Electrodeposition from Ionic Liquids

33 downloads 11457 Views 6MB Size Report
Thus these liquids could, in principle, be made water free. (Today we ...... explained by the formation of an IL-like domain with low Tg between the cationic ...... poly(pyrrole) films as a matrix for hosting palladium particles and show that these.
fm

January 11, 2008

10:34

Char Count= 0

Electrodeposition from Ionic Liquids Edited by Frank Endres, Douglas MacFarlane, and Andrew Abbott

i

fm

January 11, 2008

10:34

Char Count= 0

Related Titles Wasserscheid, P., Welton, T. (eds.)

Ionic Liquids in Synthesis 2007 ISBN 978-3-527-31239-9

Staikov, G. T. (ed.)

Electrocrystallization in Nanotechnology 2007 ISBN 978-3-527-31515-4

Paunovic, M., Schlesinger, M.

Fundamentals of Electrochemical Deposition 2006 ISBN 978-0-471-71221-3

Ohno, H.

Electrochemical Aspects of Ionic Liquids 2005 ISBN 978-0-471-64851-2

ii

fm

January 11, 2008

10:34

Char Count= 0

Electrodeposition from Ionic Liquids Edited by Frank Endres, Douglas MacFarlane, and Andrew Abbott

WILEY-VCH Verlag GmbH & Co. KGaA

iii

fm

January 11, 2008

10:34

Char Count= 0

The Editors Prof. Dr. Frank Endres Faculty of Natural and Material Sciences Clausthal University of Technology 38678 Clausthal-Zellerfeld Germany Prof. Douglas MacFarlane School of Chemistry Monash University Clayton, Victoria 3800 Australia

Prof. Dr. Andrew Abbott Chemistry Department University of Leicester Leicester LE1 7RH United Kingdom

 All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek Die Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de c 2008 WILEY-VCH Verlag GmbH & Co.  KGaA, Weinheim All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microfilm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not specifically marked as such, are not to be considered unprotected by law. Composition Aptara, Inc., New Delhi, India Printing Strauss GmbH, M¨orlenbach Bookbinding Litges & Dopf GmbH, Heppenheim Cover Design Kessler, Karlsruhe Printed in the Federal Republic of Germany Printed on acid-free paper ISBN 978-3-527-31565-9

iv

fm

January 11, 2008

10:34

Char Count= 0

v

Contents Preface Foreword

IX XIII

List of Contributors List of Abbreviations

1 1.1 1.2 1.3 1.4 1.5

2

2.1 2.2 2.3

3 3.1 3.2 3.3 3.4 3.5 3.6 3.7

XV XIX

Why use Ionic Liquids for Electrodeposition? 1 Andrew P. Abbott, Ian Dalrymple, Frank Endres, and Douglas R. MacFarlane Non-aqueous Solutions 3 Ionic Fluids 3 What is an Ionic Liquid? 4 Technological Potential of Ionic Liquids 6 Concluding Remarks 12 References 12 Synthesis of Ionic Liquids 15 Tom Beyersdorff, Thomas J. S. Schubert, Urs Welz-Biermann, Will Pitner, Andrew P. Abbott, Katy J. McKenzie, and Karl S. Ryder Synthesis of Chloroaluminate Ionic Liquids 15 Air- and Water-stable Ionic Liquids 21 Eutectic-based Ionic Liquids 31 References 42 Physical Properties of Ionic Liquids for Electrochemical Applications 47 Hiroyuki Ohno Introduction 47 Thermal Properties 47 Viscosity 54 Density 55 Refractive Index 56 Polarity 58 Solubility of Metal Salts 64

Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

fm

January 11, 2008

10:34

Char Count= 0

Contents

vi

3.8 3.9

Electrochemical Properties 66 Conclusion and Future Prospects Acknowledgement 77 References 78

4

Electrodeposition of Metals 83 Thomas Schubert, Sherif Zein, El Abedin, Andrew P. Abbott, Katy J. McKenzie, Karl S. Ryder, and Frank Endres Electrodeposition in AlCl3 -based Ionic Liquids 84 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 103 Troublesome Aspects 114 References 120

4.1 4.2 4.3 4.4

5 5.1 5.2 5.3 5.4 5.5 5.6 5.7

6 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10

7 7.1 7.2

77

Electrodeposition of Alloys 125 I.-Wen Sun, and Po-Yu Chen Introduction 125 Electrodeposition of Al-containing Alloys from Chloroaluminate Ionic Liquids 126 Electrodeposition of Zn-containing Alloys from Chlorozincate Ionic Liquids 132 Fabrication of a Porous Metal Surface by Electrochemical Alloying and De-alloying 137 Nb–Sn 139 Air- and Water-stable Ionic Liquids 140 Summary 145 References 145 Electrodeposition of Semiconductors in Ionic Liquids 147 Natalia Borisenko, Sherif Zein El Abedin, and Frank Endres Introduction 147 Gallium Arsenide 149 Indium Antimonide 149 Aluminum Antimonide 150 Zinc Telluride 150 Cadmium Telluride 151 Germanium 151 Silicon 155 Grey Selenium 160 Conclusions 164 References 164 Conducting Polymers 167 Jennifer M. Pringle, Maria Forsyth, and Douglas R. MacFarlane Introduction 167 Electropolymerization – General Experimental Techniques 171

92

fm

January 11, 2008

10:34

Char Count= 0

Contents

177

7.3 7.4 7.5 7.6

Synthesis of Conducting Polymers Characterization 191 Future Directions 203 Conclusions 207 References 208

8

Nanostructured Metals and Alloys Deposited from Ionic Liquids 213 Rolf Hempelmann, and Harald Natter Introduction 213 Pulsed Electrodeposition from Aqueous Electrolytes 215 Special Features of Ionic Liquids as Electrolytes 220 Nanocrystalline Metals and Alloys from Chlorometallate-based Ionic Liquids 222 Nanocrystalline Metals from Air- and Water-stable Ionic Liquids 227 Conclusion and Outlook 234 Acknowledgement 235 References 235

8.1 8.2 8.3 8.4 8.5 8.6

9

9.1 9.2 9.3 9.4 9.5 9.6

10 10.1 10.2 10.3 10.4 10.5 10.6

11

11.1 11.2 11.3

Electrodeposition on the Nanometer Scale: In Situ Scanning Tunneling Microscopy 239 Frank Endres, and Sherif Zein El Abedin Introduction 239 In situ STM in [Py1,4 ] TFSA 241 Electrodeposition of Aluminum 245 Electrodeposition of Tantalum 250 Electrodeposition of Poly(p-phenylene) 252 Summary 256 References 256 Plasma Electrochemistry with Ionic Liquids 259 J¨urgen Janek, Marcus Rohnke, Manuel P¨olleth, and Sebastian A. Meiss Introduction 259 Concepts and Principles 260 Early Studies 265 The Stability of Ionic Liquids in Plasma Experiments 269 Plasma Electrochemical Metal Deposition in Ionic Liquids 274 Conclusions and Outlook 282 Acknowledgement 283 References 283 Technical Aspects 287 Debbie S. Silvester, Emma I. Rogers, Richard G. Compton, Katy J. McKenzie, Karl S. Ryder, Frank Endres, Douglas MacFarlane, and Andrew P. Abbott Metal Dissolution Processes/Counter Electrode Reactions 287 Reference Electrodes for Use in Room-temperature Ionic Liquids 296 Process Scale Up 310

vii

fm

January 11, 2008

10:34

Char Count= 0

Contents

viii

11.4 11.5

Towards Regeneration and Reuse of Ionic Liquids in Electroplating Impurities 334 Appendix: Protocol for the Deposition of Zinc from a Type III Ionic Liquid 344 References 345

12

Plating Protocols 353 Frank Endres, Sherif Zein El Abedin, Q. Liu, Douglas R. MacFarlane, Karl S. Ryder, and Andrew P. Abbott Electrodeposition of Al from 1-Ethyl-3-methylimidazolium chloride/AlCl3 353 Electrodeposition of Al from 1-Butyl-3-methylimidazolium chloride–AlCl3 – Toluene 356 Electrodeposition of Al from 1-Ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide/AlCl3 358 Electrodeposition of Al from 1-Butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide/AlCl3 360 Electrodeposition of Li from 1-Butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide/Lithium bis(trifluoromethylsulfonyl)amide 362 Electrodeposition of Ta from 1-Butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide 364 Electrodeposition of Zinc Coatings from a Choline Chloride: Ethylene Glycol-based Deep Eutectic Solvent 365 References 367

