealing, MOKE, AFM, MFM

Abstract. The magnetic properties of an ultrathin cobalt film were modified by a focused femtosecond pulsed laser beam. The Co wedge, with a thickness ranging from 0 to 2 nm, sandwiched by Au films was prepared using ultra-high vacuum magnetron sputtering on a mica substrate. The modifications of the laser induced magnetic anisotropy were investigated using magneto-optic Kerr microscopy and MFM/AFM techniques. The laser induces: (i) local reorientation of magnetization from an in-plane to a perpendicular state and (ii) an increase of the coercivity field. A corresponding increase of the perpendicular magnetic anisotropy is discussed considering an improvement of the Co/Au interfaces.

Introduction Ferromagnetic ultrathin films with perpendicular magnetic anisotropy have been intensively studied in recent years because of their unique properties and applications for high density data recording [1]. For higher thicknesses the magnetic dipolar anisotropy forces magnetization into the plane of the film. However, when the film thickness is decreased, a relative increase of the surface contribution leads to perpendicular magnetic anisotropy. Thickness controlled spin reorientation transition (SRT) from an in-plane to a perpendicular preferential magnetization direction can be realized [2, 3]. A wide thickness range with perpendicular magnetization is obtained for films with high quality interfaces. Tuning the SRT is possible by changes in the overlayer [2], the underlayer, and interface roughness [4], post-deposition annealing [5, 6], the ion [7], and electron bombardment [8]. The purpose of this paper is the study of the influence of intense femtosecond laser pulses on the magnetic properties of a Co ultrathin film. The ultrafast laser is a unique tool for irreversible modification of nanostructures (e.g. melting, disintegrating) [9, 10] and for reversible control of magnetism (e.g. by inverse Faraday effect in magnetically ordered materials) [11]. Moreover, a femtosecond pulse laser can be applied to produce periodic patterning by use of interference phenomena [12, 13]. In comparison with usual thermal annealing, laser annealing is a more precise technique allowing higher spatial, energy, and time resolution; additionally, due to the short action on the top layer it significantly reduces undesirable diffusion from the sample bulk. In this paper the irreversible changes of ultrathin Co film structures and magnetic properties are reported

70

Perspectives of nanoscience and nanotechnology

Experimental details The following structure was deposited by UHV magnetron sputtering onto a mica substrate with a Au(300 nm) buffer: (i) Au(1.5 nm); (ii) a Co wedge with thickness d changing linearly from 0 to 2 nm on a 15 mm long sample; (iii) Au(3 nm). The sputtering rates were 0.06, and 0.045 nm/s, for Au and Co, respectively. X-ray diffraction investigations performed on similar samples [14] reveal that the films are (111) textured. The Ti:Sapphire femtosecond pulse laser (Chameleon Ultra, Coherent) was used for local modification of the magnetic properties of the Co. The pulse width is 140 fs and the TEM00 mode output beam was linearly polarized. The wavelength was adjusted to 700nm and in the corresponding time the average power reached 1.9 W with a pulse repetition rate of 90 MHz. The laser beam was focused on the sample surface using a plan-convex lens and the light spot diameter was estimated at 2w0 ~ 10 µm. The sample was irradiated using a computer controlled motorized xy-positioning of the sample and the automatic inner shutter of the laser to adjust the exposure time ∆L. The magnetization distribution was studied at room temperature using the polar Kerr effect using an optical microscope equipped with xenon lamp illumination and a high sensitivity cooled CCD camera connected to a computer by a frame grabber [2, 3]. The normalized differential image was calculated as [I(H) – I(-Hmax)]/[I(H) + I(-Hmax)]], with all images acquired in a zero field: the I(H) image was obtained after the positive perpendicular magnetic pulse H (usually a series of 2 s positive H pulses were applied to the initially saturated sample by -Hmax), the reference I(-Hmax) image registered after a -578 Oe field pulse. MFM measurements were carried out with a Ntegra Prima scanning probe microscopes (NT-MDT) using the “Tapping mode” for topography imaging and the “Lift Mode” (height 50 nm) for magnetic imaging. The MFM setups have the possibility for application for the permanent external magnetic field (either in-plane or outof-plane direction). Low magnetic moment MESP-LM probes (Veeco) and home-deposited low coercivity 50 nm Co probes were used to visualize the sample’s magnetic structure.

Results and discussion Figure 1 shows the remanent image of the laser irradiated cobalt wedge obtained as the difference between two images after the pulses of positive and negative maximal field. The region without the laser treatment (the lower region marked as B) will be considered first. For a simplified description of Co ordering, three magnetization states could be distinguished in the discussed thickness range: superparamagnetic, out-of-plane, and in-plane [2, 3]. The out-of-plane state with a square hysteresis loop exists between the thicknesses denoted as dL and dH (the white area on the remanent image). In this case these characteristic thicknesses for the as deposited Co region are: dL = 0.41 nm and dH = 0.93 nm. The image intensity is proportional to the Kerr rotation angle ϕmax. On increasing d, the image intensity almost linearly increases in this region [2], which is related to the linear dependence of ϕmax on thickness in the ultrathin-film range. The black area on the remanent image represents: (i) the superparamagnetic state below dL or (ii) the in-plane magnetization state above dH. The transition between white (left) and black (right) area represents the SRT between the perpendicular and in-plane magnetization state. The laser modified regions are shown in the upper part of the remanent image, Fig. 1b. A regular array of irradiated spots with the exposure time ranging from 0.2 s to 5 s and two continuous irradiated lines are clearly visible in the remanent image. The lower and upper lines were produced by continuous motion of the sample at a speed of 30 and 150 µm/s, respectively. One can distinguish two types of laser induced magnetization changes – above and below the SRT of the non-irradiated sample. Above the SRT (d > dH) there is clear evidence of the creation of the perpendicular magnetization state (the white areas).

Solid State Phenomena Vol. 140

71

Figure 1. a) schematic side view of the sample; b) remanence polar Kerr image of array of irradiated spots and two lines; c) profiles of relative remanence magneto-optic Kerr effect on Co wedge, along irradiated line (triangles, marked A in remanence image) and as-deposited line (squares, marked B). Figure 2 illustrates evolution of the magnetization reversal process in two selected spots, indicated with arrows C and D in Fig. 1b (in the third row with the exposition 1 s). The process starts with a reversed domain nucleation close to a spot edge – the SRT region. On increasing the applied field pulse H the “white” domain area increases by propagation of the domain wall from the spot edge toward the centre. Finally magnetization is reversed in the whole spot except the central region. The lowest panels in Fig. 2 show the remanent images. The magnetization reversal process is similar to that described for the as-deposited wedge type of ultrathin Co [3]. The “black” area inside any spot has no remanence and seems to be related to the presence of a non-ferromagnetic phase. Even with the field pulse as high as 4 kOe it does not increase magnetic contrast of the central region. The “white” spot area represents the laser-induced change from the in-plane to the perpendicular magnetic state. It decreased with increase of d and decrease of the exposure time ∆L.

72

Perspectives of nanoscience and nanotechnology

Figure 3 shows magnetization processes in the two regions below the SRT, marked with rectangles F and E in Fig. 1b. For the non-irradiated sample the coercivity field reaches a maximum of about 50 Oe at d = 0.8 nm. Cobalt thickness regions were selected: (i) in Fig. 3a - below the coercivity field peak, where the direction of magnetization reversal process (“white” area growth) goes from the low to the high cobalt thickness - from left into right in Fig. 3a; (ii) in Fig. 3b - above the coercivity peak close to SRT, the direction of magnetization reversal goes from the high to the low cobalt thickness region- from right into left in the Fig. 3b. Details of magnetization reversal in the as-deposited Co wedge are discussed in reference [3]. The laser annealed areas are reversed in the higher magnetic field. Moreover, the domain sizes in the laser irradiated area are larger than those in the reference as-deposited regions. The laser irradiation induces an increase of coercivity, which is usually related to the higher perpendicular magnetic anisotropy.

Figure 2. Magnetization reversal of irradiated spots for Co thickness of 1.06 nm (a) and 1.34 nm (b) above SRT (pointed by arrows D and C, respectively, in Fig. 1b, with the 1 s exposition).

Figure 3. Magnetization reversal of regions for Co thickness below SRT (marked as rectangles F and E in Fig. 1b). Three lines of spots correspond to the exposition 0.2 s, 0.5 s, and 1 s.

Figure 1c shows the profiles of remanent image, taken along irradiated (“white” area) and reference non-irradiated lines (lines A and B, respectively), plotted as the dependency on Co-thickness. The magnetization states in the irradiated and non-irradiated states are similar below dH – as both dependencies overlap. Above dL the remanence MO effect of the irradiated area monotonically increases with Co thickness. The thickness dependence of remanence shows that the laser irradiation does not give rise to a change of the polar magneto-optic Kerr effect. Consequently, it

Solid State Phenomena Vol. 140

73

was deduced that the laser annealing in the region discussed does not significantly modify the film’s structure and the magnetic moment density of the Co film. From the above experimental results it was deduced that the laser irradiation related effects are responsible for the increase of the perpendicular magnetic anisotropy observed in both thickness ranges. The increase of the magnetic anisotropy seems to be related to the improvements of Co/Au interfaces after the laser annealing. A similar effect was observed [15] for Co/Au ion sputtered multilayers where annealing at only 300ºC, however for a significantly longer period, led to an increase in the magnetic anisotropy. A strong temperature increase is expected during the laser pulse and drastic laser induced changes of the morphology of the film’s nanostructure were reported for different materials [12]. Melting of the metal was expected in the laser’s focus spot. The results of AFM measurements support this hypothesis - the large and regular crystallites are observed inside the focusing spot which indicates, that during the exposure, film melting and recrystallization occurred (Fig. 4).

Figure 4. AFM Image of laser irradiated area (1.9 W for 1 second) with black spot inside. Triangular shaped crystal and tall 300nm high mica swelling are visible. If the laser power is significantly reduced to 0.3 W with a slightly prolonged exposure time of 20 s the effect of laser annealing, but without visible surface changes can be observed (Fig. 5). It means that using a focused laser beam as a tool it is possible to make longitudinal pattering of the magnetic properties of a cobalt film without any visible changes to the surface topography. a)

b)

c)

Figure 5. Irradiated spot (0.3 W during 20 s) for Co thickness of 1 nm imaged with: a) magneto-optic remanence Kerr effect; b) AFM; c) MFM. MFM image was obtained in zero external field after 2 s pulse of perpendicular magnetic field 950 Oe. No visible changes in sample surface were detected by AFM.

74

Perspectives of nanoscience and nanotechnology

Conclusion In conclusion, annealing with femtosecond laser pulses induces an increase in the perpendicular magnetic anisotropy of an ultrathin Co film sandwiched between Au layers. The laser beam induced: (i) inside the focused spot – a nonferromagnetic region with crystallites on the film surface; (ii) outside the focusing spot – a reorientation from in-plane magnetization state into the out-ofplane state. The magnetic anisotropy increase was explained by considering that improvements of the quality of the Co/Au interfaces contributed to the surface magnetic anisotropy. The effects discussed could be used for magnetic pattering which are important for technical applications.

Acknowledgements The work was supported by Marie Curie Fellowships for “Transfer of Knowledge” (NANOMAGLAB 2004-003177), by the Polish State Committee for Scientific Research as the research project 3 T08A 03127 and by the KAN 400100653 project. The authors are grateful to Coherent GMBH for access to the Chameleon Ultra laser.

References [1] B. D. Terris and T. Thomson, J. Phys. D: Appl. Phys. 38, R199 (2005). [2] M. Kisielewski, A. Maziewski, M. Tekielak, A. Wawro, and L. T. Baczewski, Phys. Rev. Lett. 89, 087203 (2002) and references therein. [3] M. Kisielewski, Z. Kurant, M. Tekielak, W. Dobrogowski, A. Maziewski, A. Wawro, and L. T. Baczewski, phys. stat. sol. (a) 196, 129 (2003). [4] J.-R. Jeong, J. A. C. Bland, J.-W. Lee, Y.-S. Park, and S.-C. Shin, Appl. Phys. Lett. 90, 022509 (2007). [5] A. Wawro, Z. Kurant, L. T. Baczewski, P. Pankowski, J. Pelka, A. Maneikis, A. Bojko, V. Zablotskii, and A. Maziewski, phys. stat. sol. (c) 3, 77 (2006). [6] M. Urbaniak, F. Stobiecki, and B. Szymanski, J. Alloys Compd. p. (in press) (2007) and references therein. [7] J. Ferre, T. Devolder, H. Bernas, J. P. Jamet, V. Repain, M. Bauer, N. Vernier, and C. Chappert, J. Phys. D: Appl. Phys. 36, 3103 (2003). [8] R. Allenspach, A. Bischof, U. Durig, and P. Grutter, Appl. Phys. Lett. 73, 3598 (1998). [9] D. S. Ivanov and L. V. Zhigilei, Phys. Rev. B 68, 064114 (2003). [10] P. Lorazo, L. J. Lewis, and M. Meunier, Phys. Rev. B 73, 134108 (2006). [11] A. V. Kimel, A. Kirilyuk, F. Hansteen, R. V. Pisarev, and T. Rasing, J. Phys.: Condens. Matter 19, 043201 (2007). [12] C. Favazza, J. Trice, H. Krishna, R. Kalyanaraman, and R. Sureshkumar, Appl. Phys. Lett. 88, 153118 (2006). [13] J. B. Kim, G. J. Lee, Y. P. Lee, J. Y. Rhee, K. W. Kim, and C. S. Yoon, Appl. Phys. Lett. 89, 151111 (2006). [14] M. Urbaniak, F. Stobiecki, B. Szymanski, A. Ehresmann, A. Maziewski, and M. Tekielak, J. Appl. Phys. 101, 013905 (2007). [15] F. J. A. den Broeder, D. Kuiper, A. P. van de Mosselaer, and W. Hoving, Phys. Rev. Lett. 60, 2769 (1988).

CARBON NANOSTRUCTURES

Solid State Phenomena Vol. 140 (2008) pp 77-80 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.77

First Principle Investigation of Structural Properties of Potassium Doped Fullerene Clusters – Kn(C60)2. M. Sokół a and Z. Gburski b Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007, Katowice, Poland e-mail: a [email protected]; b [email protected] Keywords: Potassium-fullerene nanosystem, first principle simulation, DFT calculation, alkali doped fullerene, Kn(C60)2 cluster

Abstract. First principle simulations for the nanosystems Kn(C60)2, (n = 1, 2) composed of two fullerene (C60) molecules and one or two potassium (K) atoms have been undertaken. A very effective delocalization of the 4s1 valence electron of potassium was observed, the potassium atom in practice becomes an ion. The adsorption binding energy of potassium atom(s) is Ea = 1.923 ± 0.04 eV, - 3.819 ± 0.04 eV for K(C60)2 and K2(C60)2, respectively. The reported large values of adsorption energy should cause a significant change in electronic properties of alkali doped fullerene clusters. Introduction In recent years, fullerenes have been intensively studied because of their unusual physical and chemical properties [1-3]. The discoveries of conducting and superconducting doped fullerides [3, 4] have initiated a strong interest in experimental studding fullerene covered with alkali metals, i.e. AnC60 systems, where A is an alkali metal atom [5-10]. The relative high dielectric permittivity ε = 4.4 of solid C60 [11] determines the significant difference between the binding energies of (C60)2 and (C60)2+[12], the relatively high binding energy of C60+ in (C60)n+ clusters, and the large stabilization energies obtained for exohedral AC60 complexes [13]. From these results one can expect large values for the binding energies of ions in fullerene clusters, such as An(C60)2. Because of the difficulty involved in the preparation and collecting measurements of nanocomposites, the molecular dynamics simulations (classical or quantum called ab initio) can be exploited as valuable tools to provide an atomistically detailed description of complex molecular processes occurring in these extremely small systems. The density-functional theory (DFT) [14, 15] in its Generalized Gradient Approximation (GGA) [16-18] applying the SIESTA programming code [19] was used for the simulations. The core electrons were replaced by normconserving pseudopotentials in their fully separable form. The nonlinear exchange-correlation correction [20-22] was included for the potassium atom to improve the description of the core-valence interactions. Results It is known that the attractive nature of the fullerene-fullerene interaction potential is very strong as a result of the high atomic density on the surface of the C60 [23-26]. It was shown that the addition of alkali atoms into fullerene based bulk samples significantly changes their physical properties (conductivity, polarizability, etc.) [3, 4]. The studies of finite systems (clusters, ultrathin-films) composed of C60 are not yet so advanced. However, in recent years increasing activity in this field has been observed because the finite sized extremely small fullerene based systems are expected to exhibit new physical and chemical properties. Following this path the focus has been on the extremely small, potassium doped nanosystems K(C60)2 and K2(C60)2.

78

Perspectives of nanoscience and nanotechnology

The study commenced by searching the equilibrium configuration of (C60)2 system. For that matter, firstly the equilibrium configuration of the pair of fullerenes only (without potassium atoms) was found. Next in order to obtain the binding curve for K atom adsorbed by (C60)2 cluster, the adsorption energy varying the positions of potassium atom by 0.1 Å in Z axis, up to 8 Å was calculated (see Fig. 1).

Fig. 1. Snapshot of the equilibrium configuration (minimum energy) of a simulated K2(C60)2 nanosystem. In the case of K(C60)2 the top potassium atom is removed The adsorption energy Ea is defined as the total energy gained by the (C60)2 system due to the addition of the potassium, Ea = Et[Kx(C60)2] – Et[(C60)2] – x*Et[K] (x = 1, 2). The binding curve for K atom adsorbed by (C60)2 cluster is shown in Fig. 2.