12.1 12.2 12.3 12.4 12.5

12.6 12.7

13 13.1 13.2 13.3 13.4 13.5 13.6 13.7 13.8 13.9 13.10 13.11

Future Directions and Challenges 369 Frank Endres, Andrew P. Abbott, and Douglas MacFarlane Impurities 369 Counter Electrodes/Compartments 370 Ionic Liquids for Reactive (Nano-)materials 371 Nanomaterials/Nanoparticles 372 Cation/Anion Effects 373 Polymers for Batteries and Solar Cells 373 Variable Temperature Studies 374 Intrinsic Process Safety 374 Economics (Price, Recycling) 375 Which Liquid to Start With? 375 Fundamental Knowledge Gaps 376 Subject Index

379

319

fm

January 11, 2008

10:34

Char Count= 0

ix

Preface Around ten years ago there were only about twenty papers per year dealing with “ionic liquids” or “room-temperature molten salts”. Hence the term “ionic liquid” was unknown to most of the scientific community at that time. Furthermore, there was practically no knowledge of it in industry, and just a handful of groups worldwide were investigating ionic liquids. Ionic liquids were perceived as an academic curiosity. When one of us (F.E.) started his independent research in 1996 with the subject “room-temperature molten salts” many people cautioned him about the eccentric topic. What was the reason for these opinions? From the 1950s to about 1995 most of the people in the community performed studies with ionic liquids based on AlCl3 , often called “first generation” ionic liquids. These are hygroscopic liquids, liberating HCl and a variety of oxo-chloroaluminates upon exposure to moisture. Reproducible operation in these liquids requires either a strictly controlled inert gas atmosphere with extremely low water concentration or at least closed vessel conditions with limited contamination. Thus, these liquids were considered to be difficult to work with and of little practical importance. On the other hand as “room-temperature molten salts” they had attractive electrochemical windows and allowed the electrodeposition of noble metals and of aluminum and its alloys in micrometer thick layers. Aluminum is quite an interesting metal as it is self-passivating, thus under air it forms spontaneously an oxide layer which protects the metal underneath from further corrosion. It was John Wilkes who realized that “room-temperature molten salts” would only experience a widespread interest and uptake if they were stable under environmental conditions. Wilkes’ group published details of the first such liquid in 1992 using the BF4– and the PF–6 anions, the latter showing a miscibility gap with water. Thus these liquids could, in principle, be made water free. (Today we know that ionic liquids containing BF4– and PF6– are subject to decomposition in the presence of water.) Electrochemical studies showed that even these “early” ionic liquids had wide electrochemical windows of about 4 V with cathodic limits of –2 to –2.5 V. vs. NHE. This cathodic limit should, from the thermodynamic point of view, be wide enough to electrodeposit many reactive elements. Around 1995, Seddon realized that the expression “room-temperature molten salts” was counter-productive. The expression “molten salt” was always associated with “high temperature”, as also the editors (and many authors) of this book had Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

fm

January 11, 2008

10:34

x

Preface

Char Count= 0

to experience. The introduction of the term “ionic liquids” for low melting molten salts with melting temperatures below 100◦ C (the definition was actually coined by Walden in 1914) created a clear distinction and the new term “ionic liquid” began to appear in an unprecedented number of publications. In the late 1990s the first papers on the electrodeposition of silver and copper in ionic liquids based on BF4– and PF6– appeared in the literature. From these early papers it was immediately clear that electrodeposition from ionic liquids is not trivial and is actually more complicated than using ionic liquids based on AlCl3 . In addition to viscosity and conductivity concerns, impurities such as water, halides and organic compounds proved to be major difficulties. Thus, nearly twenty years of accumulated knowledge on AlCl3 -based ionic liquids could only be transferred with difficulty to this new class of ionic liquids, because their Lewis acidity/basicity was totally different. Thus, the electrochemistry of these second generation ionic liquids had to be re-invented, more or less. Nevertheless, progress was not slowed and in 2002 alone there were already 600 papers dealing with ionic liquids, about 10% concentrating on electrochemical aspects. In the following years more stable ionic liquids with wider electrochemical windows were developed and cathodic decomposition potentials as low as –3 V vs. NHE were reported, opening the door to the electrodeposition of many reactive elements such as Si, Ge, Ta, Al. Recently a novel class of deep eutectic solvents based on choline chloride have been developed. These can be handled easily under environmental conditions and circumvent many problems that occur in aqueous solutions. They also offer the first economically viable liquids that can be used on an industrial scale. As the interest of electrochemists and classical electroplaters in ionic liquids has risen strongly in the last few years we decided, in 2006, to collect the key aspects of the electrodeposition from ionic liquids in the present book. The book has been written by a panel of expert authors during late 2006 and the first half of 2007 and thus describes the state of the art as of that point in time. In Chapter 1 we explain the motivation and basic concepts of electrodeposition from ionic liquids. In Chapter 2 an introduction to the principles of ionic liquids synthesis is provided as background for those who may be using these materials for the first time. While most of the ionic liquids discussed in this book are available from commercial sources it is important that the reader is aware of the synthetic methods so that impurity issues are clearly understood. Nonetheless, since a comprehensive summary is beyond the scope of this book the reader is referred for more details to the second edition of Ionic Liquids in Synthesis, edited by Peter Wasserscheid and Tom Welton. Chapter 3 summarizes the physical properties of ionic liquids, and in Chapter 4 selected electrodeposition results are presented. Chapter 4 also highlights some of the troublesome aspects of ionic liquid use. One might expect that with a decomposition potential down to –3 V vs. NHE all available elements could be deposited; unfortunately, the situation is not as simple as that and the deposition of tantalum is discussed as an example of the issues. In Chapters 5 to 7 the electrodeposition of alloys is reviewed, together with the deposition of semiconductors and conducting polymers. The deposition of conducting polymers

fm

January 11, 2008

10:34

Char Count= 0

Preface

is still a little neglected in the literature, although the wide anodic decomposition limit allows even benzene to be easily polymerized to poly( p-phenylene) in ionic liquids. Chapter 8 summarizes the principles of nanometal deposition as well as the few examples of nanometal deposition in ionic liquids. Chapter 9 shows how scanning probe microscopy can be used to study the electrodeposition of metals on the submicro- and nano-scale. In situ STM is also used to probe impurities in the ultralow concentration regime. Chapter 10 is devoted to a novel field in the scene, i.e. plasma electrochemistry. By applying a glow-discharge plasma to the surface of an ionic liquid which contains metal ions, suspensions of nanoparticles can be made that might be of interest, for example, as catalysts. Chapter 11 is devoted to technical aspects such as counter electrode reactions, reference electrodes (a very complicated subject), upscaling, recycling and impurities. As industry increases the scale of production the focus on cost and purity will be of increasing importance. In Chapter 12 we provide some plating protocols, which will enable the reader to begin electrodeposition experiments in ionic liquids. In Chapter 13 we have tried to summarize the future directions of the field as we see them and challenging aspects which, in our opinion, warrant further study. Of course, as the field is in a permanent state of development, such a chapter can hardly be comprehensive, but we hope that our thoughts, which are based on many years of experience, will help to stimulate further the field of “electrodeposition from ionic liquids”. Frank Endres, Andrew Abbott and Douglas R. MacFarlane Yokohama, Japan, December 2007

xi

fm

January 11, 2008

10:34

Char Count= 0

xiii

Foreword It is always an honour to be asked to write a foreword for what is clearly an important book, but it is also a curse! What can you say that is original and interesting? – Particularly when the editors themselves have written a Preface!! But this IS an important book – electrodeposition is at the roots and heart of ionic liquid technology. It was one of the earliest applications of ionic liquids, and currently is one of the exciting areas which are developing at an amazing rate. It is a wonderful example of industrial processes developing hand-in-hand with academic research. So, I accepted this cursed honour, and am very glad that I did: the opportunity to see the chapters of this book in advance has been a privilege. So let’s start with the obvious. This book on electrodeposition from ionic liquids comes on the tail of another excellent Wiley book, edited in 2005 by Hiroyuki Ohno, entitled “Electrochemical Aspects of Ionic Liquids”, an updated revision of a 2003 Japanese volume with the title “Ionic Liquids: The Front and Future of Material Development” (CMC Press, Tokyo). Is there any overlap? Well, in the thirty-two chapters of this earlier edited book, which covers the whole spectrum of electrochemistry in ionic liquids, there were only twenty pages devoted to the topic of electrodeposition (an article by Yasushi Katayama). So, there is no significant overlap to worry about. Then, there is the whole question of the philosophy of the edited book? Has it holistic value, or is it just a random collection of articles by disparate authors? Well, the editors here have taken the same approach as Wasserscheid and Welton (“Ionic Liquids in Synthesis”, 2nd Edit., Wiley-VCH, 2007). There is a well developed plan for the book, and the chapters are integrated, and dovetail well. In addition, the authors have been carefully selected – this is a book written by the leading lights of the field. The editors have done an excellent job of producing a volume which deals with the literature, conceptual framework, and practical aspects of the subject. It was particularly pleasing to see chapters and sections dealing with the problems associated with the area, including impurities, recycling and scale-up, reference electrodes, and counter electrodes. Further, as one might expect with Andy Abbott as one of the editors, there is a clear distinction drawn between ionic liquids and deep eutectic solvents. So, is this book perfect? Well, no! One thing drove me to distraction, and it is a problem redolent of the wider literature – the choice of abbreviations Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

fm

January 11, 2008

xiv

10:34

Char Count= 0

Foreword

for the ionic liquids – or, more precisely, the lack of choice! Different chapters used different systems, and each cation and anion was represented in at least four ways within the book (meaning up to, or more than, sixteen possible abbreviations for some ionic liquids. For the simple, symmetrical and common ionic liquid cation, 1,3-dimethylimidazolium, there were five different abbreviations used: [MMIM], [mmim], [C1 mim], [C1 MIM], and [DMIM]; for 1butyl-1-methylpyrrolidinium, there were six different abbreviations used: [Py1,4 ], P1,4 , [BMP], BuMePy, [c4 mpyr], and [c4 mpyrr]. And, even more bizarrely, for the common anion bis(trifluoromethylsulfonyl)amide, six different abbreviations were used: (CF3 SO2 )2 N, NTF, Tf2 N, NTf2 , TFSI, and TFSA. Thus, in principle (I didn’t count!), 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide could have had thirty-six possible abbreviations!! This is way past ridiculous. But the problem doesn’t just lie with the editors; this is reflection of the problem in the wider literature. The burgeoning of the ionic liquid literature during the past decade has meant that there has been no period of stability during which a consensus could be reached. The question of a uniform system, and the wider question of the fundamental definition of an ionic liquid, will have to be addressed elsewhere – the problem is manifest here, however. Another minor issue is that some of the English has a distinctly Germanic ring to it – but never to the point of obscuring the meaning. To summarise then, this book is timely and edited by three of the four main experts in the field. It is planned with meticulous detail, and – of paramount importance – it is authoritative. It is inconceivable that any researcher in the future will not need access to this book, and it will be extensively cited. I congratulate Frank, Doug and Andy on a wonderful volume. Editing books of this type is a service to the community (no one does it for the royalties!), and we owe them a debt of gratitude for the huge investment of time they have made. Kenneth R. Seddon The QUILL Research Centre The Queen’s University of Belfast Belfast, B9 5AG, U.K.

fm

January 11, 2008

10:34

Char Count= 0

xv

List of Contributors Andrew P. Abbott Chemistry Department University of Leicester Leicester LE1 7RH UK Tom Beyersdorff IOLITEG GmbH & Co. KG Ferdinand-Porsche-Straße 5/1 79211 Denzlinger Germany Natalia Borisenko Faculty of Natural and Materials Sciences Clausthal University of Technology 38678 Clausthal-Zellerfeld Germany Po-Yu Chen Faculty of Medicinal and Applied Chemistry Kaohsiung Medical University 807 Kaohsiung Taiwan Richard G. Compton Oxford University Physical and Theoretical Chemistry Laboratory South Parks Road Oxford OX1 3QZ United Kingdom

Jan Dalrymple C-Tech Innovation Ltd. C-Tech United Kingdom Frank Endres Faculty of Natural and Materials Sciences Clausthal University of Technology 38678 Clausthal-Zellerfeld Germany Email: [email protected] Maria Forsyth Australian Centre of Excellence for Electromaterials Science Department of Materials Engineering Monash University Wellington Road Clayton VIC 3800 Australia Rolf Hempelmann Physical Chemistry Department Saarland University 66123 Saarbr¨ucken Germany J¨ urgen Janek Physikalisch-Chemisches Institut Justus-Liebig-Universitaet Giessen Heinrich-Buff-Ring 58 35392 Giessen Germany

Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

fm

January 11, 2008

xvi

10:34

Char Count= 0

List of Contributors

Qunxian Liu Faculty of Natural and Material Sciences Clausthal University of Technology 38678 Clausthal-Zellerfeld Germany

Will Pitner Merck KGaA PLS R&D LSS Ionic Liquids 1 Frankfurter Str. 250 64271 Darmstadt Germany

Douglas MacFarlane School of Chemistry Monash University Wellington Road Clayton VIC 3800 Australia

Manuel P¨ olleth Physikalisch-Chemisches Institut Justus-Liebig-Universitaet Giessen Heinrich-Buff-Ring 58 35392 Giessen Germany

Katy J. McKenzie Chemistry Department University of Leicester Leicester LE1 7RH UK Sebastian A. Meiss Physikalisch-Chemisches Institut Justus-Liebig-Universitaet Giessen Heinrich-Buff-Ring 58 35392 Giessen Germany Harald Natter Physical Chemistry Department Saarland University 66123 Saarbr¨ucken Germany Email: [email protected] Hiroyuki Ohno Department of Biotechnology Tokyo University of Agriculture and Technology 2-24-16 Nakacho, Koganei Tokyo 184-8588 Japan

Jennifer M. Pringle School of Chemistry Monash University Wellington Road Clayton VIC 3800 Australia Email: [email protected]. au Emma I. Rogers Oxford University Physical and Theoretical Chemistry Laboratory South Parks Road Oxford OX1 3QZ United Kingdom Marcus Rohnke Physikalisch-Chemisches Institut Justus-Liebig-Universitaet Giessen Heinrich-Buff-Ring 58 35392 Giessen Germany Karl S. Ryder Chemistry Department University of Leicester Leicester LE1 7RH UK

fm

January 11, 2008

10:34

Char Count= 0

List of Contributors

Thomas J. S. Schubert Managing Director IOLITEC GmbH & Co. KG Ferdinand-Porsche-Straβe 5/1 79211 Denzlingen Germany Debbie S. Silvester Oxford University Physical and Theoretical Chemistry Laboratory South Parks Road Oxford OX1 3QZ United Kingdom I.-Wen Sun Department of Chemistry National Cheng Kung University Tainan 70101 Taiwan Jorg Th¨ oming UFT Section of Chemical Engineering – Regeneration and Recycling University of Bremen, Leobener Str. 28359 Bremen Germany

Daniel Waterkamp UFT Section of Chemical Engineering – Regeneration and Recycling University of Bremen, Leobener Str. 28359 Bremen Germany Urs Welz-Biermann New Business- Chemicals/Ionic Liquids (NB-C) Merck KGaA NB-C, D1/311 Frankfurter Str. 250 64293 Darmstadt Germany Sherif Zein El Abedin Faculty of Natural and Materials Sciences Clausthal University of Technology 38678 Clausthal-Zellerfeld Germany Permanent address: Electrochemistry and Corrosion Laboratory National Research Centre Dokki Cairo Egypt

xvii

fm

January 11, 2008

10:34

Char Count= 0

xix

List of Abbreviations Cations: Pyrrolidinium cations:

1-Butyl-1-methylpyrrolidinium: 1-Propyl-1-methylpyrrolidinium:

[Py1,4 ], P1,4 , [BMP], BuMePy, [c4 mpyr], [c4 mpyrr] P1,3

Imidazolium Cations

1-Methyl-3-methylimidazolium: 1-Ethyl-3-methylimidazolium: 1-Propyl-3-methylimidazolium: 1-Butyl-3-methylimidazolium: 1-Butyl-3-butylimidazolium: 1-Butyl-3H-imidazolium: 1-Ethyl-3H-imidazolium: 1-Hexyl-3-methylimidazolium: 1-Octyl-3-methylimidazolium: 1-Propyl-2,3-dimethylimidazolium: 1-Butyl-2,3-dimethylimidazolium: 1-Etyl-2,3-dimethylimidazolium: 1-Hexyl-2,3-dimethylimidazolium: 1-Decyl-3-methylimidazolium: 1-Benzyl-3-methylimidazolium: 1-Hydroxyethyl-3-methylimidazolium: 1,2-Di-ethyl-3,4-dimethylimidazolium: 1-Alkyl-3-methylimidazolium: 1-(2-hydroxyethyl)-3methylimidazolium: 1-(2-methoxyethyl)-3methylimidazolium: 1-[2-(2-methoxyethoxy)ethyl]-3methylimidazolium:

[MMIM], [mmim], [C1 mim], [C1 MIM], [DMIM] [EMIM], [emim], [C2 mim], [C2 MIM] [PMIM], [pmim], [C3 mim], [C3 MIM] [BMIM], [bmim], [C4 mim] [BBIM], [bbim] [Hbim] [Heim] [HMIM], [hmim], [C6 mim], [HMPL] [OMIM], [omim], [C8 mim] [p-DiMIM], [DMPIM] [b-DiMIM], [C4 -DMIM] [e-DiMIM] [C6 -DMIM] [decyl-MIM], [C10 MIM], [C10 mim] [BZMIM] [HO(CH2 )2 MIM], [C2 OHMIM] [DEDMIM] [Cn MIM], [Cn mim] [C2 OHmim] [C3 Omim] [C5 O2 mim]

Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

fm

January 11, 2008

xx

10:34

Char Count= 0

List of Abbreviations

Pyridinium Cations:

N-Methylpyridinium N-Ethylpyridinium N-Propylpyridinium N-Butylpyridinium: N-Hexylpyridinium:

[MP] [EP], [C2 py], [EtPy] [PP] [BP], [bpyr], [bpyrr], [C4 py] [HP], [HPYR], [C16 py]

Piperidinium Cations:

N-Ethyl-N-methylpiperidinium: N-Propyl-N-methylpiperidinium: N-Butyl-N-methylpiperidinium:

[C2 mPip] [C3 mPip], [PP13 ] [C4 mPip], [PP14 ]

Phosphonium Cations:

Tri-hexyl-tetradecylphosphonium:

[Ph3 t], [P14,6,6,6 ], [P6,6,6,14 ]

Pyrazolium Cations:

N,N-Diethyl-3-methylpyrazolium

[DEMPZ]

Ammonium-Cations:

Trimethylammonium: Tetramethylammonium: 1,1,1-Trimethyl-1methoxyethylammonium: Butyl-trimethylammonium: Benzyl-trimethylammonium: Propyl-trimethylammonium: 1-Cyanomethyl-1,1,1trimethylammonium: 1,1-Dimethyl-1-ethyl-1methoxyethylammonium: 1,1-Diethyl-1-methyl-1methoxyethylammonium: Tributyl-methylammonium: Trimethyl-n-hexylammonium: Tetraethylammonium: Triethyl-hexylammonium: Tetrabutylammonium: Triethyl-hexylammonium: Hydroxyethyl-trimethylammonium: Butyl-diethyl-methylammonium:

[TMHA] [N1111 ], [TMA] [N111,2O1 ] [N1114 ], [N4111 ], [BTMA] [BTMA] [N1113 ], [N3111 ], [PTMA] [N111,1-CN ] [N112,2O1 ] [N122,2O1 ] [N4441 ], [TBMA] [N1116 ], [TMHA] [N2222 ], [TEA] [N2226 ] [N4444 ], [TBA], Bu4 N [N6222 ] [Me3 NC2 H4 OH], Ch also called choline [N1224 ]

fm

January 11, 2008

10:34

Char Count= 0

List of Abbreviations

Sulfonium Cations:

Trimethylsulfonium: Triethylsulfonium: Tributylsulfonium:

[S111 ] TES, [S222 ] TBS, [S444 ]

Anions:

Bis(trifluoromethylsulfonyl) amide:

Trispentafluoroethyltrifluorophosphate: Trifluoroacetate: Trifluoromethylsulfonate: Dicyanoamide: Tricyanomethide: Tetracyanoborate: Tetraphenylborate: Tris(trifluoromethylsulfonyl) methide: Thiocyanate:

(CF3 SO2 )2 N, NTF, Tf2 N, NTf2 , TFSI, TFSA Sometimes this anion is also called bis(trifluoromethylsulfonyl)imide or bistriflamide, bistriflimide FAP ATF, TFA OTF, OTf, TFO, Tf Also called trifluoromethanesulfonate DCA TCM TCB [BPh4 ] [CTf3 ] SCN

Other chemicals:

[CHES]: ChCl: DCM: EDOT: EG: Fc: Fc+ : GC: ITO: PC: PEDOT: TMPD: TMS:

2-(Cyclohexylamino)ethylsulfonate Choline chloride Dichloromethane Ethylenedioxythiophene Ethyleneglycole Ferrocene Ferrocinium Glassy carbon Indium-tin-oxide Propylenecarbonate Polyethylenedioxythiophene Tetramethylphenylenediamine Tetramethylsilane

Abbreviations:

AAS: ACD: AFM: ATR-FTIR:

Atomic Absorption Spectroscopy Anomaleous Codeposition Atomic Force Microscopy Attenuated Total Reflection Fourier Transform Infrared Spectroscopy

xxi

fm

January 11, 2008

xxii

10:34

Char Count= 0

List of Abbreviations

BASIL: CIS: CV: CVD: DSSC: ECALE: EC-STM: EDX, EDS, EDAX: EIS: EMF: EQCM: FAB MS: FFG-NMR: FWHM: HBD: HOPG: HO-ESY: H-REM, H-SEM: H-TEM: ICP: LCA: LED: LSV: MBE: MNDO: NHE: NMR: OCP: OPD: PECD: PED: PLED: PPP: PVD: RTIL: SAED: SIGAL: STM: TSIL: UPD: UHV: VFT, VTF: XPS: XRD:

Biphasic Acid Scavenging Utilizing Ionic Liquids Copper-indium-selenide Cyclic Voltammetry Chemical Vapor Deposition Dye Sensitized Solar Cell Electrochemical Atomic Layer Epitaxy Electrochemical in situ scanning tunnelling microscopy Energy Dispersive X-ray analysis Electrochemical Impedance Spectroscopy Electromotive Force Electrochemical Quarz Crystal Microbalance Fast atom bombardment mass spectroscopy Fixed Field Gradient Nuclear Magnetic Resonance Spectroscopy Full width at half maximum Hydrogen Bond Donor Highly Oriented Pyrolytic Graphite Heteronuclear Overhauser Effect Spectroscopy High Resolution Scanning Electron Microscopy High Resolution Transmission Electron Microscopy Inductively Coupled Plasma (Spectroscopy) Life Cycle Analysis Light Emitting Diode Linear Sweep Voltammetry Molecular Beam Epitaxy Modified neglect of diatomic overlap Normal Hydrogen Electrode Nuclear Magnetic Resonance Open Circuit Potential Overpotential deposition Plasmaelectrochemical deposition Pulsed Electrodeposition Polymer Light Emitting Diode Poly-para-phenylene Physical Vapour Deposition Room Temperature Ionic Liquid Selected Area Electron Diffraction Siemens Galvano-Aluminium Scanning Tunnelling Microscopy Task Specific Ionic Liquid Underpotential deposition Ultrahigh Vacuum Vogel-Tammann-Fulcher X-ray photoelectron spectroscopy X-ray diffraction