Fig. 2. The adsorption energy as a function of the distance (in Z axis direction, see Fig. 1) between the potassium atom and the centre of mass of both fullerene molecules. For the potassium atom placed at a distance ~ 4.5 Å above the middle of the axis connecting the centres of mass of C60 molecules (at the minimum of Ea), Ea = - 1.923 ± 0.04 eV and 3.819 ± 0.04 eV for K(C60)2 and K2(C60)2, respectively was obtained. The authors have also calculated the charge density for a pure fullerene cluster (C60)2 and for both doped systems Kn(C60)2.. It can be seen that in K(C60)2 and in K2(C60)2, the addition of potassium atom causes only a very slight modification of the charge density distribution for the pure fullerene cluster IFigs 3a, 3b)

Solid State Phenomena Vol. 140

79

Fig. 3. Contour plots of charge density of: a) K(C60)2 and b) K2(C60)2 system. Plots c) and d) show the difference between pure fullerene cluster (C60)2 and fullerene molecules doped by one or two potassium atoms, respectively. The locations of potassium atoms are shown in Fig. 1. To better understand charge density modifications the difference between a pure fullerene cluster (C60)2 and the system doped by one or two potassium atoms (see Fig. 3c, d) were also plotted. Analyzing Fig. 3c, d, it can be observed that the only one unbound (4s1) electron from the potassium atom is, in both cases, shared by the two fullerene molecules. To obtain quantitative information the Mulliken population analysis was made which confirmed an even dispersion of potassium 4s1 valence electron over both fullerenes. Strictly speaking, it has been found that only the charge q = - 0.131 e remains within the potassium atom in K(C60)2 system. In the case of K2(C60)2 each potassium atom holds the charge q = - 0.136 e from the 4s1 valence electron. Therefore, by losing one electron the potassium atoms in practice become ions (K+). The calculated density of states (DOS) diagrams is presented as Fig. 4.

Fig. 4. The density of states for: a) pure fulerene cluster (C60)2 and b), c) for fullerene cluster doped by one or two potassium atoms, respectively. Dotted lines denote the Fermi level

80

Perspectives of nanoscience and nanotechnology

The potassium doping of a pure fullerene cluster (C60)2 shifts its DOS towards lower energies, both for K(C60)2 and K2(C60)2. The addition of K atom induces several discrete peaks on the main DOS distribution. The calculated DOS for the doped samples have similar shape and peak energies, but different peak heights. Conclusions In conclusion, the calculations show that the single 4s1 electron of potassium atom is very highly delocalized (only the charge q ∼ - 0.13 e remains within the potassium atom) in Kn(C60)2 systems. Therefore, the potassium atom in practice becomes an ion. The adsorption energy of potassium atom(s) in both studied systems K(C60)2, K2(C60)2 is quite large – the consequence of high ionization of the alkali (K) atom. The adsorption of alkali metals should induce a significant change in the electronic properties of a fullerene cluster due to the charge transfer and charge fluctuation. Acknowledgments This research was supported by the computer time grant No. G30-12 from the Interdisciplinary Center for Mathematical and Computational Modeling, at Warsaw University. References [1] M. S. Dresselhaus, J. Dresselhaus, P. C. Eklund, Science of Fullerenes and Carbon Nanotubes Academic, New York (1995). [2] H. Kuzmany, B. Burger, J. Kurti, Optical and Electronic Properties of Fullerenes and FullereneBased Materials, edited by J. Shinar, Z. V. Vardeny, Z. H. Kafafi Dekker, New York (2000). [3] R. C. Haddon et al., Nature (London) 350 (1991) p. 320 [4] D.W. Murphy et al., J. Phys. Chem. Solids 53 (1992) p. 1321 [5] S. Wehrli, E. Koch, M. Sigrist, Phys. Rev. B 68 (2003) p. 115412 [6] B. Verberck, V. N. Popov, A. V. Nikolaev, D. Lamoen, J. Chem. Phys. 121 (2004) p. 321 [7] A. V. Nikolaev, K. H. Michel, J. Chem. Phys. 122 (2005) p. 64310 [8] A.Dawid, Z. Gburski, Phys. Rev. A 56 (1997) p. 3294 [9] A.Dawid, Z. Gburski, J. Mol. Struct. 410 (1997) p. 507 [10] A.Dawid, Z. Gburski, J. Mol. Struct. 704 (2004) p. 287 [11] A. F. Hebard, R. C. Haddon, R. M. Fleming, A. R. Kortan, Appl. Phys. Lett. 59 (1991) p. 2109 [12] W. Branz, N. Malinowski, A. Enders, T. P. Martin, Phys. Rev. B 66 (2002) p. 094107 [13] A. Ruiz, J. Hernández-Rojas, J. Bretón, J. M. Gomez Llorente, J. Chem. Phys. 109 (1998) p. 3573 [14] R. Car, M. Parrinello, Phys. Rev. Lett. 55 (1985) p. 2471 [15] P. Hohenbergand, W. Kohn, Phys. Rev. 136 (1964) p. 864 [16] W. Kohn, L.J. Sham, Phys. Rev. A 140 (1965) p.1133 [17] A.D. Becke, Phys. Rev. A 38 (1988) p. 3098 [18] C. Lee, W. Yang. R.G. Parr, Phys. Rev. B 37 (1988) p. 785 [19] P. Ordejón, E. Artacho, J. M. Soler, Phys. Rev. B (Rapid Comm.) 53 (1996) R10441 [20] D. Sanchez-Portal, E. Artacho, P. Ordejon, J.M. Soler, Int. J. Quantum Chem. 65 (1997) p. 453 [21] L. Kleinman, D.M. Bylander, Phys. Rev. Lett. 48, (1982) p. 1425. [22] S. G. Louie, S. Froyen, M.L. Cohen, Phys. Rev. B 26 (1982) p. 1738 [23] L. Liwei, D. Bedrov, G. D. Smith, Phys. Rev. E. 71 (2005) p. 011502 [24] A. Piątek, A. Dawid, Z. Gburski, J. Molec. Struc. 792 (2006), p.82 [25] A. Dawid, Z. Dendzik, Z. Gburski, J. Molec. Struc. 704 (1) (2004) p.173. [26] A. Piątek, A. Dawid and Z. Gburski, J. Phys. Condens. Matter, 18 (2006), p.8471

Solid State Phenomena Vol. 140 (2008) pp 81-88 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.81

MD Study of the Endohedral Potassium Ion Fullerene Cluster (K+@C60)7 A. Piątek 1,a, A. Dawid 1, b, K. Górny 1, c, R. Nowak 2, d and Z. Gburski1, e 1

Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007, Katowice, Poland Nordic Hysitron Laboratory, Helsinki University of Technology, 02015 HUT, Finland

2

e-mail:

a d

[email protected]; b [email protected]; c [email protected]; [email protected]; e [email protected]

Keywords: endohedral fullerene, K+@C60 , liquid fullerene, phase transition, MD simulation

Abstract. The nanosystem composed of only as few as seven endohedral fullerene K+@C60 molecules was simulated using the MD method. The interaction was taken to be the full site-site pairwise additive Lennard-Jones (LJ) potential, which generates both translational and anisotropic rotational motions of each endohedral fullerene. The atomically detailed MD simulations allow the dynamics of the motion of K+@C60 molecule inside the cluster to be analysed. The radial distribution function, the mean square displacement, the translational velocity correlation functions and the Lindemann index of endohedral fullerene have been calculated for several energies of the nanosystem. The solid/liquid phase transition and the existence of the liquid phase in the endohedral potassium ion fullerene cluster was found. Introduction In the era of nanotechnology, the quest to find the liquid phase of fullerite, which happened to be rather disappointing in case of bulk C60 samples [1 - 5], switched to the nanoscale. The activity regarding this issue in nanosystems has been enhanced by research on the possible application of fullerene compounds as electro-fluid shuttle memory elements, ultra-lubricators in molecular bearings and nanogears, etc [6-9]. In addition, it is known that molecular clusters display physical properties that often differ strikingly from those of bulk system [10–14]. In fact, Gallego et al. [15] showed by MD simulation that a small cluster, (C60)7, can achieve a liquid like state. They used Girifalco’s potential [16], which is obtained by considering the C60 molecule as a perfect sphere with a surface consisting of a uniform density of carbon atoms. The existence of a liquid phase of the (C60)7 cluster has been recently confirmed in more refined, atomically detailed MD simulations by applying the full 60-site pairwise additive Lennard-Jones (LJ) potential which generates more a realistic translatory-rotary motions of the molecules [17]. The question arose, however, of whether the liquid phase could appear also in endohedral fullerene clusters. This is the subject of the computer experiment undertaken by the MD method. Particularly, the (K+@C60)7 nanosystem was studied, taking into account the possible applications of endohedral fullerene nanocompounds which include electro-fluid shuttle memory elements, nanosensors [18], [19], etc. Computational procedure The Lennard-Jones (LJ) potential was used to describe the interaction between the carbon atoms for each pair of C60 molecules,

[

VL− J (rij ) = 4ε ij (σ ij / rij ) − (σ ij / rij ) 12

6

]

where rij is the distance between the ith and jth carbon atoms of the pair of interacting fullerenes,

ε and σ are two standard Lennard-Jones (LJ) potential parameters shown in Table 1.

82

Perspectives of nanoscience and nanotechnology

ε kB

C-C K+-K+

[K ]

28 35

σ [Å]

m [10-25 kg]

3.4 3.57

0.199 0.649

Table 1. The carbon and potassium mass and L-J potential parameters used in simulation, taken from [20-22] The fullerenes C60 have been treated as rigid bodies with translational and rotational degrees of freedom. The 60 carbon atoms of fullerene form a truncated icosahedron, which consists of 12 pentagonal and 20 quasihexagonal faces. There are two distinct C-C bond lengths. The values used in this work, 1.37 Å and 1.448 Å were taken from the quantum chemistry calculations [23]. The Lennard-Jones potential was also used for potassium-carbon interaction (parameters given in Table 1) and Coulomb potential between potassium ions. The classical equations of motion were integrated up to 2.4 ns by the Adams-Moulton predictor-corrector algorithm [24]. The integration time step was 2 fs which ensures good stability of the algorithm. The clusters were equilibrated for 10 7 integration steps. The appropriate correlation functions were averaged over 2. 4 ⋅ 10 4 time origins. The origins were separated by a time interval equal to 50 integration time steps. The calculations were carried out for a constant energy ensemble for zero total linear and angular momentum of the whole system. The ground state (global minimum potential energy) configuration of (K+@C60)7 was obtained by Monte Carlo (MC) simulation after 6*106 time steps. The thermodynamic limit could be a problem in the case of constant temperature MD simulations for the system consisting of a very small number of particles and it is always a problem to define a temperature [11]. That is why the simulated physical observables of interest were investigated in the same manner as Gallego et al. [15], i.e. as a function of the total energy Et of the cluster instead of temperature. Results It is known, that the ground state of the solid phase of (C60)7 cluster is a pentagonal bipyramid (PBP) [25, 15]. A pentagonal bipyramid configuration for (K+@ C60)7 cluster after 6*106 simulation steps in Monte Carlo procedure was also obtained (see Fig. 1).

Fig. 1 Snapshot of the minimum potential energy configuration of (K+@C60)7 cluster (PBP). Having initially (K+@ C60)7 in PBP configuration (in the solid state at very low temperature), the energy of the system was gradually increased in a stepwise manner, modeling heating of the system. The calculated mean square displacement of K+@ C60 for several low and high energies,

Solid State Phenomena Vol. 140

83

∆r(t)=r(t)-r(0) where r is the position of the center of mass of single endohedral fullerene, is presented in Fig. 2.

Fig. 2. The mean square displacement of the centre of mass of the K+@C60 molecule in the cluster for three energies, Et The slope of the mean square displacement can be directly connected with the translational diffusion coefficient D of a molecule by the Einstein formula ≅ 6Dt. Fig. 2 shows that both the solid and liquid state appear in the cluster studied, because D > 0 ( D ≅ 0) for high (low) energies, respectively. To look deeper, for the appearance of solid/liquid phase transition the Lindemann index δL was calculated [11]

δL =

1 2 2

(< rij > − < rij > ) 2 ∑ N ( N − 1) i < j < rij > N

2

(1)

where rij is the distance between the centre of mass of ith and jth molecules. The energy evolution of index δL is given in Fig. 3.

Fig. 3. The energy Et dependence of the Lindemann index of (K+@C60)7 cluster. One can see that the value of Lindemann index δL is very low and practically does not change with the energy for Et < -0.30 eV/molecule, thereby confirming the solid state phase of the cluster in this energy region as there is no translational diffusion. Further increasing the energy, further heating, leads to the transition from a solid state towards the liquid phase of (K+@ C60)7. Fig. 4

84

Perspectives of nanoscience and nanotechnology

shows examples of the normalized velocity autocorrelation function Cv(t)=-1 where v is the translational velocity of the center of mass of K+@ C60 , calculated for three energies Et – low, intermediate and high.

Fig. 4. The linear velocity autocorrelation function of the K+@C60 molecule in the cluster, for three energies Et . It can be seen that that for low energy the velocity of the C60 molecule exhibits a deep dip and dumped pulsations, characteristic of the solid phase. The high energy Cv(t) decays more regularly, almost exponentially – the behaviour associated with a softer phase of matter (liquid). The third correlation function Cv(t ) interpolates between the previous extremes. Following this observation, the velocity autocorrelation time τv = ∫ dt Cv(t ) has been calculated as a function of energy Et (see Fig. 5) – the jump of τv arround (-0.25 eV/molecule) is clearly visible.

Fig. 5. The energy dependence of the linear velocity correlation time of the K+@C60 molecule in the cluster. Note, that closely connected with τv is the translational diffusion coefficient D ~ τv [26]. Therefore, the plot τv(Et) gives information about two regions in the (K+ @C60)7 cluster. Firstly, the almost negligible and practically energy independent diffusion – typical of the solid phase. Secondly, quite effective and energy increased diffusion of endohedral fullerenes (liquid). The calculated angular r r r velocity autocorrelation function Cω(t) = < ω (t)٠ ω (0)> < ω (0)2 >-1 in the solid phase (Et < -0.31 eV/molecule) is shown in Fig. 6.

Solid State Phenomena Vol. 140

85

Fig. 6. The angular velocity autocorrelation function of the K+@C60 molecule for different energies Et in solid phase of the cluster. The number associated with a given curve denotes the energy Et in the unit eV/molecule. One can distinguish two types of behaviour of Cω(t). Firstly, for Et ≤ -0.31 eV/molecule the oscillatory, strongly damped pulsations of Cω(t) characteristic for hindered rotations in the solid phase [26 - 30]. The correlation function Cω(t) changes with energy, the lower the energy, the deeper is the dip of Cω(t). The negative dip of Cω(t) near t = 1 ps (Et = -0.47 eV/molecule) vanishes at Et = -0.31 eV/molecule. Secondly, an exponential relaxation for Et > -0.29 eV/molecule similar to the almost free rotations of C60 molecules in bulk fullerite sample above the structural phase transition [16]. Summarizing the facts gathered at this stage; for Et ≤ -0.29 eV/molecule there is a solid phase with emanating rotations but no translations of fullerenes (D ~ 0). Traditionally, for bulk samples one calls this kind of condensation of matter the “plastic phase”. For the reason explained earlier, the results are presented as a function of the total energy Et. Nevertheless, if the classical definition of temperature from the kinetic theory of perfect gas is applied, the formula for the temperature T reads T = 2 EK / k B (6 N − 6) , where 6N-6, states the internal degrees of freedom, kB is Boltzmann constant and is the average kinetic energy, N is the number of molecules, i.e. N=7 in this case. To give some impression of what temperature (kinetic theory ) would correspond to a given energy Et , in Fig. 7 the appropriate calibration curve i.e. the plot (Et) is shown.

Fig. 7. The dependence of the average kinetic energy on the total energy Et for (K+@C60)7 cluster.

86

Perspectives of nanoscience and nanotechnology

Note a slight change of the slope of (Et) in the phase transition region, this effect was also observed by Gallego et al. [15] in the case of Girafalco’s interaction potential between fullerenes. For example, the reported jump of τv around Et ≈ –0.24 eV/molecule (Fig. 5) would correspond to the temperature T ≈ 350 K of the cluster. In Fig. 8 we present the calculated standard deviation of 2 2 total energy δ Et = Et − Et

2

.

2 2 Fig. 8. The standard deviation of total energy δ Et = Et − Et

2

as a function of Et for

+

(K @C60)7 cluster. Note, that /kBT2 would be proportional to the specific heat at constant volume [26]. Here again, the spectacular increase of the energy fluctuation is observed for higher energies (Et > -0.24 eV/molecule). Naturally, the jumps of the observables, τv, δL and can be associated with the phase transition in the cluster. It can be seen that the phase transition of (K+@ C60)7 cluster covers a wide energy region ∆Et = 0.06 eV/molecule, from -0.26 eV/molecule to about -0.21 eV/molecule. It corresponds to the temperature range between 340 and 400 K. The endohedral fullerene’s centre of mass radial distribution function g(r) is presented in Fig. 9.

Fig. 9 The radial distribution function of the centre of mass of K+@C60 molecule in the cluster

Solid State Phenomena Vol. 140

87

For the low energy Et = -0.47 eV/molecule (solid phase) it shows two peaks associated with the pentagonal bipyramid structure of the cluster. At the high energy Et = -0.17 eV/molecule the sharp peaks of g(r) shift and broaden out confirming the destruction of pentagonal bipyramid structure and liquification of the cluster.

Conclusions The atomistically detailed MD simulations enabled evidence of the solid/liquid phase transition and the existence of the liquid phase in a (K+ @C60)7 cluster to be discovered. This may be of some importance, taking into account the possible future applications of endohedral fullerene nanocompounds. For example, the application of an external electric field to the (K+ @C60)7 cluster in the liquid phase can induce the guided transport of endohedral potassium ion fullerenes.

Acknowledgement The support of the Academy of Finland through consortia NAKAMA and NANOTOMO - research projects No. 124059 and 117867, respectively, is gratefully acknowledged.