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

1

1 Why use Ionic Liquids for Electrodeposition? Andrew P. Abbott, Ian Dalrymple, Frank Endres, and Douglas R. MacFarlane

With any great voyage of discovery the explorer should always be asked at the outset “Why are you doing this?” To answer the question “Why use ionic liquids for electrodeposition?” it is first necessary to look at current best practice and find its limitations. It is widely recognised that in 1805 Italian chemist, Luigi Brugnatelli made the first experiments in what we now know as electroplating. Brugnatelli used the newly discovered Voltaic Pile to deposit gold “I have lately gilt in a complete manner two large silver medals, by bringing them into communication by means of a steel wire, with a negative pole of a voltaic pile, and keeping them one after the other immersed in ammoniuret of gold newly made and well saturated” [1]. The process was later improved by John Wright who found that potassium cyanide was a beneficial electrolyte to add for silver and gold plating as it allowed thick adherent deposits to be obtained. Until the middle of the 19th century the production of jewellery and the gilding of decorative items were the main uses of electrodeposition. With an increased understanding of electrochemistry, the practice of metal deposition spread to non-decorative metals such as nickel, brass, tin, and zinc by the 1850s. Even though electroplated goods entered many aspects of manufacturing industry very little changed about the physical processes involved in electrodeposition for about 100 years. It was only with the advent of the electronics industry in the middle of the 20th century that significant changes occurred in the hardware and chemistry of the plating solutions. The post-war period saw an increase in gold plating for electronic components and the use of less hazardous plating solutions. This trend has continued with increased control of hazardous materials to the environment. Improved solution composition and power supply technology has allowed the development of fast and continuous plating of wire, metal strips, semiconductors and complex substrate geometries. Many of the technological developments seen in the electronics industry depend upon sophisticated electroplating including the use of exotic metals and this is one of the drivers for new technology within the electroplating sector. The other

Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

2 1 Why use Ionic Liquids for Electrodeposition?

main driver is the search for alternative technologies for metals such as chromium, nickel and cadmium. Anti-corrosion and wear-resistant coatings are predominant markets in the electroplating sector and environmental directives will evidently limit their usage in the future. The main metals of that commercially deposited are Cr, Ni, Cu, Au, Ag, Zn and Cd together with a number of copper and zinc-based alloys [1]. The whole electroplating sector is based on aqueous solutions. There are some niche markets based on organic solvents such as aluminum but these are very much exceptions. Metals outside this list are generally deposited using plasma or chemical vapor deposition techniques (PVD and CVD). These methods allow the coating of most substrates (metal, plastic, glass, ceramic etc.) not only with metal but also with alloys or compounds (oxide, nitride, carbide, etc.), without damaging the environment. Although these techniques are technically interesting, it is regrettable that they always involve high capital investment and it is difficult to prepare thick coatings, thus they are only applied to high value niche markets. Clearly the key advantages of using aqueous solutions are: Ĺ Ĺ Ĺ Ĺ Ĺ Ĺ

Cost Non-flammable High solubility of electrolytes High conductivities resulting in low ohmic losses and good throwing power High solubility of metal salts High rates of mass transfer.

For these reasons water will remain the mainstay of the metal plating industry, however, there are also limitations of aqueous solutions including: Ĺ Limited potential windows Ĺ Gas evolution processes can be technically difficult to handle and result in hydrogen embrittlement Ĺ Passivation of metals can cause difficulties with both anodic and cathodic materials Ĺ Necessity for complexing agents such as cyanide Ĺ All water must eventually be returned to the water course.

These prevent aqueous solutions being applied to the deposition of several technically important materials. The key technological goals include replacement of environmentally toxic metal coatings, deposition of new alloys and semiconductors and new coating methods for reactive metals. The main driving force for non-aqueous electrolytes has been the desire to deposit refractory metals such as Ti, Al and W. These metals are abundant and excellent for corrosion resistance. It is, however, the stability of their oxides that makes these metals difficult to extract from minerals and apply as surface coatings.

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

1.2 Ionic Fluids 3

1.1 Non-aqueous Solutions

There is clearly a range of alternative non-aqueous solutions that could be used. Ideally, to obtain the properties required for an electrolyte solution, polar molecules have to be used and these should preferably be small to obtain the requisite high fluidity. Unfortunately, all polar molecules result from having electronegative elements which by their nature makes them good electron donors. Accordingly, they will strongly coordinate to metal ions making them difficult to reduce. While a number of metals have been deposited from polar organic solvents these tend to be the rather noble metals and the processes offer few advantages over aqueous solutions. Some studies have been made using non-polar organic solvents, predominantly aromatic hydrocarbons but these suffer from the serious disadvantage that the dissolved electrolytes are highly associated and the solutions suffer from poor conductivity. The solutions do, however, have wide potential windows and it has been demonstrated that metals such as aluminum and titanium (Ti at least in very thin layers) can be deposited from them. One of the most successful non-aqueous processes is the SIGAL process developed in the late 1980s for the deposition of aluminum from toluene [2, 3]. The aluminum source is triethyl aluminum which is pyrophoric and, despite the high flammability of the electrolyte solution, the process has been commercialized and is currently the only electrochemical method for the deposition of aluminum. A review of electrochemistry in non-aqueous solutions is given by Izutsu [4].

1.2 Ionic Fluids

Clearly an alternative to molecular solvents is the use of ionic fluids. Ionic materials usually melt at high temperatures due to their large lattice energies. Hightemperature molten salts have been extensively used for the electrowinning of metals such as Li, Na, Ti and Al at temperatures of up to 1000 ◦ C [5–7]. They have wide potential windows, high conductivities and high solubilities for metal salts, in fact they have most of the advantages of aqueous solutions and overcome most of the limitations of aqueous solutions, but clearly they suffer from the major limitation that the operational conditions are difficult to achieve and limit the range of substrates that can be used for deposition. The alternative to high-temperature molten salts is to use an ionic substance that melts at a low temperature. While this may sound like an oxymoron it is logical to suppose that the melting point of an ionic substance is related to ionic size and if the ions are made large enough the material will eventually melt at ambient conditions. A significant amount of work was carried out in the middle of the 20th century with the aim of developing lower temperature molten salts. One of the key aims was a lower temperature melt for aluminum deposition which led

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

4 1 Why use Ionic Liquids for Electrodeposition?

to the formation of Li+ / K+ / AlCl3 eutectics which have freezing points close to 100 ◦ C [8]. These low freezing points arise due to the large chloroaluminate anions (AlCl4 − and Al2 Cl7 − ) that form in the eutectic mixtures and have low lattice energies. The use of quaternary ammonium salts, particularly pyridinium and imidazolium salts, has pushed the freezing point down to ambient conditions. The term “ionic liquids” was coined to distinguish these lower temperature ionic fluids from the high temperature analogues which are composed predominantly of inorganic ions. The synthesis and properties of a range of ionic liquids are briefly summarized in the following chapter while the history and chemical properties of these liquids are covered in several well known reviews [9–12]. Several applications of ionic liquids are being tested and these are as diverse as fuel desulfurization [13] and precious metal processing [14] but few have yet come to practical fruition.  R BASF’s BASIL process [15] and the Dimersol process [16] have both been commercialized. The former uses the ionic liquid as a phase transfer catalyst to produce alkoxyphenylphosphines which are precursors for the synthesis of photoinitiators used in printing inks and wood coatings. The imidazole acts as a proton scavenger in the reaction of phenyl-chlorophosphines with alcohols to produce phosphines.  R The Dimersol process uses a Lewis acid catalyst for the dimerization of butenes to produce C8 olefins which are usually further hydroformylated giving C9 alcohols used in the manufacture of plasticizers. Several other processes are also at the pilot plant scale and some ionic liquids are used commercially as additive e.g. binders in paints.