References [1] A. Cheng, M. I. Klein and C. Caccamo, Phys. Rev. Lett. 71 (1993), p.1200 [2] M. H. J. Hagen, E. J. Meijer, G. C. A. M. Mooij, D. Frenkel and H. N. W. Lekkerkerker, Nature (London) 365 (1993), p.425 [3] N. W. Ashcroft, Nature (London) 365 (1993), p387 [4] L. Mederos and G. Navascues, Phys. Rev. B 50 (1994), p.1301 [5] C. Caccamo, D. Costa and A. Fucile, J. Chem. Phys. 106 (1997) p255. [6] C. Rey and L. J. Gallego, Phys. Rev. E, 53 (1996), p.2480 and the references therein [7] S. B. Legoas, R. Giro, D. S. Galvado, Chem. Phys. Lett. 386 (2004), p.425 [8] Y. Zhu, S. Granick, Phys. Rev. Lett. 93 (2004), p.096101 [9] M. Skrzypek and Z. Gburski, Europhys Lett. 59 (2002), p.305 [10] P. Jena, B. K. Rao and S. K. Khanna, The Physics and Chemistry of Small Clusters, Plenum, New York, 1987 [11] R. S. Berry, T. L. Beck, H. L. Davis and J. Jelinek, Adv. Chem. Phys. 70B (1987), p.75 [12] A. Dawid and Z. Gburski, Phys. Rev. A, 58 (1998), p.740 [13] A. Dawid and Z. Gburski, Phys. Rev. A, 56 (1997), p.3294 [14] Y. Zhou, M. Karplus, K. D. Ball, R. S. Berry, J. Chem. Phys. 116 (2002), p.2323 [15] L. J. Gallego, J. Garcia-Rodeja, M. M. G. Alemany and C. Rey, Phys. Rev. Lett., 83 (1999), p.5258 [16] L. A. Girifalco, J. Phys. Chem. 96 (1992), p.858 [17] A. Piątek, A. Dawid and Z. Gburski, J. Phys. Condens. Matter, 18 (2006), p.8471 [18] J. W. Kang, H. J. Hwang, Computational Materials Science, 33 (2005), p.338 [19] N. Y. Pan, J. S. Shih, Sensors and Actuators B, 98 (2004), p.184 [20] W. A. Steele, The Interaction of Gases with Solid Surfaces (New York: Pergamon) 1974 [21] A. Cheng and M. Klein, Phys. Rev. B, 45 (1992), p.1889 [22] J. W. Kang, H. J. Hwang, Physica E, 23 (2004), p.36 [23] G. E. Scusseria, Chem. Phys. Lett., 176 (1991), p.423 [24] D. C. Rappaport, The Art of Molecular Dynamic Simulation (Cambridge: Cambridge University Press) 1995 [25] E. Sawada and S. Sugano, Z. Phys, D 14 (1989), p.247 [26] J. P. Hansen and I. P. Mc Donald, Theory of Simple Liquids (London: Academic) 1986

88

Perspectives of nanoscience and nanotechnology

[27] A. Piątek, A. Dawid, Z. Gburski, J. Molec. Struc. 792 (2006), p.82 [28] Z. Dendzik, M Paluch, Z Gburski, J Zioło, Journal of Physics: Condensed Matter, 9 (23) (1997) L339 [29] D. Chrobak, K. Nordlund and R. Nowak, Phys. Rev. Lett. 98, (2007) p.045502. [30] A. Dawid, Z. Dendzik, Z. Gburski, J. Molec. Struc. 704 (1) (2004) p.173.

Solid State Phenomena Vol. 140 (2008) pp 89-96 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.89

Molecular Dynamics Simulation Study of the Liquid Crystal Phase in a Small Mesogene Cluster (5CB)22 W. Gwizdała a, A. Dawid b and Z. Gburski c Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007, Katowice, Poland e-mail:

a

[email protected]; b [email protected]; c [email protected]

Keywords: Mesogene cluster, nanoscale liquid crystal, molecular dynamics simulation (MD), 5CB

Abstract. The molecular dynamics (MD) technique was used to investigate the nano droplet composed of twenty mesogene molecules 4-cyano-4-n-pentylbiphenyl (5CB). The 5CB molecules were treated as rigid bodies, the intermolecular interaction was taken to be the full site-site pairwise additive Lennard-Jones (LJ) potential plus a Coulomb interaction. The radial distribution functions in the temperature range from 150 to 400 K, were calculated as well as the linear and angular velocity autocorrelation functions. In addition the total dipole moment autocorrelation function and dielectric loss of (5CB)22 mesogene cluster were calculated and the liquid crystal ordering in the nanoscale system was studied up to its vaporization temperature. Introduction In recent decades, there have been many simulation studies of the liquid crystals [1]. Before commencing any simulation of the physical properties of liquid crystals, the initial choice to be made is usually between the use of phenomenological approaches, such as the mean-field density functional theory, or molecular level methods such as that of Gay-Berne [2 ], or Lebwohl-Lasher [3]). Recently, a number of molecular dynamics studies of bulk liquid crystals samples have been carried out using fully atomistic models [4 - 6]. These simulations allow for a detailed description of the molecular structure and are essential for a quantitative investigation of liquid crystalline properties, which are sensitive to the molecular details. In establish if, and how, the peculiarities of the dynamics of mesogene molecules near the surface of nanoscale droplets influence the tendency for ordering in liquid crystal mesogenes was the reason for investigating the extremely small mesogene ensemble composed of only twenty two mesogenes. The phenyl-based nematogen 5CB (4-cyano-4-n-pentylbiphenyl) was chosen because of its known nematic and smectic phases in a bulk sample close to room temperature [7, 8]. Simulation details The representative model of 4-cyano-4-n-pentylbiphenyl (5CB) molecule is shown in Fig. 1. The Van der Waals potential (Lennard-Jones + Coulombic interaction) was used to describe the interaction between atoms (sites) of interacting mesogene molecules in (5CB)22 cluster. 12 6 V (rij ) = 4ε ij (σ ij / rij ) − (σ ij / rij )  + Zi Z j e2 / 4πε 0 rij  

(1)

where rij is the distance between the ith and jth atoms of the pair of different molecules, Zn is the electric charge of nth site.

90

Perspectives of nanoscience and nanotechnology

Fig. 1 The model of 5CB molecule (hydrogen atoms aren’t displayed). The Lennard-Jones interaction parameters between two different sites were calculated using the Lorentz-Berthelot Rule [9] (see Table 1). εii [kJ mol-1] m [10-25 kg]

site

σii [nm]

N

0.3250

0.72

0.232581

C1

0.3750

0.44

0.199438

C2, C7, C8, C13

0.3750

0.46

0.199438

CH

0.3750

0.46

0.216174

CH2

0.3905

0.50

0.232911

CH3

0.3905

0.50

0.249648

Table 1 The potential model parameters of 5CB, taken from [10].

Site charges distribution are given in Table 2. atom

charge |e|

N

-0.450

C1

0.100

C2

0.300

C3, C4

-0.075

C5, C6

-0.070

C7

0.250

C8

0.250

C9, C10

-0.070

C11, C12

-0.060

C13

0.100

C14-C18

0

Table. 2. The distribution of 5CB sites charges, taken from [10] (see Fig. 1)

Solid State Phenomena Vol. 140

91

The mesogenes were treated as rigid bodies with no internal degrees of freedom. The classical equations of motion were integrated up to 5 ns by the Adams-Moulton predictor-corrector algorithm [9], the integration time step used was 1 fs which ensured good total energy conservation. The starting configuration was prepared via Monte Carlo potential energy minimization. The average energy was adjusted as desired by a process of velocity scaling. The cluster was equilibrated for 10 6 MD steps. The time-dependent autocorrelation functions were calculated by averaging the appropriate physical variables over 10 4 time origins. The origins were separated by a time interval equal to 50 integration time steps. The calculations were carried out for a constant energy ensemble for zero total linear and angular momentum. Results A representative snapshot of the instantaneous configuration of (5CB)22 at the temperature of 150 K is shown in Fig.2 which reveals the evident spatial ordering of 5CB molecules within the volume of the cluster. Their long axes of symmetry are arranged in more or less similar, apparently correlated directions. To check the level of molecular order in the cluster, the authors constructed the order tensor Qij = 3cos Θi cos Θ j − δ ij / 2 where i, j = x, y, z are indices referring to the laboratory frame. The bracket represents an average over the whole sample and simulation time. The Θ is the angle between the molecular long axis and the eigenvector nˆ of the order tensor Qij corresponding to the maximum eigenvalue. The nˆ vector is usually called the director of a sample. The largest eigenvalue of the order tensor Qij is the second-rank order parameter P2 [10 16] ( 0 ≤ P2 ≤ 1), P2 =0.72, 0.62, 0.54 for T= 150 K, 300 K and 350 K, respectively was obtained. These high values of P2 indicate liquid crystal ordering in the nanodroplet.

Fig. 2 The snapshot of the simulated system at a temperature of 150 K. r The plot of the mean square displacement of 5CB molecule is shown in Fig. 3, where r r r r ∆ r (t)= r (t) - r (0) and r is the position of the centre of mass of 5CB, indicates that at 150 K the solid phase of the cluster exists whereas at higher temperatures the translational diffusion of molecules occurs.

92

Perspectives of nanoscience and nanotechnology

Fig. 3. The short-time part of the mean square displacement of the centre of mass of 5CB molecule at several temperatures. Hence at higher temperature the cluster is not in the solid state. That is made clear by the Einstein relationship which connects the slope of the linear part of the mean square displacement with the r translational diffusion coefficient D of a molecule where > 6Dt. The calculated diffusion coefficient D increases with rising of temperature, for example D = 0.37 · 10-6 cm2/s for T = 300 K and D = 1.73 · 10-6 cm2/s for T = 400 K. The soft, spatially ordered phase of mesogene molecules means a liquid crystal phase. The radial distribution function g(r) of the centre of mass of the 5CB molecule is shown as Fig. 4.

Fig. 4 The radial distribution function of 5CB mesogene at several temperatures The appearance of sharp peaks at 150 K confirms the solid state of the cluster, for higher temperatures the distribution g(r) becomes much broader, characteristic for soft matter, the liquid phase. To obtain further insight into mesogene dynamics the normalized velocity autocorrelation function r r r r Cv(t) = < υ (t)٠ υ (0)> < υ (0)2 >-1, where υ is the translational velocity of the center of mass of 5CB was calculated. Fig. 5 shows that - in spite of the huge mass of the 5CB molecule - its velocity changes rapidly, being almost entierly decorrelated after 2 ps.

Solid State Phenomena Vol. 140

93

Fig. 5 The translational velocity autocorrelation function of 5CB molecule at several temperatures The oscillations of Cv(t) characteristic for the solid phase at low temperature disappear at higher r r r temperatures > 300 K. The angular velocity autocorrelation function Cω(t) = < ω (t)٠ ω (0)> < ω (0)2 >-1 , presented in Fig. 6, decays even faster than Cv(t). The negative dip of Cω(t) at t ≈ 0.3 ps (T=300 K) almost vanishes at T=400 K. Overall, Cω(t) changes with temperature, the lower the temperature the greater is the dip of Cω(t).

Fig. 6. The angular velocity autocorrelation function of 5CB molecule at several temperatures. In a standard dielectric experiment the frequency dependence of the dielectric loss ε’’(ν), which is the imaginary part of complex dielectric permittivity [17]

ε * (ν ) = ε ' (ν ) − iε " (ν ) , i = − 1 is measured. In case of pure dipolar absorption in the classical limit (i h  0) ε ' ' (ν ) is related to the cosine Fourier transform of the total dipole moment r M (t ) autocorrelation function,

r

r

r

φ (t ) = M (t ) ⋅ M (0) M 2 (0)

−1

94

Perspectives of nanoscience and nanotechnology



ε ' ' (ν ) ≈ ν ∫ dtφ (t ) cos(2πνt )

(2)

0

N r r M = ∑ µi , where µri is the electric dipole moment of ith molecule. i =1

The normalized correlation function φ (t ) of 5CB cluster at several temperatures is presented in Fig. 7.

Fig. 7 The total dipole moment autocorrelation function of (5CB)22 cluster The correlation function φ (t ) is significantly dependent on the temperature. The higher the temperature of the system, the faster is the dipolar relaxation, because mesogene molecules encounter more vigorous motion within the layer. Figure 8 presents the simulated normalized ∞

dielectric loss

ε n ' ' (ν ) = ν ∫ dtφ (t ) cos(2πνt )

of the mesogene thin film.

0

Fig. 8 The normalized dielectric loss of (5CB)22 cluster. The sensitivity of ε n ' ' (ν ) to temperature change is clearly observed. The higher the temperature, the greater is the absorption as a result of the increasing mobility of 5CB molecules.

Solid State Phenomena Vol. 140

95

Summary MD simulations of the mesogene cluster (5CB)22 over a range of temperature have been presented. The solid phase of (5CB)22 is observed at a very low temperature T = 150 K and the partly ordered liquid phase at higher temperatures > 300 K. Note, that even the presence of a surface (finite size) effect was not able to destroy the tendency of the rod-like 5CB molecules to develop spatial and orientational ordering in a cluster. The maximum dielectric loss of (5CB)22 cluster in the vicinity of frequency region νmax ≈ 5 · 1010 Hz is predicted. The authors preliminary studies may serve as guidance for the future experiments with various small mesogene clusters. These kinds of nanosystems might be considered as a component of a future generation of mesogene based displays. References

[1]. J. W. Doane, Mater. Res. Bull.,16 (1991) 22 and references therein. [2] J. G. Gay and B. J. Berne, J. Chem. Phys. 74 (1981) 3316. [3] P. A. Lebwohl and G. Lasher, Phys. Rev. A 6 (1972) 426. [4] M Tsige, M. P. Mahajan, C. Rosenblatt, and P. L. Taylor, Phys. Rev. E, 60 (1999) 638. [5] Z. Wang, J. A. Lupo, S. Patnaik, R. Pachter, Computational and Polymer Science, 11 (2001) 375. [6] A. Piątek, A. Dawid , Z. Gburski, J. Phys.: Condensed Matter, 18 (37) (2006) 8471. [7] T. Manisekaran, T. K. Bamezai, N. K. Sharma, J. Shashidhara Prasad, Liquid Crystals, 23 (1997) 597. [8] S. M. Risser and K. F. Ferris, Mol. Cryst. and Liq. Cryst. 373 (2002) 143 [9] D. C. Rappaport, The Art of Molecular Dynamic Simulation, Cambridge University Press, Cambridge, 1995. [10] A. V. Komolkin, A. Laaksonen, and A. Maliniak, J. Chem. Phys. 101 (1994) 4103. [11] P. Pasini,C. Zannoni, Advances in the Computer Simulations of Liquid Crystals: Proceedings of the NATO Advanced Study Institute on Advances in the Computer Simulations of Liquid Crystals Erice, Italy 11-21 June 1998. [12] M. Tsige, Milind P. Mahajan, C. Rosenblatt, P. L. Taylor Phys. Rev. E, 60 (1999), 1. [13] S. Pałucha, P. Brol, M. Kośmider, Z. Dendzik, Z. Gburski, J. Molec. Struct. 704 (2004) 263. [14] A. Dawid, Z. Gburski, Phys. Rev. A, 56 (1997) 3294. [15] A. Dawid, Z. Dendzik, Z. Gburski, J. Molec. Struct. 704 (2004) 173. [16] C. G. Joslin, C. G. Gray, Z. Gburski, Molec. Phys.53 (1) (1984) 203. [17] A. K. Jonscher, Dielectric Relaxation in Solids, Chelsa Dielectric, London, 1983.

Solid State Phenomena Vol. 140 (2008) pp 97-102 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.97

Preparation and Characterization of Polymer/Multi-walled Carbon Nanotube Nanocomposites M. Sarah Mohlala a and Suprakas Sinha Ray b National Centre for Nano-Structured Materials, Council for Scientific and Industrial Research, Pretoria, 0001, Republic of South Africa e-mail: a [email protected]; b [email protected] Keywords: Polycaprolactone; Poly(butylene adipate-co-polycaprolactam); Nanocomposites; Thermal properties

Abstract. This paper describes the preparation, characterization and properties of nanostructured composite materials based on poly(butylene adipate-co-polycaprolactam) (PBA-co-PCL)/multiwalled carbon nanotubes (MWCNTs) and polycaprolactone (PCL)/MWCNTs. The polymer/MWCNTs nanocomposites were prepared by mixing the polymers with various amounts of MWCNTs using both solution and melt blending processes. The dispersion of MWCNTs into the polymer matrix was analyzed by transmission electron microscopy (TEM) and the thermal stability of the nanocomposites was studied by thermal gravimetric analysis (TGA). Differential scanning calorimetry (DSC) was used to study the crystallization and melting behaviour of the polymer matrices containing the MWCNTs. Introduction Carbon nanotubes (CNTs) are known to possess unique mechanical, optical, electrical and thermal properties as well as having good chemical stability [1-4]. Due to their unique properties, many researchers have endeavoured to fabricate advanced CNT-based composite materials that possess enhancement of one or more of these properties. The electrical, mechanical, thermal and physical properties of the polymeric materials can be improved by the incorporation of minute amount of CNTs [5,6]. The promising area of composite research involves an enhancement of the mechanical and thermal properties of a polymer using CNTs as the reinforcing material. The study of the behaviour [7] and dispersion of CNTs [6] in the polymer matrix have been reviewed by other researchers. Aliphatic polyesters such as PBA-co-PCL and PCL are biodegradable polymers with semi crystalline characteristics. Both polymers can degrade by hydrolysis of their ester linkages in their polymer chains. Therefore, the aliphatic polyesters have are considered for a wide range of possible applications, such as biodegradable packaging materials, disposable films, implantable biomaterials and microparticles for drug delivery. The polyester, poly(butylene succinate-co-adipate), mixed with layered silicates has been reported as forming nanocomposites with improved thermal and mechanical properties [8]. Recently, Chen et al. reported on the preparation of PCL/MWCNTs composites by ultrasonically mixing the PCL and MWCNTs in a THF solution [9]. It was shown that the MWCNTs were well separated and uniformly distributed in the polymer matrix. Investigations on the preparation of PCL/CNT composites using functionalized MWCNTs have been studied [10]. From this study, it was demonstrated that the functionalized MWCNTs were also well separated and randomly distributed in the PCL matrix. The authors had no knowledge of any study on the preparation of PCL/MWCNT nanocomposites using the melt process method, and no work has been reported on PBA-co-PCL/MWCNT nanocomposites. Therefore, in this study PCL/MWCNT nanocomposites were fabricated by mixing the PCL and pure MWCNTs using the melt process, while PBA-co-PCL/MWCNT nanocomposites were produced by mixing PBA-coPCL and MWCNTs in chloroform using the solution blending method.

98

Perspectives of nanoscience and nanotechnology

Experimental Procedure Preparation of polymer composites. PCL , PBA-co-PCL and MWCNTs with diameters in the range of 10-20 nm were purchased from Sigma Aldrich (S.A). The polymers were dried at 70 °C before they were used. PBA-co-PCL and MWCNTs (0.2, 0.1, and 0.02 wt. %) were mixed together in an attempt to prepare polymer–MWCNT composites using solution processing. PBA-co-PCL was dissolved in 50ml of chloroform using a magnetic stirrer at room temperature for 4 hrs. The MWCNTs were sonicated in 20ml of chloroform for 2 hrs. The PBA-co-PCL was then transferred to the nanotube/chloroform suspension and the mixture was further sonicated for 2 hrs. The suspension was the transferred to a watch glass and air dried in a fume hood overnight to give a fractured film of the composite. To obtain the desired composites, raw pellets of PCL were melted at 80 °C, at the 60 rpm speed, in a thermo mixer (Haake Rheomix Os). Various amounts of PCL were then mixed with 2, 1 and 0.5 wt. % of MWCNTs at 80 °C and mixed for 10 min at a speed of 60 rpm. The dispersion of carbon nanotubes in the polymer matrices were examined by TEM (Joel JEM 100S), operated at an accelerating voltage of 80 kV. The samples were prepared by ultrasonicating the nanocomposites samples suspended in methanol and then placed on a Cu grid for analysis. The morphology of the nanocomposite materials were studied by SEM (Leo 1525 FE-SEM). The samples were prepared by fracturing the samples under nitrogen and they were carbon coated prior imaging to avoid charging. The thermal stabilities of the materials were carried out by thermogravimetric analysis using TA Instruments Q500 equipment at a heating rate of 10°C/min under nitrogen from room temperature to 800°C. The melting and crystallization temperatures of pure polymers and nanocomposites were analyzed by Differential Scanning Calorimetry Q200 system (TA instruments) in the temperature range of -60 to 125°C for PBA-co-PCL and -60 to 80°C for PCL at a heating rate of 10°C/min in an atmosphere of nitrogen. Results and discussions PBA-co-PCL/MWCNTs nanocomposites. TEM images of PBA-co-PCL/MWCNTs nanocomposites with two different weight ratios of MWCNTs are shown in Figure 1. It can be clearly seen that the MWCNTs are well separated and randomly distributed in the polymer matrix. This observation indicates that solution blending method is a suitable method for dispersing MWCNTs homogeneously into the PBA-co-PCL matrix.

Figure 1. TEM images of PBA-co-PCL/MWCNTs nanocomposite (a) 0.1 wt. % MWCNTs loading and (b) 0.02 wt. % MWCNTs loading.

Solid State Phenomena Vol. 140

99

In order to study the effect of MWCNTs on the thermal stability of PBA-co-PCL matrix, TGA analysis of pure PBA-co-PCL and PBA-co-PCL/MWCNTs nanocomposite samples was conducted. Figure 2 shows the TGA curves of PBA-co-PCL and PBA-co-PCL/MWCNTs nanocomposites obtained at a heating rate of 10°C/min in a nitrogen atmosphere. TGA results for 0.02 wt.% showed to have a slightly higher decomposition temperature than 0.1 and 0.2 wt.%, showing that the addition of small amount of CNTs influences the thermal properties more than larger amounts.

0.02 wt.% 0.1 wt.% 0.2 wt.% pure PBA-co-PCL

100

Weight %

80

60

40

20

0 100

200

300

400

500

600

700

o

Temperature ( C)

Figure 2. TGA analysis of PBA-co-PCL/MWCNTs nanocomposites

When MWCNTs were added to the PCL by the melt process, good dispersion of CNTs in the polymer matrix was obtained as shown in Figure 3.

Figure 3. SEM image of PCL-2 wt.% MWCNT nanocomposite Surprisingly, there was no enhancement in the thermal stability of the polymer (Figure 4), instead the addition of MWCNTs reduced the thermal stability of the PCL. This could be due to the fact that MWCNTs contain a lot of Van der Waals forces among themselves, therefore making them difficult to disperse into the polymer matrix. The study by Chen et al. revealed that PCL/MWCNTs nanocomposites prepared by ultrasonication method in THF, have improved thermal properties relative to pristine PCL [8].

100

Perspectives of nanoscience and nanotechnology

a

Heat Flow/ Endo down

0.5 wt.% 1 wt.% 2 wt.% pure PCL

-40

-20

0

20

40

60

80

o

Temperature ( C)

Figure 4. TGA analysis of PCL/MWCNTs nanocomposites.

DSC results represented in Figure 5 show that the addition of MWCNTs in the polymer matrix lowered the Tm and increased the Tc of the PCL. b

Heat Flow/ Exo Up

0.5 wt.% 1 wt.% 2 wt.% pure PCL

.

-40

-20

0

20

40

60

o

Temperature ( C)

Figure 5. DSC analysis of PCL/MWCNTs nanocomposites (a) melting and b) crystallization behaviour. This indicates that even though the addition of MWCNTs into the PCL did not enhance the thermal stability of PCL, there was an interaction between PCL and MWCNTs, therefore further studies will be carried out to establish this behaviour. Conclusions The results of TEM show that nanocomposites with well separated and randomly distributed PBAco-PCL/MWCNT have been fabricated by mixing the PBA-co-PCL and MWCNTs in chloroform with ultrasonic treatment. The addition of small amounts of MWCNTs into the pure polymer showed a slight enhancement of the thermal stability of PBA-co-PCL. There was no improvement

Solid State Phenomena Vol. 140

101

in the thermal stability of the PCL when small amounts of MWCNTs were added to the polymer using the melt mixing process. Future work involves studying the detailed mechanical properties of prepared polymer/MWCNTs nanocomposites. Functionalized MWCNTs will be mixed with polymers (PCL and PBA-co-PCL) in an attempt to improve the interaction between the polymer and CNTs in order to prepare polymers with improved thermal and mechanical properties. Acknowledgements The authors express their thanks to the Council for Scientific and Industrial Research (CSIR) and the Department of Science and Technology (DST) for funding the project. References [1] S.B. Sinnott, R. Andrews: Crirical Reviews in Solid State and Materials Sciences Vol. 26 (2001), pp. 145. [2] E.T. Thostenson, Z. Ren, T-W. Chou: Composites Science and Technology Vol. 61 (2001), pp. 1899. [3] F. Li, M. Cheng, S. Bai, G. Su, M.S. Dresselhaus: Applied Physiscs Letters Vol. 77 (2000), pp. 3161. [4] J.P. Lu: Journal of Physics and Chemistry Solids Vol. 58 (1997), pp. 1649. [5] W. Tang, M.H. Santare, S.G. Advani: Carbon Vol. 41 (2003), pp. 2779. [6] X-L. Xie, Y-W. Mai, X-P. Zhou: Materials Science and Engineering R Vol. 49 (2005), pp. 89. [7] I. Szleifer, R. Yerushalmi-Rozen: Polymer Vol. 46 (2005), pp. 7803. [8] S. Sihna Ray, M. Bousmina: Polymer Vol. 46 (2005), pp.12430. [9] E-C. Chen, T-M. Wu: Polymer Degradation and Stability Vol. 92 (2007), pp. 1009. [10] K. Saeed, S-Y. Park, H-J. Lee, J-B. Baek, W-S. Huh: Polymer Vol. 47 (2006), pp. 8019.

Solid State Phenomena Vol. 140 (2008) pp 103-108 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.103

Formation of Carbon Fibres in High-voltage Low-current Electrical Discharges A.T. Sobczyk.1, a, A. Jaworek 2, b, E. Rajch2, c, M. Sozańska3 ,d 1

2

Institute of Fluid Flow Machinery, Polish Academy of Sciences, ul. Fiszera 14, 80-952 Gdansk, Poland

Institute of Physics, Pomeranian Academy, ul. Arciszewskiego 22B, 76-200 Slupsk, Poland 3

Silesian University of Technology, Department of Materials Science, ul.Krasinskiego 8, 40-019 Katowice, Poland

e-mail: a [email protected]; b [email protected]; d [email protected]

c

[email protected],

Keywords: carbon fibres, electrical discharges in gases, corona discharge

Abstract. A novel method to synthesize carbon fibres using low-current electrical-discharge plasma in hydrocarbon vapours is presented in the paper. The low-current arc discharge of positive polarity was generated between a stainless steel needle, and a nickel alloy plate, over a voltage range from 2 kV to 30 kV. The discharge was stabilised by a high series resistance (1 – 12.5 MΩ). The experiments were carried out in an argon atmosphere at normal temperature and atmospheric pressure. The arc discharge of current in the range of 1 to 4 milliamps was found to be a potentially effective method for the production of carbon fibres. The diameter of the fibres varied from about 20 to 120 µm with a growth rate of about 0.5 mm/s. Introduction Carbon fibres are usually synthesized by spinning from a polymer precursor. Wet spinning is used for polymers, for example, polyacrylonitrile (PAN) dissolved in an organic solvent, while melt spinning is used for molten substances, for example, pitch [1]. The polymer fibre is next pyrolised to remove hydrogen or other elements from its chemical structure, and the remaining carbon atoms form carbon fibre due to sp2 and sp3 bonds between them. Vapour-grown carbon fibre (VGCF) methods [2,3], which are based on the pyrolysis of hydrocarbon vapours, such as methane, benzene, or acetylene are used less frequently. Electrical discharges have many technological applications in processes such as thin film deposition, ion implantation, or the removal of NOx or hydrocarbons from flue gases. Nowadays, high-current arc discharge is one of the most effective methods used to obtain single- and multiwalled carbon nanotubes (CNT) and carbon fibre synthesis [4]. In 1991 Iijima first discovered carbon nanotubes in a deposit, which remained after a high-current arc discharge. In recent years, low-current electrical discharges were also investigated as a method for CNT synthesis. Li et al. [5] synthesised CNTs in corona-discharge plasma at atmospheric pressure. MWCNTs with a diameter of about 40 nm were produced using an anodic aluminium oxide template on a stainless steel plate and cobalt as a catalyst. The discharge was generated from a tungsten needle in a mixture of methane and hydrogen. The electrode system was powered by an ac voltage of 8 kV/25 kHz. Sano and Nobuzawa [6], using a tungsten needle as discharge electrode and a flat-ended graphite rod as the ground electrode, obtained CNTs of diameter in the range from 10 nm to 30 nm on the tip of the needle. The discharge was generated in an atmosphere of ethylene and hydrogen at a voltage of 1.6 kV. The authors of these papers, however, do not make any reference to the synthesis of carbon

104

Perspectives of nanoscience and nanotechnology

fibres which were obtained in our experiments for higher voltages and larger inter-electrode distances. Carbon fibres were synthesised using a low-current discharge by Brock and Lim [7]. Their experiment was similar to ours but was carried out by a point-to-plate corona discharge of negative polarity, at a voltage of 3 kV, in n-heptane vapour using nitrogen as a carrier gas. In this experiment, the growth rate of the fibre was of the order of millimetres per second. The fibres produced had diameters in the range from 25 to 90 µm. An advantage of using electrical discharge for the synthesis of carbon fibres is that it is a highly efficient and low cost process. The paper presents a novel method of carbon fibre synthesis using low-current electrical-discharge plasma in hydrocarbon vapours. The experiments were carried out using argon as the carrier gas at normal temperature and pressure. The low-current arc discharge was found to be an effective method for the production of carbon fibres. Experimental The electrical discharge was generated between two electrodes: a needle electrode and a plate, in a reactor chamber 170 mm long and cross section 68 mm × 68 mm as shown in Figure 1.

Fig. 1. Diagram of the experimental arrangement. The discharge electrode was a stainless steel needle (70% Fe, 18% Cr, 10% Ni, 1% Si), 1 mm in diameter, while the plate electrode, of dimensions 67 mm x 90 mm, was made of nickel alloy (75% Ni, 17% Fe, 5% Mo, 2% Mn). The distance between the electrodes was set at 15 mm. The fibres were synthesized from cyclohexane as a feedstock vapourized by argon flowing at a flow rate of about 0.08 dm3/s, through a gas-washing bottle. The cyclohexane concentration in the argon was about 8000 ppm. The cyclohexane was supplied by Chempur (Poland) and Argon of purity 99.99% was purchased from Messer Polska (Poland). The discharge electrode was connected to a high voltage DC power supply (Spelmann HV SL 300). The voltage was controlled in the range from +2 kV to +30 kV. A series ballast resistor R=1 MΩ to 12.5 MΩ was used to stabilise the discharge current. The current, which was measured with a moving-coil ammeter, varied from 0.1 mA to 5.6 mA. The discharges were generated at atmospheric pressure and room temperature. If not otherwise stated, the time for carbon fibre synthesis for each run in these experiments was 30 seconds. The morphology of the carbon structures was examined with a Hitachi S3400N Scanning Electron Microscope. Results and discussion Two plasma columns of a low-current arc discharge (the bright zones at the photographs) for various discharge conditions are shown in Figure 2.

Solid State Phenomena Vol. 140

105

b.

a.

Fig. 2. Photographs of plasma column of low-current arc discharge during growth of carbon fibre in argon with 8000 ppm of cyclohexane: (a) voltage U=20 kV, discharge current i= 1.78 mA; (b) voltage U= 30kV, discharge current i=5.6 mA. The arrows indicate the tip of the needle; the tip of carbon fibre stretching out from the needle point, which is not visible in Figure 2a, but can be seen in the reflected light in Figure 2b; and the soot layer at the plate electrode. The fibre was relatively straight with a smooth surface. With positive polarity of the discharge electrode, the carbon fibres grew at voltages within the range from 17 kV to 30 kV. The current for these voltages was in the range from 1.5 mA to 5.6 mA respectively. The current-voltage characteristics of the discharge in argon are shown in Figure 3. 10 increasing supply voltage decreasing supply voltage

discharge current [mA]

1

discharge in cyclohexane

0.1

0.01

0.001

0.0001 0

1

2

3

4

5

6

7

8

9

10

voltage across the electrodes [kV]

Fig. 3. Current-voltage characteristics of positive corona discharge in Ar. The group of scattered points (triangles at upper part of the plot) refer to the discharge in argon with 8000 ppm of cyclohexane. For low voltages, the current was of the order of a few microamps. For a sufficiently high voltages, the discharge was changed to a low-current arc discharge. Due to thermal ionisation in the plasma column, the discharge current increased significantly and the voltage drop between the electrodes

106

Perspectives of nanoscience and nanotechnology

decreased. The magnitude of the discharge current was limited only by the series resistance. When cyclohexane is added to the gas, the characteristic in the arc mode is not reproducible and the magnitude of the current depended on the stage of carbon fibre synthesis. For this case only scattered current-voltage points are shown in the plot (Figure 3). The low-current arc discharge provided the conditions suitable for carbon fibre synthesis. It was noticed, that the current was not constant during the growth of the carbon fibre but increased slightly, by up to 10% as indicated in Figure 4. 4.5

discharge current [mA]

4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 0.0 0

5

10

15

20

25

30

35

40

time of synthesis [s]

Fig. 4. Example of changes in the discharge current during the growth of carbon fibre. Nominal discharge current = 3.5 mA. The increase in the discharge current can be explained by a decrease in the length of plasma column as the carbon fibre increases in length. Because the carbon fibre has higher conductivity than the plasma, the total resistance of the discharge circuit decreases, and the current increases. The dependence of the growth rate of the carbon fibre on the discharge current is illustrated in Figure 5. 0.5

growth rate [mm/s]

0.4

0.3

0.2

0.1

0 1

2

3

4

5

6

discharge current [mA]

Fig. 5. Growth rate of carbon fibres synthesised in low-current arc discharge vs. discharge current. The fibre length was determined from the photographs taken at the beginning, and at the end of the process of synthesis. For a discharge current in the range from 1.2 mA to 1.8 mA, the growth rate increased with the increase in current, and the fibese produced were homogenous. For the current

Solid State Phenomena Vol. 140

107

increasing from 2 mA to 2.8 mA, the growth rate of carbon fibre decreased. The optimum conditions for carbon fibre synthesis were found to be a current of about 1.8 mA at a voltage of 21 kV. The most intense growth of carbon fibre was obtained under these conditions. After a synthesis time of 30 seconds the length of carbon fibre was about 4 mm and its diameter about 70 µm. For higher currents, increasing from 3.6 mA to 5.6 mA, the length of fibres ranged from about 7 mm to 12.5 mm. Further investigations, for voltages higher than 30 kV, were impossible because of the limitations of the power supply used. SEM images of two carbon fibres synthesised at a voltage of 20 kV are shown in Figure 6.

a.

b. Fig. 6. Photographs of carbon fibre tip synthesized at a voltage of 20 kV. (a) for discharge current of 1.78 mA; (b) for discharge current of 3.46 mA.

It is evident from Figure 6 that substantial changes in morphology of carbon fibres occur with an increase in the discharge current. For the lower currents, 1.78 mA, the surface of the carbon fibre was relatively smooth, while for the higher currents of 3.46 mA, the fibre surface developed a rough, pebble-like texture. The diameter of the fibre was also larger for the higher discharge current. Changes in the morphology of the cross section of the fibres synthesised at different discharge currents are shown in figure 7.

a.

b. Fig. 7. Photographs of carbon fibre cross section synthesized at a voltage of 20kV. (a) for a discharge current of 1.78 mA; (b) for a discharge current of 3.46 mA.