1.3 What is an Ionic Liquid?

The recognised definition of an ionic liquid is “an ionic material that is liquid below 100 ◦ C” but leaves the significant question as to what constitutes an ionic material. Some authors limit the definition to cations with discrete anions e.g. BF4 − , NO3 − . This definition excludes the original work on chloroaluminate systems and the considerable work on other eutectic systems and is therefore unsatisfactory. Systems with anionic species formed by complex equilibria are difficult to categorise as the relative amounts of ionic species depend strongly on the composition of the different components. Ionic liquids have also been separated into first and second generation liquids [10]; where first generation liquids are those based on eutectics and second generation have discrete anions [17]. Others have sought to further divide the first generation liquids into separate types depending on the nature of the Lewis or Brønsted acid that complexes [18]. While there is some dispute whether eutectics with Brønsted acids constitute ionic liquids at all there are others who seek to widen the description of ionic liquids to include materials such as salt hydrates [19]. In general, ionic liquids form because the charge on the ions is delocalized and this gives rise to a reduction in lattice energy. The majority of ionic liquids are

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

1.3 What is an Ionic Liquid?

described by the equilibrium: cation + anion + complexing agent ↔ cation + complex anion

(1.1)

Potentially, complex cations could also be formed using species such as cryptands or crown ethers: cation + anion + complexing agent ↔ complex cation + anion

(1.2)

The confusion arises from the magnitude of the equilibrium constant. For discrete anions such as BF4 − and even ((CF3 SO2 )2 N)− the equilibrium lies clearly to the right of Eq. (1.1). For some eutectic-based liquids the equilibrium constant is also to the right e.g. Cat+ Cl− + AlCl3 ↔ Cat+ + AlCl− 4

(1.3)

But the addition of more Lewis acid produces other anionic species. Cat+ Cl− + 2AlCl3 ↔ Cat+ + Al2 Cl− 7

(1.4)

The use of less Lewis acidic metals e.g. ZnCl2 or SnCl2 will lead to a small amount of Cl− . The species formed between the anion and the complexing agent becomes weaker when a Brønsted acid e.g. urea is used [18]. Cat+ Cl− + urea ↔ Cat+ + Cl− · urea

(1.5)

Others have claimed that, in the extreme, water can act as a good Brønsted acid and, in the extreme, hydrate salts can act as ionic liquids [19]. LiClO4 + 3.5 H2 O ↔ Li+ · xH2 O + ClO− 4 · yH2 O

(1.6)

Ionic liquids with discrete anions have a fixed anion structure but in the eutecticbased liquids at some composition point the Lewis or Brønsted acid will be in considerable excess and the system becomes a solution of salt in the acid. A similar scenario also exists with the incorporation of diluents or impurities and hence we need to define at what composition an ionic liquid is formed. Many ionic liquids with discrete anions are hydrophilic and the absorption of water is found sometimes to have a significant effect upon the viscosity and conductivity of the liquid [20–22]. Two recent approaches to overcome this difficulty have been to classify ionic liquids in terms of their charge mobility characteristics [23] and the correlation between the molar conductivity and fluidity of the liquids [24]. This latter approach is thought by some to be due to the validity of the Walden rule η = constant

(1.7)

5

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

6 1 Why use Ionic Liquids for Electrodeposition?

in ionic liquids, where  is the molar conductivity and η is the viscosity. This is, however a misrepresentation of Eq. (1.7) which was found empirically and is only strictly valid for a specific ion at infinite dilution and constant temperature. The Walden rule is a useful tool for approximate classification of ionic liquids but it actually follows from the Nernst–Stokes–Einstein equation (See Chapters 2.3 and 11.3) [23]. Most importantly, deviations from the Walden rule do not necessarily show that a salt is not an ionic liquid but more usually occur where ionic species deviate from the model of centro-symmetric spherical ions with similar ionic radii. The Walden rule can, however, be used to give evidence of different charge transfer mechanisms e.g. a Grotthus mechanism for protonic ionic liquids [24]. In this book a broad-church of ionic liquids will be assumed, encompassing all of the above types because, in the discipline of electrodeposition, it is the resultant deposit that is important rather than the means. As will be seen later there is also a very fine line between a concentrated electrolyte solution and an ionic liquid containing diluents.

1.4 Technological Potential of Ionic Liquids

A series of transition- and main group-metal-containing ionic liquids have been formulated and the feasibility of achieving electrodeposition has been demonstrated for the majority of these metals, Figure 1.1 shows the elements in the periodic table that have been deposited using ionic liquids. Details of these systems are given in the subsequent chapters and concise summaries exists in recently published reviews [18,25,26]. It must be stressed that while the deposition of a wide range of metals has been demonstrated from a number of ionic liquids the practical aspects of controlling deposit morphology have not been significantly addressed due to the complex nature of the process parameters that still need to be understood. Despite the lack of reliable models to describe mass transport and material growth in ionic liquids, there are tantalizing advantages that ionic liquid solvents have over aqueous baths that make the understanding of their properties vitally important. Some of these advantages include: Ĺ Electroplating of a range of metals impossible to deposit in water due to hydrolysis e.g. Al, Ti, Ta, Nb, Mo, W. As an example, the deposition of Al by electrolysis in a low-temperature process has long been a highly desirable goal, with many potential applications in aerospace for anti-friction properties, as well as replacing Cr in decorative coatings. The deposition of Ti, Ta, Nb, Mo, W will open important opportunities in various industries, because of their specific properties (heat, corrosion, abrasion resistance, low or high density etc.). Ĺ Direct electroplating of metals on water-sensitive substrate materials such as Al, Mg and light alloys with good adherence should be possible using ionic liquids. Ĺ There is potential for quality coatings to be obtained with ionic liquids rather than with water. Currently available metallic coatings suffer from hydrogen

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

1.4 Technological Potential of Ionic Liquids 7

Fig. 1.1 Summary of the elements deposited as single metals or alloys.

Ĺ

Ĺ

Ĺ

Ĺ

embrittlement; a major problem caused by gaseous hydrogen produced during water electrolysis. During electroplating with ionic liquids, negligible hydrogen is produced, and coatings will have the better mechanical properties. Metal ion electrodeposition potentials are much closer together in ionic liquids compared with water, enabling easier preparation of alloys and the possibility of a much wider range of possible electroplated alloys, which are difficult or impossible in water. Ionic liquids complex metals and therefore offer the possibility to develop novel electroless plating baths for coating polymers (e.g. in electronics) without the need for the toxic and problematic organic complexants used in water. Although the cost of ionic liquids will be greater than aqueous electrolytes, high conductivity and better efficiency will provide significant energy savings compared with water, and capital costs will be much lower than the alternative techniques PVD and CVD. When used in electropolishing and electropickling processes, strongly acidic aqueous electrolytes create large quantities of metal-laden, corrosive effluent solution, whereas in ionic liquid electrolytes the metals will precipitate and be readily separated and recycled.

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

8 1 Why use Ionic Liquids for Electrodeposition?

Ĺ The replacement of many hazardous and toxic materials currently used in water, e.g. toxic form of chromium(VI), cyanide, highly corrosive and caustic electrolytes, would save about 10% of the current treatment costs. Ĺ Nanocomposite coatings – nanoparticles giving improved properties compared to microparticles e.g. thermal and electrical conductivity, transparency, uniformity, low friction. Ĺ An increased range of metal coatings on polymers is accessible by electroless plating using ionic liquids containing reducing agents

In the longer term, specialist, ionic liquids will enable technically complex highadded-value products to be introduced, e.g. semiconductor coatings, special magnetic alloys, nanoparticle composite coatings with special erosion/corrosion properties, metal foams for energy storage activated surfaces for self-sterilization purposes (e.g. through photo-catalysis), etc. Also, metals have significantly different reduction potentials in ionic liquid solutions compared to water. For example the difference in reduction potential between Cr and Pt in ionic liquids may be as little as 100 mV whereas in aqueous solutions it is in excess of 2 V. One consequence of this characteristic is that alloy coatings may be prepared more readily and that it should be possible to develop many novel alloy coatings. A fundamental advantage of using ionic liquid electrolytes in electroplating is that, since these are non-aqueous solutions, there is negligible hydrogen evolution during electroplating and the coatings possess the much superior mechanical properties of the pure metal. Hence essentially crack-free, more corrosion-resistant deposits are possible. This may allow thinner deposits to be used, thus reducing overall material and power consumption still further. The electrodeposition of metals from ionic liquids is a novel method for the production of nanocrystalline metals and alloys, because the grain size can be adjusted by varying the electrochemical parameters such as over-potential, current density, pulse parameters, bath composition and temperature and the liquids themselves. Recently, for the first time, nanocrystalline electrodeposition of Al, Fe and Al–Mn alloy has been demonstrated. The properties of the new electrolyte media could also provide much higher health and safety standards for employees in the workplace, i.e. elimination of hazardous vapors, elimination of highly corrosive acidic/alkaline solutions and substantially reduced use of toxic chemicals. These issues are dealt with later in the book. Current aqueous processing systems have a strongly negative impact on the environment (risk of groundwater contamination, soil pollution), which obliges the treatment of wastewater and the dumping of the ultimate waste in landfill. The metal finishing industry in general estimates that at least 15% of turnover is related to the cost of treatment for environmental protection. Legislation within the framework of sustainable development is increasingly stringent (e.g. European Commission directive 96/61/EC concerning “integrated pollution prevention and control”). Thus, industries using metal finishing processes must search for new techniques to achieve these environmental goals. In addition to the growing costs

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

1.4 Technological Potential of Ionic Liquids 9

and negative effects on competitiveness, it is a question of survival in the years to come. The technology developed by this project is generic to most metal-plating systems and, as such, should represent a significant advancement for the environmental sustainability of the metal finishing and electronics manufacturing sectors. Ionic liquids are based on large non-centrosymmetric organic cations with complex anions, which are liquid at room temperature. The range of new ionic liquids has insignificant vapor pressure (thus odorless), some are non-toxic (and even completely biodegradable) and most are highly conductive compared with organic electrolyte solutions. This statement has to be tempered, naturally, because compared to the current state of the art, i.e. concentrated inorganic acids, the conductivities are at best 10 to 100 times lower. An advantage might be that ionic liquids can be operated at temperatures above 100 ◦ C where ionic conductivities of up to 0.2 −1 cm−1 are achievable. The ongoing development of ionic liquids might lead to even better conducting liquids. There are, however numerous risk elements in the development of ionic liquids: Ĺ Coatings must achieve quality standards and a large amount of process development is required. Ĺ A life cycle analysis (LCA) and an environmental impact study have not been completed for any of this technology. Ĺ Issues of scale-up and integration design of generic prototype systems have not been addressed systematically. Ĺ Some applications are at a fundamental research stage with associated higher risk, i.e. electroless coating, semiconductors, anodising, nanocomposite coatings. Ĺ Process economics are expected to be favorable for high-added-value products, but there are likely to be applications where economics are less favorable. Ĺ For improved existing products, customer acceptance is likely to be a significant factor, i.e. reluctance to change product specifications.

The potential impact is extremely broad and fundamental in nature, because the research will explore a totally innovative approach to metal finishing technology, which has never been exploited previously. The use of this completely different type of solvent/electrolyte system, entirely changes the normal behavior of metal finishing processes seen in traditional aqueous electrolytes and an extensive range of entirely new processes and products can be expected. The following chapters discuss the history, development and physical properties of low-temperature ionic substances but in this section it is useful to discuss the differences that arise in changing from a molecular to an ionic environment and the implications that this will have for electrodeposition processes occurring at an electrode surfaces. There are several physical plating parameters that are different in an ionic liquid from those in an aqueous solution. Temperature: Ionic liquids have wide liquid regions, typically in the range −50 to 250 ◦ C which allow more thermodynamic control than is possible in aqueous solutions. This may have potential benefits for the development of new alloys.

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

10 1 Why use Ionic Liquids for Electrodeposition?

Diluents: Ionic liquids can be diluted with a range of organic and aqueous solvents and these significantly affect conductivity, viscosity and metal speciation. The effects have not as yet been characterized and a significant amount of fundamental data still needs to be obtained. A significant amount of work has been carried out on the chloroaluminate-based ionic liquids, although their use in other ionic liquids has been generally ignored. Cation: Cationic structure and size will affect the viscosity and conductivity of the liquid and hence will control mass transport of metal ions to the electrode surface. They will also be adsorbed at the electrode surface at the deposition potential and hence the structure of the double layer is dominated by cations. Some studies have shown that changing the cationic component of the ionic liquid changes the structure of deposits from microcrystalline to nanocrystalline [27]. While these changes are undeniable more studies need to be carried out to confirm that it is a double layer effect. If this is in fact the case then the potential exists to use the cationic component in the liquid as a built-in brightener. Double Layer Structure: Surprisingly few studies have been carried out into the double layer structure of ionic liquids. This is partially due to experimental difficulties but also to interpretation of the resulting impedance spectra. What is clearly evident, however, is that the double layer in an ionic liquid cannot be described by applying the models used for aqueous solutions [28–30]. A study using imidazolium bistriflamide, (F3 CSO2 )2 N− and BF4 − salts suggested that a model of alternating anion and cation layers may be applicable to the data [29, 30]. Baldelli [31, 32] concluded that the double layer is one ion layer thick using sum frequency generation spectroscopy and electrochemistry to probe the electric field at the ionic liquid/electrode interface. The double layer capacitance in an ionic liquid is considerably smaller than in an aqueous solution and less than that predicted by having a perfect Helmholtz layer at the interface, which could result from the presence of ion pairs at the electrode surface at all potentials. Most likely the double layer structure is also influenced by cation/anion interaction. While the structure at the electrode/ionic liquid interface is uncertain it is clear that in the absence of neutral molecules the concentration of anions and cations at the interface will be potential dependent. The main difference between aqueous solutions and ionic liquids is the size of the ions. The ionic radii of most metal ions are in the range 1–2 Å whereas for most ions of an ionic liquid they are more typically 3–5 Å. This means that in an ionic liquid the electrode will be coated with a layer of ions at least 6–7 Å thick. To dissolve in an ionic liquid most metal species are anionic and hence the concentration of metal ions close to the electrode surface will be potential dependent. The more negative the applied potential the smaller the concentration of anions. This means that reactive metals such as Al, Ta, Ti and W will be difficult to deposit as the effective concentration of metal might be too low to nucleate. It is proposed, as one explanation, that this is the reason that aluminum cannot be electrodeposited from Lewis basic chloroaluminate ionic liquids. More reactive metals such as lithium can however be deposited using ionic liquids because they are cationic and therefore

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

1.4 Technological Potential of Ionic Liquids 11

present at high concentrations close to the electrode surface at large negative overpotentials. The strategy to electrodeposit reactive metals must therefore be to either make cationic metal complexes or to work with metal salts at high concentrations. Anode material: In aqueous solutions the anodic processes are either breakdown of the electrolyte solution (with oxygen evolution at an inert anode being favored) or the use of soluble anodes. The use of soluble anodes is limited by the passivation of many metals in aqueous solutions. In ionic liquids, however, the first option is not viable due to the cost and the nature of the anodic breakdown products. New strategies will therefore have to be developed to use soluble anodes where possible or add a sacrificial species that is oxidized to give a benign gaseous product. Preliminary data have shown that for some metals the anodic dissolution process is rate limiting and this affects the current distribution around the cathode and the current density that can be applied. Electrolytes: The above issue of double layer structure is important to the mechanism of nucleation and growth in ionic liquids, it may therefore be possible to control the structure at the electrode/solution interface by addition of an inert electrolyte. In this respect most Group 1 metals are soluble in most ionic liquids, although it is only generally lithium salts that exhibit high solubility. In ionic liquids with discrete anions the presence of Group 1 metal ions can be detrimental to the deposition of reactive metals such as Al and Ta where they have been shown to be co-deposited despite their presence in trace concentrations. Brighteners: Brighteners are added to most aqueous electroplating solutions and work by either complexing the metal ions and decreasing the rate of nucleation or by acting as an interfacial adsorbate, blocking nucleation and hindering growth. Aqueous brighteners have not been studied in depth in ionic liquids and it is doubtful that they will function in the same way as they do in water because of the difference in double layer structure and mass transport. In unpublished work we have surveyed the use of aqueous brightener compounds and applied them to the deposition of zinc and chromium from Type 2 or Type 3 eutectics (see also Chapter 2). None of these were found to be effective in ionic liquids. A small amount of work has been carried out into brighteners that complex the metal ions in solution (see Chapter 11.3) but again no systematic studies have been carried out. Brighteners which rely on electrostatic or hydrophobic interactions may function in ionic liquids but their efficacy is likely to be surface and cation/anion specific. To date all systems that have produced bright metallic finishes have been found to have a nanocrystalline structure which may be due to a progressive nucleation mechanism. This is currently under investigation and if confirmed it will help significantly with the design of future brightener systems. As with other solutes in ionic liquids, the general rule of like dissolving like is applicable i.e. ionic species will generally be soluble as will species capable of interacting with the anion. Aromatic species tend to exhibit poor solubility in ionic liquids consisting of aliphatic cations and vice versa.

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

12 1 Why use Ionic Liquids for Electrodeposition?