For the lower discharge current, the structure in the cross section was homogeneous with the carbon layers perpendicular to the axis of the fibre. For the higher discharge current, the fibre consisted of

108

Perspectives of nanoscience and nanotechnology

two different co-axial structures: an external shell and a central core. The structure of the core was similar to that of the carbon fibre obtained for a current of 1.9 mA, but the shell was porous. A comparison of these results with those obtained using the VGCF method [3] indicates that larger diameter of fibres are produced by the electrical discharge method of synthesis. Additional experiments with other types of discharges: glow discharge, pre-breakdown streamers, and breakdown streamers indicated that these discharges were ineffective as methods for synthesising carbon fibres. Using discharges of negative polarity also failed to produce carbon fibres. It was noticed during the experiments that synthesis of carbon fibre did not start before a thin layer of soot had been deposited onto the plate electrode. It can therefore be supposed that two conditions are necessary for carbon fibre synthesis in an electrical discharge: the presence of hydrocarbon vapours and a soot layer on the passive electrode. To elucidate this process, the plate electrode was covered with a soot layer, and a discharge was generated in a pure argon atmosphere. However, synthesis of carbon fibre was not produced in these conditions. The role of the soot layer in the process is not fully understood and it requires further investigations to learn more about the mechanisms of synthesising carbon fibres in an electrical-discharge plasma. Summary In these experiments, carbon fibres were synthesized in a point to plate low-current electrical discharge of positive polarity using cyclohexane with argon as the carrier gas. The discharge current was varied from 1 mA to 5.6 mA. For these currents, the length of carbon fibres produced varied in the range 0.5 mm to 12 mm and their diameters varied from about 20 µm to 120 µm. The growth rate of the carbon fibre was up to about 0.5 mm/s. The influence of the conditions used for synthesis on the properties of the synthesized carbon fibre; such as current, composition of the gas, pressure and the distance between the electrodes requires further investigation. References [1] [2] [3] [4] [5] [6] [7]

D. D. L. Chung: Carbon Fiber Composites. Butterworth Heineman MA: Newton, (1994) 13-53. J. R. Bradley, G. G. Tibbetts: Carbon Vol. 23 (1985), p. 423-433. G. G. Tibbetts, C. P. Jr Beetz: J. Phys. D: Appl. Phys.; Vol. 20 (1987), p. 292. A. Huczko: Appl. Phys. A Vol. 74 (2002), p. 617-638. M. W. Li, Z. Hu, X. Z. Wang, Q. Wu, Y. Chen, Y. L. Tian: Diam. Relat. Mat. 13 (2004), p. 111. N. Sano, M. Nobuzawa. Carbon Vol. 43 (2005), p. 2224. J. R. Brock, P. Lim: Appl. Phys. Lett.; Vol. 58 (1991), p. 1259..

Solid State Phenomena Vol. 140 (2008) pp 109-116 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.109

A Titanium-decorated Fullerene Cluster – a Molecular Dynamics Simulation A. Piątek 1, a , R. Nowak 2, b, Z. Gburski 1, c 1

Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007, Katowice, Poland 2

Nordic Hysitron Laboratory, Helsinki University of Technology, 02015 HUT, Finland

e-mail: a [email protected]; b [email protected]; c [email protected] Keywords: decorated fullerene, phase transition, MD simulation

Abstract. A small titanium-decorated fullerene cluster (C60[TiH2]6)7 was studied by MD simulation over a wide range of energy, from the solid state to the vaporization of the nanosystem. The low energy, solid state structure of the cluster was obtained as a deformed pentagonal bipyramid. Several physical characteristics: the radial distribution function, the mean square displacement, the translational velocity autocorrelation function, translational diffusion coefficient, Lindemann index, etc., were calculated for a wide range of energy in the system. Introduction Recently, it was predicted that a single TiH2 atomic group attached to carbon nanostructures, such as C60 or nanotubes, can adsorb up to three Hz hydrogen molecules [1-3]. In view of this, a detailed understanding of physical and chemical properties of Ti-decorated fullerene systems becomes a subject of great interest, as they have potential for hydrogen storage [3]. In this paper, the focus is on the dynamical properties of an extremely small nanosystem, a cluster, composed of only as few as seven C60[TiH2]6 aggregates. It was shown [4 - 6] that a very small pure fullerene cluster (C60) 7 exhibits unusual properties. There are no other reports about wide temperature range liquid phase of fullerene clusters of different sizes. Simulation details The Lennard-Jones (LJ) potential was applied to describe the interaction between carbon atoms for 12 6 each pair of C60 molecules, V (rij ) = 4ε (σ / rij ) − (σ / rij ) ,where rij is the distance between the ith

[

]

and jth carbon atoms of the pair of interacting fullerenes, ε and σ are the two standard LennardJones (LJ) potential parameters shown in Table 1. ε kB

[K ]

σ [Å]

m [10-25 kg]

C-C

28

3.4

0.199

Ti-Ti

205.4

3.8

0.795

H-H

12.5

2.81

0.016

Table 1. The carbon mass and L-J potential parameters used in simulation, taken from [7-9]

110

Perspectives of nanoscience and nanotechnology

The fullerenes C60 have been treated as rigid bodies with translational and rotational degrees of freedom. The 60 carbon atoms of fullerene form a truncated icosahedron, which consists of 12 pentagonal and 20 quasi-hexagonal faces. There are two distinct C-C bond lengths. The values used in this work, 1.37 Å and 1.448 Å were taken from quantum chemistry calculations [10]. Following previous first principle energy calculations [3], the six TiH2 groups were located on twofold axes of the pairs of hexagons forming the C60 molecule. The LJ parameters for the reciprocal titanium, hydrogen and carbon interactions are also given in Table 1. The classical equations of motion were integrated up to 2.4 ns by the Adams-Moulton predictor-corrector algorithm [11]. The integration time step was 2 fs which ensures good stability of the algorithm. The clusters were equilibrated for 10 7 integration steps. The appropriate correlation functions were averaged over 2. 4 ⋅ 10 4 time origins. The origins were separated by a time interval equal to 50 integration time steps. The calculations were carried out for a constant energy ensemble for zero total linear and angular momentum of the whole system. The ground state configuration of (C60[TiH2]6)7 was reached by Monte Carlo (MC) simulation after 5 *106 time steps. The thermodynamic limit could be a problem in case of constant temperature MD simulations for a system consisting of a very small number of particles and it is always a problem to define the temperature in extremely small clusters [12]. That explains the reason for studying the calculated physical quantities of specific interest in the same manner as Gallego et al. [4], i.e. as a function of the total energy Et of the cluster instead of temperature. Results It is known, that the ground state (global minimum of potential energy) of the solid phase of (C60)7 cluster is a pentagonal bipyramid (PBP) [13, 4]. In the case of (C60[TiH2]6)7 the lowest potential energy configuration also happened to be similar, but not exactly, like aPBP . It can be called a deformed pentagonal bipyramid (DPBP). Because of the presence of TiH2 atomic groups attached to the fullerene it is not a perfect “buckyball” and the minimum energy configuration is not a perfect PBP. The DPBP configuration was reached after 106 simulation steps in the Monte Carlo procedure (see Fig. 1).

Fig. 1. Snapshot of the minimum potential energy configuration (deformated pentagonal bipyramid) of (C60[TiH2]6)7 cluster. Having the (C60[TiH2]6)7 cluster initially in a PBP configuration, (solid state at low energy/temperature), the energy of the system was gradually increased in a stepwise manner , heating the system. Fig. 2 shows the normalized velocity autocorrelation function

Solid State Phenomena Vol. 140

111

Cv(t)=-1 where v is the translational velocity of the centre of mass of C60[TiH2]6 molecule , calculated for three energies Et – low, intermediate and high.

Fig. 2. The linear velocity autocorrelation function of C60[TiH2]6 molecule in the cluster, for three energies Et It can be seen that for the low energy condition the velocity of C60[TiH2]6 molecule exhibits a deep dip and strongly dumped pulsations, characteristic for the solid phase, it becomes fully decorrelated after 2.5 ps. The high energy Cv(t ) decays much more regularly – the behavior associated with a softer phase of matter (a liquid). The third correlation function Cv(t ) interpolates between the previous extremes. Following this observation, the velocity autocorrelation time τv = ∫ dt Cv(t ) was calculated, as a function of energy Et (see Fig. 3).

Fig. 3. The energy dependence of the linear velocity correlation time of C60[TiH2]6 molecule in the cluster The sharp increasing of τv arround (-0.12 eV/molecule) is clearly visible. Note, that closely connected with τv is the translational diffusion coefficient D ~ τv [14]. Therefore, the plot τv(Et) provides information about two regions in the cluster. Firstly, the solid phase where the almost negligible, and practically no, energy independent diffusion appears. Secondly, a phase

112

Perspectives of nanoscience and nanotechnology

having increased diffusion of the decorated fullerene, which is associated with the increased energy r r – the liquid state.The calculated angular velocity autocorrelation function Cω(t) = < ω (t)٠ ω (0)> r < ω (0)2 >-1 is presented in Fig. 4.

Fig. 4. The angular velocity autocorrelation function of C60[TiH2]6 molecule for different energies Et in solid phase of the cluster. Two types of behaviour of Cω(t) can be readily distinguished. Firstly, for Et ≤ -0.12 eV/molecule the damped pulsations of Cω(t) characteristic for hindered rotations in the solid phase [16-19]. The correlation function Cω(t) changes with the energy, the lower the energy, the deeper is the dip of Cω(t). The negative dip of Cω(t) near t = 0.3 ps (Et = -0.42 eV/molecule) vanishes at (Et = -0.12 eV/molecule). Secondly a regular relaxation for Et > -0.12 eV/molecule, which is similar to the reorientations in a liquid-like phase. To prove even stronger the appearance of phase transition which separates the solid and liquid states of (C60[TiH2]6)7, the Lindemann index δL was calculated [12]

δL =

1 2 2

(< rij > − < rij > ) 2 ∑ N ( N − 1) i < j < rij > N

2

(1)

The energy evolution of index δL is given in Fig. 5a. One can see that the value of Lindemann index δL is very low and only slightly change with energy for Et < -0.15 eV/molecule, confirming therefore the solid state phase (no translational diffusion) of the cluster in this energy region. Further increasing of energy (heating) leads towards to phase transition and the liquid phase of (C60[TiH2]6)7 [ 12, 15, 16], the indication of this is the substantial raise up of Lindemann index δL for Et > -0.15 eV/molecule .

Solid State Phenomena Vol. 140

113

5a

5b

Fig. 5. The energy Et dependence of the Lindemann index of ; a) (C60[TiH2]6 )7 and b) pure fullerene cluster (C60)7 . In Fig. 5b the index δL for pure (C60)7 cluster, calculated in [5] is presented. Note that the solid/liquid phase transition in pure (C60)7 cluster appears at much lower energy, compared to the (C60[TiH2]6)7 nanosystem. Apparently, the Ti-decorated fullerenes have greater mutual attraction than pure fullerenes. The phase transition from a solid to a softer phase in (C60[TiH2]6)7 exists at

114

Perspectives of nanoscience and nanotechnology

much higher energy, compared to the pure (C60)7 cluster. In Fig. 6 presents the calculated standard 2 2 deviation of the total energy δ Et = Et − Et

2

.

Fig. 6. The deviation of total energy Et for (C60[TiH2]6 )7 cluster. Note, that /kBT2 would be proportional to the specific heat at constant volume [14]. However, since there is no explicit use of the concept of temperature for the very small system, the authors stay with the presentation of . Here again, the spectacular increase of the energy fluctuation is observed for higher energies (Et > -0.1 eV/molecule). Naturally, the observed jumps of observables τv, δL and can be associated with the phase transition in the cluster. Next the mean square displacement of Ti-decorated fullerene was calculated for several energies, ∆r(t)=r(t)-r(0), where r is the position of the center of mass of single C60[TiH2]6 (Fig. 7).

Fig. 7. The mean square displacement of the centre of mass of the C60[TiH2]6 molecule in the cluster for the three energies The slope of the mean square displacement can be directly connected with the translational diffusion coefficient D of a molecule via Einstein formula ≅ 6Dt. From figure 7 we see

Solid State Phenomena Vol. 140

115

that translational mobility of C60[TiH2]6 significantly increases with the growth of energy. The energy dependence of D is given in Fig. 8, we see that the translational diffusion coefficient of C60[TiH2]6 significantly increases with the growth of Et .

Fig. 8. The energy dependence of the translational diffusion coefficient in the liquid phase of (C60[TiH2]6 )7 cluster Note the bend in the D(Et ) plot around Et ≈ -0.12 eV/molecule which marks the region of the phase transition. The endohedral fullerene’s centre of mass radial distribution function g(r) is presented in Fig. 9.

Fig. 9. The radial distribution function of the centre of mass of C60[TiH2]6 molecule in the cluster.

116

Perspectives of nanoscience and nanotechnology

For the low energy Et = -0.43 eV/molecule (solid phase) it shows several peaks associated with the deformed pentagonal bipyramid structure of the cluster. At the high energy Et = -0.06 eV/molecule the multitude of sharp peaks vanishes and the broadening of g(r), characteristic for the liquid phase, is observed. Conclusions The atomistically detailed MD simulations show that the solid phase of (C60[TiH2]6)7 is presented for low energies Et . Evidence is provided of the phase transition of solid/liquid expanded in the energy range -0.20 < Et < -0.05 eV/molecule and the appearance of the liquid phase in the relatively narrow energy region from Et > -0.05 eV/molecule to vaporization of the cluster. The liquid phase of (C60[TiH2]6)7 exists at much higher energies (Et > -0.05 eV/molecule), compared to the pure fullerene cluster (C60)7 (Et > -0.35 eV/molecule ). The influence of a molecular hydrogen H2 atmosphere on the properties of (C60[TiH2]6)n (n=7, 13, 25, ….) nanosystems could be the subject of future studies, directly related towards the issue of hydrogen storage. Acknowledgement The support of the Academy of Finland through consortia NAKAMA and NANOTOMO - research projects No. 124059 and 117867 respectively, is gratefully acknowledged.

References [1] T. Yildirim, J. Íñiguez and S. Ciraci, Phys. Rev. B 72 (2005), p.153403 [2] T. Yildirim and S.Ciraci, Phys. Rev. Lett. 94 (2005), p.175501 [3] E. Durgun, S. Ciraci, W. Zhou and T. Yildirim, Phys. Rev. Lett. 97 (2006), p.226102 [4] L. J. Gallego, J. Garcia-Rodeja, M. M. G. Alemany and C. Rey, Phys. Rev. Lett., 83 (1999), p.5258 [5] A. Piątek, A.Dawid, Z. Gburski, J. Phys. Condens. Matter 18 (2006), p. 8471 [6] A. Piątek, A. Dawid, Z. Gburski, J. Phys. Condens. Matter 18 (2006), p.11397 [7] H. Dominguez, A. G. Goicochea, N.Mendoza, J. Alejandre 2006 Journal of Colloid and Interface Science 297 (2006), p.370 [8] W, A, Steele, The Interaction of Gases with Solid Surfaces (New York: Pergamon) 1974 [9] A. Cheng and M. Klein, Phys. Rev. B 45 (1992), p.1889 [10] G. E. Scusseria, Chem. Phys. Lett. 176 (1991), p. 423 [11] D. C. Rappaport, The Art of Molecular Dynamic Simulation (Cambridge: Cambridge University Press) 1995 [12] P. Jena, B. K. Rao and S. K. Khanna, The Physics and Chemistry of Small Clusters, Plenum, New York, 1987 [13] E. Sawada and S. Sugano, Z. Phys, D 14 (1989), p.247 [14] J. P. Hansen and I. P. Mc Donald, Theory of Simple Liquids (London: Academic) 1986. [15] A. Dawid and Z. Gburski, Phys. Rev. A, 58 (1998), p.740 [16] A. Dawid and Z. Gburski, Phys. Rev. A, 56 (1997), p.3294 [17] A. Piątek, A. Dawid, Z. Gburski, J. Molec. Struc. 792 (2006), p.82 [18] Z. Dendzik, M Paluch, Z Gburski, J Zioło, Journal of Physics: Condensed Matter, 9 (23) (1997) L339 [19] A. Dawid, Z. Dendzik, Z. Gburski, J. Molec. Struc. 704 (1) (2004) p.173

NANOSTRUCTURES FOR MEDICINE

Solid State Phenomena Vol. 140 (2008) pp 119-126 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.119

Nanotechnology for Treating Damaged Organs Jing Lu a and Thomas J. Webster b Brown University, Division of Engineering, 184 Hope Street, Providence, RI 02917, USA e-mail : a [email protected]; b [email protected] Keywords: Stents, Nanomaterials, Vascular, NanoPatterning

Abstract. Atherosclerosis, which is caused by endothelial dysfunction, vascular inflammation, and the build-up of lipids, cholesterol, calcium, and cellular debris within the intima of the vessel wall, is one of the most important complications of health. Vascular stenting is the procedure of implanting a thin tube into the site of a narrow or blocked artery due to atherosclerosis. However, the application of vascular stents using conventional metals is limited because the implantation process will cause significant injury to the vascular wall and endothelium, which functions as a protective biocompatible barrier between the tissue and the circulating blood, resulting in neointima hyperplasia followed by the development of long-term restenosis. The objective of this in vitro study was to investigate the endothelial cell function, especially their adhesion behaviour, on highly controllable features on nanostructured surface. Considering the importance of the endothelium and its properties, highly controllable nanostructured surface features of titanium, a popular vascular stent metal, were created using E-beam evaporation to promote endothelialization and to control the direction of endothelial cells on vascular stents. Endothelial cells are naturally aligned with the blood flow in the body. In this manner, the present in vitro study provides much promise for the use of nanotechnology for improving metallic materials for vascular stent applications. Introduction Atherosclerosis, which is caused by endothelial dysfunction, vascular inflammation, and the build-up of lipids, cholesterol, calcium, and cellular debris within the intima of the vessel wall, is one of the most important complications of health. Approximately 58 million people have been affected with this disease in the U.S. [1]. Vascular stenting is the procedure of implanting a thin tube into the site of a narrow or blocked artery due to atherosclerosis. It has been proven to be superior to balloon angioplasty in most types of coronary lesions and is currently the most frequently performed percutaneous coronary intervention for the treatment of coronary artery disease. Metals, including titanium, stainless steel, Nitinol and CoCr alloys, have been widely used as vascular stents. However, the application of vascular stents using conventional metals, or those with micron grain sizes which are smooth at the nanometer level, is limited because the implantation process causes significant injury to the vascular wall and endothelium, which functions as a protective biocompatible barrier between the tissue and the circulating blood, resulting in neointima hyperplasia followed by the development of long-term restenosis. It is for these reasons that numerous studies focus on methods to accelerate the endothelialisation, or coverage by endothelial cells, of the implanted stents. The repair of the disrupted endothelium involves both the migration

120

Perspectives of nanoscience and nanotechnology

of surrounding mature endothelial cells into the injured area and the attraction of circulating immature cells called endothelial progenitor cells to the site, which then develop into endothelial-like cells. Therefore investigators are modifying the traditional metals used as vascular stents by coating them with antiproliferative agents such as silicon carbide [2], expanded-polytetrafluoroethlyene [3], tantalum [4] and hyaluronan [5] to improve the interactions between the cells and the stents. However, these coatings may be destroyed by blood flow due to the normal shear stresses applied to the vascular stents. Recently much attention has been paid to modifying vascular stent surfaces from a conventional nano-smooth topography to a random nanostructured topography which mimics the surface of natural vascular tissue. The results from such in vitro studies indicate that vascular stents with random nanostructured surface features may invoke vascular cell responses which are promising for improving vascular stent applications [6]. This paper reports on the investigation of another manner to modify stent surfaces using controllable, as opposed to random, nanostructural features. Since endothelial cells align in vascular tissue, a highly controllable nanosurface on metals may be desirable. The objective of the investigation was, therefore, to investigate endothelial cell functions on highly controllable nanostructured surface features with emphasis on their adhesion. Materials and Methods Preparation of Nanopatterned Surface Features with E-beam Evaporation. A micropattern of nanostructured surface features on titanium (Ti) was produced by E-beam evaporation. The process was as follows: initially, a thin layer of 99% Ti was deposited on a Si wafer by E-beam evaporation to make ultra-smooth Ti substrates while Au-coated grids with grooves 30 µm wide (SPI Supplies) were placed on the substrates. Another Ti layer was then deposited into the grooves by E-beam evaporation. The evaporation process took place above the melting point of pure Ti at a deposition rate greater than 0.5 nm per second. After that, the Au grids were removed and the substrates were rinsed. Thus micropatterns of Ti nanostructured surface features were created [7] (Figs. 1 and 2).

Fig. 1. Schematic of the method of producing the Ti nanopatterned surfaces by E-beam evaporation.

Solid State Phenomena Vol. 140

121

Figure 2: Schematic diagram of E-beam evaporation (adapted and redrawn with the permission of www.khlab.msl.titech.ac.jp). Control Materials. Random nanostructured surface features in the form of pressed discs of diameter 12mm and 0.5mm thick were produced using a simple uniaxial, single-ended compacting hydraulic press (Carver Inc.) at room temperature. Compacts with random nanostructured features were used as controls. 99.38% commercially pure Ti nanopowder (purchased from Reade International Inc.) were loaded into a hardened tool steel die and were pressed at 10 GPa pressure to obtain compacts according to standard techniques [8]. Etched borosilicate glass coverslips (Fisher brand) were also used as control materials. Before use, they were degreased in acetone and ethanol and etched in 1 M NaOH for 1 hour. All substrates were rinsed with acetone for 15 minutes, methanol for 15 minutes and finally in deionized water for a further 15 minutes. They were then sterilized by autoclave for 30 minutes before cell culture experiments as described below. Endothelial Cell Culture. Rat aortic endothelial cells (RAEC) were purchased from VEC Technologies (Greenbush, NY, USA) and were cultured in MCDB-131 Complete Medium (VEC Technologies) using 0.2% gelatin coated Petri dishes under standard cell culture conditions (a humidified, 5% CO2/95% air environment at 37ºC). The medium was replenished every second day. Cell Adhesion and Proliferation. For cell adhesion tests, the RAEC were seeded with 4500 cells/cm2 in MCDB-131 Complete Medium onto the nanopatterned surfaces, random nanostructured surfaces, and etched glass coverslips. After 4 h, non-adherent cells were rinsed away using three applications of phosphate buffered saline solution while the viability of the adherent cells was determined using a LIVE/DEAD Viability/Cytotoxicity Kit for mammalian cells according to the manufacturer’s instructions (Molecular Probes, Chicago, IL, USA). Using a fluorescence microscope, the live and dead cells were determined quantitatively using 530 nm wavelength for calcein AM and 560 nm wavelength for Ethidium homodimer-1. They were also evaluated qualitatively by scanning electron microscopy. All experiments were run in triplicate.

122

Perspectives of nanoscience and nanotechnology

Results Substrate Characterization. Scanning electron microscopy (SEM) was used to examine the surfaces of the substrates and the images are shown in Figure. 3.

(a)

(b)

(c) (d) Figure 3. (a), (b) Scanning electron microscope (SEM) images of nanopatterned surface features (a) bar = 20 µm; (b) bar = 2 µm; and (c), (d) random nanostructured surface features ((c) bar = 10 µm; (d) bar = 200 nm). The SEM images revealed the micropatterns of the nanostructured surface features created by E-beam evaporation. The spacing and width of the controllable nanostructured surfaces were about 30 µm as expected from the Au grid spacing. The SEM images of the Ti compacts (prepared by cold pressing) showed numerous nanometer Ti particles on the substrates, which created random nanostructured surface features. RAEC Adhesion Behavior. The results of cell adhesion tests indicated that the RAEC density on the nanopatterned Ti surface features was greater than that on the random nanostructured Ti surface features as shown in Figure 4. Moreover, cell viability on the nanopatterned substrates also greatly improved.

Solid State Phenomena Vol. 140

123

Figure 4: Increased rat aortic endothelial cell (REAC) adhesion on the nanopatterned Ti compared to random nanostructured Ti features. The samples labelled as polished Ti are the same as smooth Ti. The samples labelled as line width = 30 µm are those that were nanopatterned. Data = +/- STD; N = 3; * p < 0.1 (compared to polished Ti and random nanostructured Ti). RAEC Morphology. Fig. 5 shows the morphological differences between endothelial cells on the nanopatterned and random nanostructured substrates, indicating that the nanopatterns were able to significantly enhance cell spreading and control the direction of the cells. However, on the random nanostructured surfaces there was no directional preference of the cells.

Nanopatterned area

(a) (b) Figure 5. SEM image of cell morphology (a) nanopatterned surface features (b) random nanostructured surface Discussion This study demonstrated improved cell adhesion and cell orientation occurred on micropatterns of nano features on the Ti surface when compared to a Ti surface with random nanostructured surface

124

Perspectives of nanoscience and nanotechnology

features. The technique of creating nanopatterns by E-beam evaporation offers a technically sound and economically promising way to easily coat existing vascular stents with nanopatterns. However, the method used to fabricate the random structured nanometer surface features cannot be used as a coating process since it compacted the Ti nanoparticles together at high pressure using a hydraulic press and the particles may delaminate when subjected to high shear stress. It is also important to consider the mechanism by which greater vascular cell adhesion occurred on the nanopatterned surface features. Palmaz et al. established the influence of micron topographical patterns on the endothelialization of metallic stent surfaces, hypothesizing that such micron patterns may reduce the time for endothelialization of stents and, thus, the risk of restenosis [9]. They attributed this increase to a physical effect, in which the changes in the alignment and shape of cells are related to a simple cell conformation process. Some other studies have also demonstrated that cell shape and orientation are related to cell gene expression. Changes in cell shape may affect much of a cell’s metabolism. In addition, as stated earlier, the nanopatterned surface features resemble the interface between the circulating blood in the lumen and the rest of the vessel wall better than the random nanostructured surface features. However, there is some contradiction concerning the influence of surface patterns on cell alignment. While this has been largely untested for nanomaterials, it has been extensively studied for conventional materials. For example, some studies indicated that the degree of cell alignment increases with groove size [9], but other studies showed that the larger the concentration of ridges on the grooved surface, the greater the effect on cell morphology [10]. It follows that the influence of the dimension of nanopatterns is another important subject needing to be studied further on these novel substrates. In addition, as demonstrated here, the nanopatterns make a contribution to cell shape and orientation, which are associated with cellular responses to implanted materials. These E-beam produced surfaces can be applied to other implantable materials such as those used for improving bone prostheses where the preferred orientation of calcium containing mineral and collagen is vital. Faster formation of bone-like cell multilayers on microgrooves 150 µm wide than on smooth surfaces, have already been observed [11]. With the help of E-beam evaporation techniques, Ti-based bone prostheses can easily be coated with micropatterns of nanomaterials to enhance their performance and possibly result in a generation of anisotropic bone juxtaposed to an implant surface. Conclusions Considering the importance of the endothelium and its properties, highly controllable nanostructured surface features were created to promote endothelialization and to control the morphology of the endothelial cells on vascular stents. The quantitative analysis of this study showed enhanced endothelial cell adhesion on these nanopatterned surfaces compared with those on random nanostructured surfaces. The topography of endothelial cells was observed using scanning electron microscopy, indicating that besides being well spread on the nanopatterned surface, the cells aligned along the nanolines, or grooves. In each sub-region, they were parallel with each other, similar to their appearance on natural vessel walls. This could help endothelial progenitor cell proliferation and differentiation to establish new endothelium for vascular stent success.

Solid State Phenomena Vol. 140

125

Acknowledgements The authors would like to thank the Coulter Foundation for funding. References 1. 2. 3. 4. 5. 6.

National Institute of Health, 2005 P. Schuler, D. Assefa, J. Ylanne, et al : Cell Commun. Adhes. Vol. 10(1) (2003), p.17-26. M. Cejuna, R. Virmani, R. Jones, et al. : J. Vasc. Interv. Radiol. Vol. 13 (2003), p. 823-30. J.Y. Chen, Y.X. Leng, X.B. Tian, et al.: Biomaterials Vol. 23(12) (2002), p. 2545-52. W.G. Pitt, R.N. Morris, M.L. Mason: J. Biomed. Mater. Res. 2004; 68(1):95-106. S. Choudhary, M. Berhe, K.M. Haberstroh and T.J. Webster: International Journal of Nanomedicine Vol. 1(1) (2006), p. 41-49. 7. D. Khang, M. Sato, et al. : International Journal of Nanomedicine Vol. 1(1) (2006), p. 65-72. 8. T.J. Webster, J.U. Ejiofor: Biomaterials Vol. 25 (19) (2004), p. 4731-4739. 9. J.C. Palmaz, A. Benson, A. Eugene: J. Vasc. Interv. Radiol. Vol. 10(4) (1999), p. 439-44. 10. P. Clark, P. Connolly, A.S.G. Curtis, J.A.T. Dow, C.D.W. Wilkinson: J. Cell Sci. Vol. 99 (1991), p. 73-7. 11. A. Khakbaznejad, B. Chehroudi, D.M. Brunette: J. Biomed. Mater. Res. Vol. 70(2) (2004), p. 206-18.

Solid State Phenomena Vol. 140 (2008) pp 127-132 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.127

Electrospray Nanocoating of Microfibres. A. Jaworek1, a, A. Krupa1, b, A.T. Sobczyk1, c, M. Lackowski1, d, T. Czech1, e, S. Ramakrishna 2, f, S. Sundarrajan 2, g and D. Pliszka 2, h 1

2

Institute of Fluid Flow Machinery, Polish Academy of Sciences, Fiszera 14, 80-952 Gdansk, Poland

National University of Singapore, Nanoscience & Nanotechnology Initiative, Faculty of Engineering, Block E3, 05-29, 2 Engineering Drive 3, Singapore 117576

e-mail: a [email protected]; b [email protected]; c [email protected]; d [email protected]; e [email protected]; f [email protected]; g [email protected]; h [email protected] Keywords: nanocoating, electrospray, microfibers, nanotechnology

Abstract. The paper presents experimental results of electrospray deposition of nanopowder onto microfibers. The process is designed to form fibrous filters with an enhanced collection efficiency in the submicron range by covering the fabric with a catalytic material. Polyamide fibres were coated with Al2O3, ZnO, MgO, or TiO2 nanoparticles. The structures obtained were porous at the nanometer scale which increased the total surface area of the catalyst. Introduction Electrospraying is a method of liquid atomisation achieved by subjecting a liquid to electrical stress at a capillary nozzle. The liquid meniscus maintained at high electric potential elongates into a jet and disrupts into fine droplets. The droplets obtained by this method are charged and can be of submicron size. The size of the droplets and their production rate can be readily controlled by adjustment of the liquid flow rate and the voltage applied to the nozzle. Since the last decade, there is increasing interest in the application of electrospray processes in nanotechnology [1-4]. It was discovered by many authors that the deposition efficiency of a charged spray onto an object is significantly higher than for an uncharged spray. In the case of nanotechnology applications, the droplets can be directly deposited onto a substrate which facilitates surface coating or direct writing in a submicron scale. The process of jet formation and its disintegration into droplets is known in the literature as the mode of spraying [5, 6]. The spraying modes can be categorized into two groups: 1. Dripping modes, in which only fragments of liquid are ejected from the capillary outlet by the deformation and detaching of the liquid meniscus. These fragments can be formed as large regular drops (dripping mode), fine droplets (microdripping mode) or a single or multiple spindles (spindle and multispindle modes), or irregular fragments of liquids. 2. Jetting modes by which the meniscus elongates into a long fine jet, The jet can be smooth and stable (cone-jet mode) or can move in any regular way: rotate around the capillary axis (precession mode) or oscillate in a plane (oscillating mode). Sometimes a few jets on the circumference of the capillary can be observed (multijet mode). The case when the jet branches is known as a ramified jet. In this paper the experimental results of deposition of nanopowder on microfibers using electrospraying are presented. The aim of the process is the formation of fibrous filters with enhanced collection efficiency in the submicron range by covering a tight unwoven fabric with a catalytic material. The deposited metal-oxides used as catalyst will facilititate the production of materials for protective technologies such as masks, garments, or respirators. The currently used

128

Perspectives of nanoscience and nanotechnology

materials are not very efficient in giving protection against small-sized biological agents or chemical compounds, and they are frequently heavy. The composite materials made from nanoparticles deposited onto nanofibers and sandwiched between two layers of fabric will provide light weight, low cost, and breathable structures with improved protection, duration and performance. Experimental The experiments were carried out in a system consisting of a stainless-steel capillary nozzle and an aluminium plate (a gutter) (Fig. 1).

Fig. 1. (a) schematic of experimental stand, (b) photograph of electrospray system. The dimensions of the capillary were 0.45 mm o.d. and 0.25 mm i.d. and a length of 10 mm. The distance between the nozzle tip and the plate was 25 mm. The diameter of the aluminium plate was 150 mm. The substrate was a commercial fabric made of woven 30 µm diameter polyamide fibres stretched over a small 1 mm thick metal ring of o.d 15 mm and i.d 10 mm. The nozzle was connected to a high voltage supply, SPELMANN SL600W/30kV/PN, of positive polarity with the plate electrode earthed. An electric heater under the plate facilitated the liquid’s evaporation. The spray plumes were recorded using a SONY DSC-F585V CCD camera. Metal oxides suspensions of Al2O3, ZnO, MgO, and TiO2 in methanol supplied from Chempur (Poland) were used in the experiments. The liquid was supplied from a syringe mounted above the capillary nozzle. MgO particles of pure grade 40.3 g/mol were purchased from POCH Gliwice (Poland); TiO2 particles of molecular weight of 79.90 g/mol were supplied by Eurochem BGD (Poland); Al2O3 particles of 99.9% purity, and ZnO particles of 99.0% purity were purchased from Alfa Aesar. The suspensions were electrosprayed for about 15 minutes at an estimated flow rate of 20 ml/h. The structures obtained were examined with a Zeiss EVO-40 scanning electron microscope.

Results and discussion In the experiments, the suspended particles in methanol were electrosprayed and deposited onto 30 µm diameter polyamide fibres. The cone-jet and multijet spray modes are shown in Figs. 2a and b. The smaller droplets in the multijet mode produced smaller particle agglomerates. The results presented were obtained for the multijet mode. Low particle concentration and the ease of solvent

Solid State Phenomena Vol. 140

129

evaporation enabled a uniform tight layer to be deposited on the fibres. The layer morphology varies depending on the structure of deposited particles.

Fig. 2. (a) cone-jet mode of electrospraying (voltage 5.1 kV), (b) multijet mode of electrospraying (voltage 7.8 kV). Liquid: MgO suspension in methanol. Electrode distance: 20 mm, exposure: 3 µs. Figs. 3 to 6 show SEM micrographs of the fragments of the layers consisting of Al2O3, ZnO, MgO, and TiO2 particles deposited onto the polyamide fibres.

Fig. 3. SEM images of Al2O3 layer deposited by electrospraying onto a 30 µm dia polyamide fibre. (Al2O3 suspension in methanol, voltage=24 kVdc). Left: a fragment of the coated fibre. Right: the spindle-like structure of the deposit The layers are formed despite the dielectric nature of the fibre. This is due to a thin film being formed on the fibre by the methanol-metal-oxide suspension before complete evaporation of the solvent. The film is sufficiently conducting for the electrical charges to leak to earth. Because the electric field lines terminate on the rear of the fibre it is almost evenly covered over the whole surface. It was observed that each material forms a different porous structure on the fiber after evaporation of the solvent. The Al2O3 layer shown in Fig. 3 was uniform and composed of spindlelike particles a few µm long and 1 µm in diameter. The ZnO deposit was irregular, and composed of ZnO flakes about 1-2 µm across as show in Fig 4. The flakes were deposited on the fibre in island-like groups.

130

Perspectives of nanoscience and nanotechnology

Fig. 4. SEM images of ZnO layer deposited by electrospraying on a 30 µm dia polyamide fibre. (ZnO suspension in methanol, voltage=25 kVdc). Left: a fragment of the coated fibre. Right: the flaky structure of the deposit. The MgO particles which are shown in Fig. 5 formed a tight deposit on the fibre which was composed of individual particles having a size of about 100-200 nm.

Fig. 5. SEM images of MgO layer deposited by electrospraying on a 30 µm dia polyamide fibre. (MgO suspension in methanol, voltage=5.1 kVdc). Left: a fragment of the coated fibre. Right: the crystalline structure of the deposit. The TiO2 particles deposited onto the polyamide fibres formed a thin and irregular deposit as given in Fig. 6. The particles forming the layer were not strongly adherent to the polyamide and the layer was not uniform. Only small agglomerates of particles with a size of about 100-200 nm can be observed on the substrate

Fig.6. SEM images of TiO2 layer deposited by electrospraying on a 30 µm dia polyamide fibre. (TiO2 suspension in methanol, voltage=15 kVdc). Left: a fragment of the coated fibre. Right: the crystalline structure of the deposit.

Solid State Phenomena Vol. 140

131

Formerly, electrospraying was used for covering various materials, usually metal or Si substrates, with thin layers. Usually, the deposit is formed from a precursor. Only a few papers have presented the results of deposition of layers directly formed with metal oxides. These papers are summarized in Table 1. Layer // substrate alumina (Al2O3) // Al silica (SiO2) // glass zirconia (ZrO2) // polymer sponge; zirconia (ZrO2) // zirconia; zirconia (ZrO2) // silicone release paper zirconia + alumina composite (Al2O3+ZrO2) // quartz titania (TiO2) // FTO glass

Particles size (sinter temperature) 100 nm (100-250oC) 20 nm 5 nm 410 nm (1450oC) 410 nm (1500oC) 200 nm

Solvent ethanol or ethanol + butyl carbitol ethylene glycol ethylene glycol ethanol + 0.5% dispersant ethanol + 0.5% dispersant butyl acetate + ethanol

Flow rate (spray time) 0.8 ml/h or 1.5 ml/h (60 min) 36 ml/h

References

22 ml/h 3.3 ml/h (2 h)

[8] [14]

0.36 ml/h (153 s for 100 layers) 0.6-45 ml/h (0.2 g/h)

[10]

[7] [9]

[11]

500 nm (Al2O3), 400 nm (ZrO2) (1200oC)

glycerol (for 0.25-250 ml/h alumina), olive oil (for zirconia) + 1 wt% dispersant

[12]

500 nm (TiO2 nanorods)

ethanol

[13]

1 ml/h

Table 1. Metal oxides electrosprayed from a suspension as submicron layers. The structure of the Al2O3 layer on an Al substrate obtained by Chen et al. [7] was composed of individual particles of the sol of about 100 nm. The morphology of SiO2 particles deposited on a glass substrate by Jayasinghe and co-workers [8,9] was grainy. The grains of a size of hundreds of nm were built from the sol particles 5 or 20 nm. Wang et al. [10] and Teng et al. [11] produced zirconia structures which were formed from irregularly arranged particles. A composite material consisting of Al2O3 and ZrO2 particles was produced by Balasubramanian et al. [12]. The TiO2 layer produced by Fujihara et al. [13] was composed of nanorods about 500 nm long. Other layers were produced by the use of various kinds of precursors [14]. Summary The paper provides experimental results of electrospray deposition of nanoparticles of metal-oxide on a polyamide fibre. The structures are porous on the nanometer scale which increases the total surface area of the catalyst. The reason for choosing electrospraying as a method of deposition is that an electrospray can operate in atmospheric conditions and deposit uniform micro- and nanothin films on large areas with an easily controlled deposition rate and film thickness by controlling he voltage and liquid flow rate. Electrospraying is a single-step, low-energy, and low-cost material processing technology, which can deliver products possessing unique properties. The substrate is not damaged after the spraying process. Optimization of the processing conditions will result in a low number of voids, flaws and cracks in the coating and give a sufficiently good homogeneity of the layer.

132

Perspectives of nanoscience and nanotechnology

Acknowledgements The work was co-sponsored by Polish Ministry of Science and Higher Education Project Grant No. 83/SIN/2006/02 and A*STAR Project No. 062 120 0017, within the Joint Singapore-Poland Science &Technology Co-Operation programme "Fabrication of novel nanocomposite filter membranes for understanding basic principles and their advanced technology application". References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14]

S. Ramakrishna, K. Fujihara, W.E. Teo, T.Ch. Lim, Z. Ma: An introduction to electrospinning and nanofibers (World Scientific, Singapore 2005) A. Jaworek: J. Mater. Sci. Vol. 42 No.1 (2007a), p. 266 A. Jaworek: Powder Technol. Vol. 176 No.1 (2007b), p. 18 W.E. Teo, S. Ramakrishna: Nanotechnology Vol. 17 (2006), p. R89 A. Jaworek, A. Krupa: J. Aerosol Sci. Vol. 30 (1999a), p. 873 A. Jaworek, A. Krupa: Exp. Fluids Vol. 27 (1999b), p. 43 C.H.Chen, M.H.J. Emond, E.M. Kelder, B. Meester: J. Schoonman, J. Aerosol Sci. Vol. 30 No.7 (1999), p. 959 S.N. Jayasinghe: Physica E Vol. 33 (2006), p. 398 S.N. Jayasinghe, M.J. Edirisinghe, D.Z. Wang: Nanotechnology Vol. 15 (2004), p. 1519 D.Z. Wang, M.J. Edirisinghe, S.N. Jayasinghe: J. Am. Ceram. Soc. Vol. 89 No.5 (2006), p. 1727 W.D Teng., Z.A. Huneiti, W. Machowski, J.R.G. Evans, M.J. Edirisinghe, W. Balachandran: J. Mater. Sci. Lett. Vol. 16 (1997), p.1017 K. Balasubramanian, S. N Jayasinghe., M. J. Edirisinghe: Int. J. Appl. Ceram. Technol. Vol. 3 No.1 (2006), p. 55 K. Fujihara, A. Kumar, R. Jose, S. Ramakrishna, S. Uchida: Nanotechnology Vol. 18 (2007), paper No. 365709 Q. Z. Chen, A. R. Boccaccini, H.B. Zhang, D.Z. Wang, M.J. Edirisinghe: J. Am. Ceram. Soc. Vol. 89 No.5 (2006), p. 1534

Solid State Phenomena Vol. 140 (2008) pp 133-140 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.133

Nanomaterials in Dental Applications Małgorzata Lewandowska a, Joanna Siejka-Kulczyk b, Mariusz Andrzejczuk c and Krzysztof J. Kurzydłowski d Warsaw University of Technology, Faculty of the Material Science and Engineering, Woloska 141, 02–507 Warsaw, Poland e-mail: a [email protected]; b [email protected]; [email protected]; d [email protected] Keywords: dental fillings, ceramic–polymer composites, nanocomposites, dental ceramic, zirconia

Abstract. Currently, nanopowders and nanocomposites reinforced with nanofillers are one of the most rapidly developing groups of materials possessing excellent prospects for a wide range of industrial and medical applications. This paper presents several examples of the nanomaterials developed in the Faculty of Materials Science and Engineering of Warsaw University of Technology which can have dental applications. Ceramic-polymer composites are the most popular materials for dental fillings. The influence of the nanofiller additions on the relevant properties of ceramic-polymer dental composites are discussed in the paper. The other group of nanomaterials applicable as dental materials are based on nanostructured yttrium stabilized zirconium oxide ceramic. This ceramic is widely used for the fabrication of crowns, bridges, inlays and other dental elements for which high strength, durability and esthetical appearance is required. The effect of the nano-grain size of the ceramic powder on the sintering parameters, microstructure and properties of the zirconia is discussed. Ceramic – polymer composites used as dental fillings Ceramic–polymer composites have the great advantage, compared to traditional dental materials, that they are aesthetically very pleasing as shown in Figure 1. Increasingly, amalgam, with its dark colour and the potential toxicity of mercury, is being replaced by the composite material. The conventional composites are reinforced with ceramic powders, having an average particle size ranging from a half to tens of micrometers. Although there are unquestionable advantages of composite materials, their properties, such as polymerization shrinkage and mechanical parameters, still need improving.

a b Fig. 1. (a) Composite and (b) amalgam fillings.

134

Perspectives of nanoscience and nanotechnology

One possibility for improving the properties of dental composites is to introduce a nanofiller. It may improve the mechanical, tribological and physico–chemical properties, for example by reducing the shrinkage during polymerization [1]. Nanofillers, of which the most popular is highly dispersed silica, can be produced by flame hydrolysis. One of the major problems related to the use of nanofillers is the nonhomogenous dispersion in the polymer matrix caused by powder agglomeration and inadequate wetting by the organic phase. In the reported study, commercial nanosilica filler, Aerosil R709 produced by the Degussa Company, was used. The nanosilica filler, which consisted of spherical particles of average size 40nm, was supplied after surface treatment with silane. The density of nanosilica was 2.2 g/cm3. As a microfiller, glass particles with an average size of 5 µm, elaborated and produced by the Glass and Ceramic Institute in Warsaw, was used. The chemical composition of the microfiller in wt % was 25.39 BaO, 22.11 SiO2, 18.76 Al2O3, 17.7 SrO, 9.07 F, 5.22 P2O5, 2.28 Na2O. The density of the microfiller was 3.0335 g/cm3. T he microfiller was also silanized for this experiment using 1.5 wt.% of silane. The composites with various filler contents (shown in Table 1) were prepared by manual mixing of the micro- and nanofiller with the polymer matrix. The composition of the polymer matrix was as follows: − Bis–GMA (2,2-bis-[4-(2-hydroxy-3-methacryloxypropoxy)phenyl]propane) - 58, 81 − TEGDMA (triethylene glycol dimethacrylate) – 40,49 − Photoinitiator (camphorquinone)– 0,16 − Activator 2 – (diethyloamino)ethyl methacrylate– 0,49 − Inhibitor (butylated hydroxytoluene, BHT) – 0,05 The composites were cured by exposing them to blue lights for a period of 160 seconds. Composite Sample Designation Microfiller vol. fraction [%] Nanosilica vol. fraction [%]

1 40 5

2 40 10

3 40 15

4 40 20

5 55 5

6 50 10

Table 1. The filler contents of the materials investigated Fig. 2 shows the microstructures of composites with various amounts of filler. It can be seen that the homogeneity of the composite’s microstructure depends on the content of ceramic phase. For low amounts of filler, Samples 1 and 2, agglomeration of the nanosilica can be observed, as shown in Fig. 2.1 and 2.2. Samples 3 – 5 are more homogenous (Fig. 2.3 – 2.5). Sample 6 again seems to be non – homogenous (Fig. 2.6).

Solid State Phenomena Vol. 140

1

2

3

4

5

6

135

Fig. 2. Microstructure of dental composites with various filler contents. The degree of microstructural homogeneity has a significant influence of the mechanical properties of composites. The flexural strength (Rf) using the three – point bend tests was used as a measure of the mechanical properties. Testpieces 25x2x2 mm were subjected to bend tests at a strain rate of 0.75 mm/min in accordance with ISO Standard (ISO – 4049). The results are shown in Figure 3. 120 3

4

a

1

60

5

100

2

80

120

Rf [MPa]

Rf [MPa]

100

40

4

6

b

80 60 40 20

20

0

0 40

45

50

55

volume fraction of filler [%]

60

65

0

5

10

15

20

25

volume fraction of nanosilica [%]

Fig.3. Flexural Strength, Rf, determined by 3 point bending tests showing: a. Influence of total filler volume fraction, b. nanosilica volume fraction.

136

Perspectives of nanoscience and nanotechnology

The flexural strength is affected by both the total filler volume fraction and the nanosilica content (Fig. 3). It rapidly increases as filler volume fraction increases to a filler volume of 55 % of filler and 15 vol. % of nanosilica (sample 3). With greater amounts of filler, Rf decreases, which may be due to the agglomeration of the nanoparticles. It should be noted that composites designated 3, 4 and 5 fulfilled ISO standard 4049 in respect of flexural strength values. Another factor influencing the mechanical properties of composites is the bonding between the particles and the matrix. Fillers in dental composites are typically silanated to improve their bonding to the organic matrix and to increase their “service life” [2, 3]. In addition, silanes improve the dispersion of the fillers in the matrix monomers [4]. Composites with silanated fillers exhibit superior values of Rf as shown in Fig. 4. [5], wear resistance and a higher resistance to water sorption compared to composites containing non – silanated fillers [6].

Fig. 4. Influence of nanosilica silanization on flexural strength (Rf) of composites [5] The presence of nanofillers can also influence the wear behaviour of dental composites. The loss of macrofillers in dental service results in surface roughness. The polymer matrix then wears more rapidly than the fillers which may protrude and even break (Fig.5). As a result, the composite and antagonistic teeth wear more rapidly. In addition, surface roughness promotes plaque adhesion [7]. It has already been shown that nanosilica improves wear resistance of dental composites [9].

Fig. 5. Wear scheme of the composite [8] Low water sorption is one of the requirements for dental composites. In order to estimate it, the specimens were dried to constant mass (m1), and than exposed to distilled water at 37ºC for one week. After exposure they were dried on blotting paper and weighed (m2). They were then dried to a constant mass and reweighed (m3).

Solid State Phenomena Vol. 140

137

The water sorption was calculated from following equation: Wsp = (m2 – m3)/ V [µg/ mm3]

(1)`

where: m2 - mass of dried samples after exposure; m3 - specimens mass after exposure and drying to constant mass; V – volume of specimen The results of the water sorption tests are shown in Fig. 6. These show that water sorption is strongly influenced by the filler volume fraction but only slightly by the nanosilica content.

Wsp [micrograms/ mm 3]

50

a

40

1 2

4

30

3

20 10 0 40

45

50

55

60

65

vololume fraction of filler [%]

Wsp [micrograms/ mm 3]

.

b

50 40 30

5

6

4

10

20

20 10 0 0

30

volume fraction of nanofiller [%]

Fig. 6. The influence of: a. total filler volume fraction b. nanosilica volume fraction on the water sorption (Wsp) In summary one can conclude that nanosized silica can be successfully introduced as a filler in dental composites. Nanosilica has a beneficial effect on the properties of the composites, improving the mechanical strength and wear resistance. Zirconia-based dental ceramic Yttria doped tetragonal zirconia polycrystals (Y-TZP) is currently used as a ceramic biomaterial in dental applications for the production of crowns and bridges [10]. Aesthetical aspects, biocompatibility and good mechanical properties, such as high strength and high fracture toughness, have increased zirconia’s share of the market for dental materials [11]. The superior properties of the Y-TZP ceramic result from their nanosized grains. The use of very fine powders with a crystallite size smaller than 40 nm enables a fine grained, homogenous material with good mechanical properties to be obtained using a lower sintering temperature. The use of fine grained powders also allows production of a material, which can possess novel properties. To take advantages of a nanostructured material, it is essential to precisely control the sintering process during all ceramic zirconia restoration preparations. The microstructures of fine grained Y-TZP ceramic are shown in Figure 7. 1225oC

1325oC

1450oC

Fig. 7. SEM micrographs of sintered Y-TZP powder compacts.

138

Perspectives of nanoscience and nanotechnology

The images reveal that microstructure is reasonably homogenous with few visible fine pores. The measured average equivalent diameter of grains increased with increased sintering temperature from 0.12 µm at a sintering temperature of 1225oC to 0.27 µm for the samples sintered at 1450oC (Fig. 8). It should be noted that sintering above 1400oC produced a fully-dense ceramic.

equivalent diameter [um]

0,40

0,30

0,20

0,10

0,00 1200

1250

1300 1350 1400 1450 sintering temperature [oC]

1500

Fig. 8. Temperature-dependent grain size for the zirconia ceramics [12]. It is known that the strength of materials strongly depends on their grain size. The increase in strength with the increase of sintering temperature results in densification and a reduction of porosity. In the study it was found that pellets sintered at 1400oC had the highest flexural strength shown in Fig. 9. 1100

strength [MPa]

1000 900 800 700 600 1200

1300

1400

1500

o

sintering temperature [ C]

Fig. 9. Temperature-dependent strength for the zirconia ceramics. A higher sintering temperature, 1450oC), caused grain growth and, according to the Hall-Petch relationship, a consequent decrease in the mechanical strength [13, 14]. The grain size of Y-TZP influences also its degradation behaviour. Tetragonal zirconia phase is inherently metastable at room temperature and may undergo stress-induced transformation to the more stable monoclinic phase. Such a transformation can drastically decrease some mechanical properties and result in degradation of the ceramic. The amount of the monoclinic phase increases with the aging time when the grain size exceeds a critical value. Watanabe et al. [15] estimated the critical grain size for tetragonal phase retention at 0.3 µm for zirconia containing 3 mol.% of Y2O3. Current ISO standards require that the intercept length of Y-TZP grains must be lower than 0.6 µm. It is suggested, however, that an intercept length of 0.4 µm with a standard deviation lower than 0.2 µm is more appropriate [11].

Solid State Phenomena Vol. 140

139

Present investigations also confirmed the grain size dependent phase stability. Zirconia materials sintered at 1400ºC and higher temperatures (with a grain size higher than 0.2 µm) experienced the tetragonal to monoclinic transformation on autoclaving. According to these results, specimens of zirconia with equivalent grain diameters less than 0.2 µm exhibited full stability under the hydrothermal conditions used in the test [13, 16]. Conclusion It was shown that nanosized particles may successfully be used for the fabrication of dental materials. They offer the possibility to improve the mechanical strength and durability. This applies to both ceramic – polymer composites and Y – TZP ceramic. Acknowledgments This study was financed by Polish Ministry of Science and High Education (Grant No. R08 027 01). References: [1] N. Moszner, U. Salz: Prog. Polym. Sci. 2001, 26, 568 – 569 [2] R. L. Kaas, J.L. Kardos: Polym. Eng. Sci. 1971, 11, 11 – 18 [3] J. Luo, R. Seghi, J. Lannutti: Materials Science and Engineering C 1977, 5, 15 - 22 [4] S. Debnath, S. L. Wunder, J. I. McCool, G. R. Baran: Dental Materials 2003, 19, 441 – 448 [5] K. Cygan, M. Lewandowska: Proceedings of XXXIV School of Materials Science, Cracow, Poland [6] J. W. Wang, H. Ploehn: J Appl Polym Sci. 1996, 59, 345–57 [7] S. Klapdohr, N. Moszner, Monatshefte für Chemie 2005, 136, 27 – 33 [8] E. Jodkowska: „Materiały złoŜone i pośrednie systemy wiąŜące w odtwarzaniu ubytków w zębach bocznych” Tour Press International, Warsaw 1993 (in Polish) [9] J. Romaniuk, M. Lewandowska, K. J. Kurzydłowski, J. R. Dąbrowski: Engineering of Biomaterials 2005, 47 – 53, 178 - 181 [10] H. Luthy, F. Filser, O. Loeffel, M. Schumacher, L.J. Gauckler, C.H.F. Hammerle: Dental Materials 2005, 21, 930-937 [11] J. Chevalier, Biomaterials 2006, 27, 535-543 [12] T. Kosmac, M. Andrzejczuk, K.J.Kurzydłowski: Ceramic engineering and science proceedings 2007, vol. 27; ISSU 2, pages 83-92 [13] E.O. Hall, Proc. Phys. Soc. (Lond.) 1951, B64, 747–753 [14] N.J. Petch, J. Iron Steel Inst. 1953, 174, 25–28 [15] M. Watanabe, S. Iio, I. Fukuura “Ageing behaviour of Y-TZP” In Advances in Ceramics, Vol. 12, Science and Technology of Zirconia II, ed. N. Claussen, M. Ruhle & A. H. Heuer. The American Ceramic Society, Inc., Columbus, Ohio, 1984, pp.391-8 [16] M. Andrzejczuk, M. Lewandowska, T. Kosmac, K. J. Kurzydłowski: Advances in Materials Science 2007, vol. 7, number 2 (12) 76-85

Solid State Phenomena Vol. 140 (2008) pp 141-146 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.141

The Influence of Graphene Sheet on the Dynamics of Cholesterol Molecules in the Lodgment Located Near a Transmembrane Protein – MD study P. Raczynski a, A. Dawid b and Z. Gburski ,c Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007 Katowice, Poland e-mail:

a

[email protected]; b [email protected]; c [email protected]

Keywords: transmembrane protein, graphene sheet, cholesterol, molecular dynamics

Abstract. Molecular dynamics (MD) simulations have been made for a cluster of cholesterols localized near the transmembrane protein at the physiological temperature of 310 K. It was observed that the cholesterol molecules form a lodgment on the surface of protein. Additional studies were made of the influence of graphene sheet on several physical observables of cholesterol molecules including: the radial distribution function, the mean square displacement, diffusion coefficient and the linear and angular velocity autocorrelation functions. Introduction Cholesterol is an important component of the cells of mammals. It is found in higher concentrations in tissues which either, produce more, or have more, densely-packed membranes, for example, the liver, the spinal cord and the brain. Recent research shows that cholesterol has an important role for the brain synapses and in the immune system, including providing protection against cancer [1 – 4]. Independent of the permanent presence of cholesterol in a cell membrane, it is transported through the blood as a component of water-soluble carrier aggregates known as lipoproteins. Although cholesterol is essential for the proper functioning of cell membranes, excess cholesterol may precipitate to form cholesterol lodgments (domains) in the inner lining of blood vessels. This triggers the formation of plaque deposition in atherosclerosis disease [5, 6]. In the work reported here a preliminary investigation was made, via computer simulation, of the influence of graphene sheet on the dynamics of cholesterol molecules forming a lodgment around a selected extracellular domain protein. Simulation details We used the standard Lennard-Jones (LJ) interaction potential V(rij) between carbon atoms of the graphene sheet [7] and the atoms (sites) of rigid-body cholesterol C27H45OH, phospholipid and protein. Namely, V (rij ) = 4ε [(σ / rij )12 − (σ / rij ) 6 ] , where rij is the distance between the atoms ith and jth, ε is the minimum of potential at a distance 21/6 σ . The cholesterol molecules include many atomic sites, but in line with the common procedure for large molecules [8], CH, CH2 and CH3 atomic groups were treated as supersites (pseudoatoms). The L-J parameters for these groups and the other atoms involved are taken from [9 – 11] and are presented in Table 1.

142

Perspectives of nanoscience and nanotechnology

Atoms

ε/kB [K]

σ [Å]

m [10-25 kg]

C (for protein) 58.2 3.85 C (for graphene) 28.6 3.85 O 88.7 2.95 N 53.22 3.49 209.76 3.6 S H 12.4 2.81 CH 43.30 3.8 CH2 67.55 3.92 CH3 101.04 3.88 Table 1. Lennard – Jones potential parameters

0.2 0.2 0.27 0.17 0.53 0.02 0.22 0.23 0.25

Moreover, the dipole moments of cholesterol (OH bonds)were included by putting the charge 0.376 e on oxygen and 0.376 e on hydrogen atoms of OH bonds [12]. The L-J potentials parameters between unlike atoms and pseudoatoms were calculated by the Lorentz-Berthelot rules σ A− B = (σ A + σ B ) / 2 and ε A− B = ε Aε B [8], where A, B are C, O, N, S, H, CH, CH2 and CH3 atoms or pseudoatoms. 1KF9 was chosen as an example of human extracellular domain protein [13], (see also Protein Data Bank [14]) and the lodgment consisted of forty cholesterol molecules. The classical equations of motion were integrated by predictor-corrector the Adams-Moulton algorithm [15]. The integration time step was 0.4 fs which ensured total energy conservation within 0.01%. The total simulation time was 1.3 ns (5.0*106 time steps) but the results were collected only for a 400 ps because very long-time simulations were divided into smaller parts. The initial distribution of molecules was generated by the Monte-Carlo (MC) algorithm [15]. Results A snapshot of the instantaneous configuration of our system at the physiological temperature of 310 K is given in Fig. 1.

Fig. 1.

An example of the instantaneous configuration of the protein 1KF9 + (C27H45OH)40 at T ≈ 310 K. The cholesterol molecules are represented by the larger circles.

Solid State Phenomena Vol. 140

143

It can be seen that the cholesterol molecules gather together near the 1KF9 surface, forming a cholesterol rich domain (lodgment) on the protein surface. The calculated radial distribution function g(r) for the centre of mass of cholesterol is shown in Fig. 2.

Fig. 2. The radial distribution function of the centre of mass of the cholesterol molecule in the cholesterol domain. The first peak at 0.72 nm corresponds to the near cholesterol neighbours, the second smaller peak around 1.3 nm is associated with the longer distance cholesterol neighbours (second shell). For the cholesterol intermolecular distance greater than 1.3 nm the value of g(r) systematically decreases, approaching zero at about 5 nm. This quantitatively proves the appearance of a cholesterol lodgment of 5 nm diameter, observed in Fig. 1. r 2 ∆ r (t ) The mean square displacement of the centre of mass of cholesterol at 310 K, where r r 2 r r 2 ∆r (t ) = r (t ) − r (0) and r is the position of the centre of mass of a single molecule, is

shown in Fig. 3.

Fig. 3. The mean square displacement of the centre of mass of the cholesterol molecule in the cholesterol domain.

144

Perspectives of nanoscience and nanotechnology

The slope of

r 2 ∆ r (t )

is connected with the translational diffusion coefficient via the Einstein

relation.

r 2 ∆r (t ) ≈ 6 Dt The plot of

r 2 ∆ r (t )

(1) tells us that cholesterol molecule shows translational mobility when the

graphene sheet is near to the domain. The value of the diffusion coefficient, calculated from the r 2 ∆ r (t ) linear part of plot is D = 2.8 * 10-6 cm2/s. To our knowledge, this is the first reported estimation of the diffusion coefficient of a cholesterol molecule in a cholesterol domain located near the surface of 1KF9 (human growth factor) protein. r r r r r −1 r v C ( t ) = v ( t ) v (0) ⋅ v (0) v (0) The translational velocity autocorrelation function v , where (t ) is the translational velocity of cholesterol molecule is presented in Fig 4.

Fig. 4. The linear autocorrelation function of the centre of mass of the cholesterol molecule in the cholesterol domain. The correlation function C → (t ) decays fast and featureless, reaching zero around 1.5 ps. v

ur ur ur ur −1 The angular velocity autocorrelation function Cωur (t ) = ω (t )ω (0) ⋅ ω (0)ω (0) which is associated ur with reorientational motion of molecule, where ω (t ) is the angular velocity of cholesterol is shown in Fig. 5.

Fig. 5. The angular velocity autocorrelation function of the cholesterol molecule in the cholesterol domain.

Solid State Phenomena Vol. 140

145

The minimum of initial deep of C → (t ) appears at t ≈ 1.1 ps. The intermolecular collisions perturb ω

both the translational and rotational motions of cholesterol molecules in the domain. The change of angular velocity (rotational dynamics) takes place more rapidly than the linear (translational mobility). The graphene sheet strongly influences the cholesterol lodgment, it could even snatch some cholesterol molecules from the lodgment (Fig. 6).

Fig. 6. An example shows that the graphene sheet pulled out the cholesterols localized in the lodgment. The cholesterol molecules are represented by the larger circles. Conclusions The MD calculations show that cholesterol molecules in such a specific environment at the physiological temperature do not disperse over 1KF9 protein surface but gather together and form a well defined cholesterol lodgment. This lodgment is not a solid state phase at the physiological temperature. Cholesterol molecules perform both translational and rotational motions within the domain. The graphene sheet could even pull some cholesterol molecules from the lodgment. The presented results are based on the phenomenological intermolecular potential (Lennard – Jones atom – atom interaction). To take into account the interactions and deformations of electron densities of molecules and some specific interaction which might result from this, ab initio MD simulationsshould be carried out in the future. This will be when the development of computing technology reaches a level that the ab initio treatment of such a computationally demanding system composed of many large, biological molecules would be effective.

146

Perspectives of nanoscience and nanotechnology

Acknowledgement This work was supported in part by The Polish Ministry of Science and Higher Education Grant No. 1 P03B 002 30. The authors wish to thank the Interdisciplinary Centre for Mathematical and Computational Modelling of Warsaw University for their kind allotment of supercomputer time. References [1] M.P. Bokoch, A. Devadoss, M.S. Palencsár and J.D. Burgess: Analitica Chimica Acta Vol. 519 (2004), p. 47-55 [2] S.B. Long, P.J. Casey and L.S. Beese: Nature Vol. 419 (2002), p. 645-650 [3] S.B. Long, P.J. Casey and L.S. Beese: Structure Vol. 8 (2000), p. 209-222 [4] P. Raczynski, A. Dawid, Z. Dendzik and Z. Gburski: Journal of Molecular Structure Vol. 750 (2005), p. 18-21 [5] D.M. Small: N. Engl. J. Med. Vol. 302 (1980), p. 1305–1307 [6] D.M. Small: Arteriosclerosis Vol. 8 (1988), p. 103–129 [7] M.P. Allen and D.J. Tildesley: Computer simulation of liquids (Oxford University Press, Oxford 1989) [8] D.C. Rapaport: The art of molecular dynamics simulation (Cambridge University Press, Cambridge 1995) [9] X. Daura, A.E. Mark and W.F. van Gunsteren: Journal of Computational Chemistry Vol. 19 (1998), p. 535-547 [10] Thomas la Cour Jansen. Theoretical Simulation of Nonlinear Spectroscopy in the Liquid Phase. PH.D. thesis. 2002 [11] T. Kuznetsova and B. Kvamme: Energy Conversion and Management Vol. 43 (2002), p. 26012623 [12] D.H. Phelps and F.W. Dalby: Physical Review Letters Vol. 16 (1966), p. 3-4 [13] C.A. Schiffer, M. Ultsch, S. Walsh, W. Somers, A.M. De Vos and A.A. Kossiakoff: J.Mol.Biol. Vol. 316 (2002), p. 277-289 [14] Information on http://www.pdb.org [15] D. Frenkel and B. Smith: Understanding molecular simulation (Academic Press, New York 2002)

Solid State Phenomena Vol. 140 (2008) pp 147-152 © (2008) Trans Tech Publications, Switzerland doi:10.4028/www.scientific.net/SSP.140.147

Computer Simulation of the Dynamics of Homocysteine Molecules Surrounding a Carbon Nanotube P. Raczynski a, A. Dawid b, Z. Dendzik c and Z. Gburski d Institute of Physics, University of Silesia, Uniwersytecka 4, 40-007 Katowice, Poland e-mail:

a

[email protected]; b [email protected], [email protected]; d [email protected]

Keywords: homocysteine, carbon nanotube, molecular dynamics

Abstract. Excessive amounts of homocysteine in the human body have been considered recently as a factor which increases the risk of developing diseases of the cardiovascular system. The nanosystem composed of homocysteine molecules covering a single walled carbon nanotube have been studied by MD technique. The translational and rotational velocity correlation functions have been calculated for several temperatures, including the physiological temperature of 309 K. The qualitative interpretation of translational and reorientational dynamics of homocysteine molecules in this specific environment is presented. Introduction The functions that homocysteine (C4H9NO2S) plays in the humans body are the subject of a current debate. For example, a high level of the blood serum homocysteine is considered to be a marker of potential cardiovascular disease, that is a risk factor for suffering a heart attack or stroke. It is not yet clear whether high serum homocysteine is itself the problem or if it as merely an indicator of other existing problems [1-4]. Elevated levels of homocysteine have been linked to increased fractures in elderly persons [5, 6]. The molecular level mechanisms for these and other activities of C4H9NO2S are not fully understood. Knowledge of the properties of homocysteine systems may be of help, when considering the role it plays in the complex, biological environment. For example, homocysteine is a component of umbilical cord blood and the studies of the low temperature properties of the homocysteine nanosystem can be related to the issue of cryopreservation of stem cells and stem-cell-based therapies. It is considered that stem-cell-based therapy is one of the most promising issues of modern and future medical treatments [7-10]. This paper reports on the application of molecular dynamics (MD) method to study an ensemble of homocysteine molecules surrounding a carbon nanotube. The dynamics of homocysteine molecules in such a specific environment has not been reported previously. Recently, among the anticipated applications of carbon nanotubes is the use of these materials for the design and development of biological sensors, in this case the homocysteine nanosensor (carbon nanotube based). Simulation details For this work the standard MD simulations for the statistical NVT ensemble were made with the periodic boundary conditions [11] at various temperatures including the physiological temperature, specifically 100, 200 and 309 K. The standard Lennard-Jones (LJ) interaction potential V(rij)

σ

σ

between the atoms (sites) of homocysteine was used [11]. Namely, V (rij ) = 4ε [( )12 − ( ) 6 ] , rij rij

where rij is the distance between the atoms ith and jth, ε is the minimum of potential at a distance 21/6 σ, kB is the Boltzmann constant. The L-J potentials parameters ε and σ are given in Table 1. [12-14].

148

Perspectives of nanoscience and nanotechnology

Atom

ε/kB [K]

σ [Å]

m [10-25 kg]

C O N S H

58.2 88.7 53.22 209.76 12.4

3.85 2.95 3.49 3.6 2.81

0.2 0.27 0.12 0.53 0.02

Table 1. Lennard – Jones potential parameters The rigid-body homocysteine molecule contains 17 atomic sites. Moreover, the dipole moments of homocysteine have been included by putting the charges -0.376 e on the oxide, -0.157 e on the sulphur, -0.53 e on the nitrogen and 0.376 e, 0.157 e, 0.256 e, 0.256 e on the bonded hydrogens, values obtained from ArgusLab [15]. The L-J potentials parameters between unlike atoms were calculated by the Lorentz-Berthelot rules σ A− B = (σ A + σ B ) / 2 and ε A− B = ε Aε B [16, 17], where A, B are different atoms. The classical Newtonian equations of motion were integrated up to 5 ns by a predictor-corrector Adams-Moulton algorithm [11] for the ensemble of 100 homocysteine molecules + a single walled armchair (10, 10) carbon nanotube with the standard periodic boundary conditions [16]. The integration time step was 0.4 fs which ensured total energy conservation within 0.01%. The total simulation time was 150 ps (3.75 *105 time steps). The initial distribution of molecules was generated by the MonteCarlo (MC) algorithm [17]. Results

r 2 The first quantity to be discused is the mean square displacement (MSD) ∆ r (t ) of the centre of r r r r mass of homocysteine, where ∆r (t ) = r (t ) − r (0) and r is the position of centre of mass of a single molecule. Fig. 1 shows MSD for the homocysteine surrounded carbon nanotube at T = 100, 200 and 309 K.

Fig. 1. The mean square displacement of the center of mass of homocysteine at: a) T = 100 K, b) T = 200 K, c) T = 309 K.

Solid State Phenomena Vol. 140

The slope of MSD increases with an increase of temperature. The greater slope of

149

r 2 ∆ r (t ) means

a higher the mobility of molecule, the non-zero value of the slope at T = 100 K indicates a liquid r 2 phase in the system (see insert in Fig. 1). The higher value of ∆ r (t ) for T = 200 and 309 K, by at least two orders of magnitude indicates the gaseous phase at these temperatures. This was confirmed this by calculating the Lindemann index δL [18]: 2

δL =

1 2 2

(< rij > − < rij > ) 2 ∑ N ( N − 1) i < j < rij > N

(1)

where rij is the distance between the centre of mass of ith and jth molecules. The calculated value of the index δL = 0.18 at T = 100 K is low and typical for a liquid phase, compared to δL ≈ 0.32 and 0.34 at T = 200 and 309 K, respectively. The practical independence of the Lindemann index on temperature, and a value higher than 0.1, is characteristic for the gaseous phase in the system [18]. It would be useful to know the value of translational diffusion coefficient D of homocysteine in the liquid/gaseous phase and this can be established because D is linked to MSD by the relation r 2 r 2 ∆ r (t ) is: D = 1.0 ∆r (t ) ≈ 6 Dt [17]. The calculated diffusion coefficient, from the slope of * 10-4, 6.0 * 10-4 and 3.6 * 10-3 cm2/s for T = 100, 200 and 309 K, respectively. Note, that D for the liquid phase is an order of magnitude lower than for the gas. The linear velocity autocorrelation r r r r r −1 function Cvr (t ) = v(t )v(0) ⋅ v(0)v(0) where v (t ) is the translational velocity of homocysteine molecule is shown in Fig. 2.

Fig. 2. The linear velocity autocorrelation function of the center of mass of homocysteine at: a) T = 100 K, b) T = 200 K, c) T = 309 K. The correlation function C → (t ) decays almost exponentially C → (t ) ∼ exp(-t/τv) but the correlation v

time τv depends significantly on temperature (see Fig. 3).

v

150

Perspectives of nanoscience and nanotechnology

Fig. 3. Correlation time τv of the linear velocity autocorrelation function as a function of temperature. In order to obtain a τ(T) plot several additional simulations were made for the intermediate temperatures, up to 400 K. After a nonlinear increasing in the region 100