1.5 Concluding Remarks

What is clear from this introduction is that the journey into the area of metal deposition from ionic liquids has tantalizing benefits. It is also clear that we have only just begun to scratch the surface of this topic. Our models for the physical properties of these novel fluids are only in an early state of development and considerably more work is required to understand issues such as mass transport, speciation and double layer structure. Nucleation and growth mechanisms in ionic liquids will be considerably more complex than in their aqueous counterparts but the potential to adjust mass transport, composition and speciation independently for numerous metal ions opens the opportunity to deposit new metals, alloys and composite materials which have hitherto been outside the grasp of electroplaters.

References 1 (2000) Modern Electroplating, 4th edn (eds M. Schlesinger and M. Paunovic), John Wiley & Sons, Inc., New York. 2 Peled, E. and Gileadi, E. (1976) J. Electrochem. Soc., 123, 15–19. 3 Simanavicius, L. (1990) Chemija, 3, 3–30. 4 Izutsu, K. (2002) Electrochemistry in Non-aqueous Solutions, Wiley-VCH, Verlag GmbH. 5 Kruesi, W.H. and Fray, D.J. (1993) Metall. Trans. B., 24, 605. 6 Fray, D.J. and Chen, G.Z. (2004) Mater. Sci. Technol., 20, 295. 7 Grjotheim, K., Krohn, C., Malinovsky, M., Matiasovsky, K., and Thonstad, J. (1982) Aluminum Electrolysis, 2nd edn, Aluminium-Verlag, Dusseldorf. 8 Lantelme, F., Alexopoulos, H., Chemla, M., and Haas, O. (1988) Electrochim. Acta, 33, 761. 9 Wasserscheid, P. and Welton, T. (2003) Ionic Liquids in Synthesis, Wiley-VCH, Verlag GmbH. 10 Welton, T. (1999) Chem. Rev., 99, 2071. 11 Wasserscheid, P. and Keim, W. (2000) Angew. Chem. Int. Ed., 39, 3772. 12 Earle, M.J. and Seddon, K.R. (2000) Pure Appl. Chem., 72, 1391. 13 Zhang, S. and Conrad Zhang, Z. (2002) Green Chemistry, 4, 376. 14 Whitehead, J.A., Lawrence, G.A., and McCluskey, A. (2004) Green Chem., 6, 313.

15 Maase, M. (2005) in Multiphase Homogeneous Catalysis (eds B. Cornils et al.), Wiley-VCH, Weinheim, Germany, p. 560. 16 Chauvin, Y., Olivier, H., Wyrvalski, C.N., Simon, L.C., de Souza, R., and Dupont, J. (1997) J. Catal., 165, 275. 17 Chiappe, C. and Pieraccini, D. (2005) J. Phys. Org. Chem., 18, 275–297. 18 Abbott, A.P. and McKenzie, K.J. (2006) Phys. Chem. Chem. Phys., 8, 4265–4279. 19 Xu, W. and Angell, C.A. (2003) Science, 302, 422. 20 Billard, I., Mekki, S., Gaillard, C., Hesemann, P., Moutiers, G., Mariet, C., Labet, A., and Buenzli, J.G. (2004) Eur. J. Inorg. Chem., 6, 1190–1197. 21 Jarosik, A., Krajewski, S.R., Lewandowski, A., and Radzimski, P. (2006) J. Mol. Liq., 123, 43–50. 22 Widegren, J.A., Saurer, E.M., Marsh, K.N., and Magee, J.W. (2005) J. Chem. Thermodyn., 37, 569–575. 23 Abbott, A.P., Harris, R.C., and Ryder, K.S. (2007) J. Phys. Chem. B, 111, 4910–4914. 24 Yoshizawa, M., Xu, W., and Angell, C.A. (2003) J. Am. Chem. Soc., 125, 15411. 25 Endres, F. and Zein El Abedin, S. (2006) Phys. Chem. Chem. Phys., 8, 2101. 26 Zein El-Abedin, S. and Endres, F. (2002) Phys. Chem. Chem. Phys., 4, 1640.

c01 (JWBG008-Endres)

December 25, 2007

13:13

Char Count=

References 27 Moustafa, E.M., Zein El Abedin, S., Shkurankov, A., Zschippang, E., Saad, A.Y., Bund, A., and Endres, F. (2007) J. Phys. Chem. B, 111, 4693. 28 Gale, R.J. and Osteryoung, R.A. (1980) Electrochim. Acta, 25, 1527. 29 Nanjundiah, C., McDevitt, S.F., and Koch, V.R. (1997) J. Electrochem. Soc., 144, 3392.

30 Nanjundiah, C., Goldman, J.L., McDevitt, S.F., and Koch, V.R. (1997) Proc. Electrochem. Soc., 96-25, 301. 31 Baldelli, S. (2005) J. Phys. Chem. B, 27, 109. 32 Rivera-Rubero, S. and Baldelli, S. (2004) J. Phys. Chem. B, 108, 15133.

13

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

15

2 Synthesis of Ionic Liquids Tom Beyersdorff, Thomas J. S. Schubert, Urs Welz-Biermann, Will Pitner, Andrew P. Abbott, Katy J. McKenzie, and Karl S. Ryder

As is well known in the Ionic Liquids Community 109 to 1018 ionic liquids, binary and ternary mixtures have been predicted to be – theoretically – achievable. Of course, this is an incredible number and it will hardly be possible to synthesize all these liquids and investigate all of them in detail for electrochemical purposes. This chapter presents an introduction to some ionic liquids that are interesting for electrochemistry. As the field is still ongoing this chapter can only give an introduction to the principles of ionic liquids synthesis. Section 2.1 briefly summarizes the major aspects of first generation ionic liquids based on AlCl3 , Section 2.2 gives a short introduction to the synthesis of air- and water-stable ionic liquids of the third generation, and Section 2.3 introduces a class of deep eutectic solvents/ionic liquids based on comparatively well-priced educts such as choline chloride. For a more detailed introduction to the chemistry of ionic liquids we would like to refer readers to the 2nd edition of “Ionic Liquids in Synthesis”, ed. by Peter Wasserscheid and Tom Welton (ISBN: 978–3-527–31239-9).

2.1 Synthesis of Chloroaluminate Ionic Liquids 2.1.1 Introduction

Ionic liquids (IL) are a new class of salt-like materials that are entirely composed of ions and that are liquid at unusually low temperatures. For the most commonly used definition of the term ionic liquid the boiling point of water was chosen as a reference point, most likely for emotional reasons: “The term ionic liquids refers to compounds consisting entirely of ions and existing in the liquid state below 100 ◦ C.” In many cases the melting point is even below room temperature. The history of ionic liquids began with the synthesis of ethylammonium nitrate reported in 1914 by Walden [1]. This material is probably the first described in the literature that fulfills the definition of ionic liquids used today. In this context it Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

16 2 Synthesis of Ionic Liquids

should be noted that at that time Walden had of course no idea of this definition or the whole concept of ionic liquids. Consequently, it is not surprising that at the time no attention was paid to the potential of this class of materials. A major breakthrough was achieved in 1951 with the report of Hurley and Wier. They noticed that a mixture of N-ethylpyridinium bromide (EtPyBr) and AlCl3 with a eutectic composition of 1:2 {X(AlCl3 ) = 0.66}1) of EtPyBr to AlCl3 became liquid at unusually low temperatures [2]. They investigated these melts with regard to their potential use in the electrodeposition of aluminum at ambient temperature [3]. Several studies were carried out on this system, however, its use was very limited since it is only liquid at a mole fraction of X(AlCl3 ) = 0.66 and the ease of oxidation of the bromide ion limits the electrochemical stability. In the following years the main interest in ionic liquids was focused on electrochemical applications [4–6]. In 1978 Osteryoung and coworkers replaced EtPyBr with N-butylpyridinium chloride (BuPyCl) and found that the properties of the resulting ionic liquids improved significantly [7, 8]. The new chloroaluminate melts were found to be liquid at room temperature over a composition range from X(AlCl3 ) = 0.66 to 0.43. In addition the anodic limit had improved by changing from bromide to chloride. The main disadvantage of these systems was the relative ease of both chemical and electrochemical reduction of the buytlpyridinium cation [9]. Wilkes and coworkers performed MNDO (modified neglect of diatomic overlap) calculations on a variety of organic cations in 1982 and found that N,N  -dialkylimidazolium cations are more stable than the N-butylpyridinium cation due to the higher electron affinity of these cations [10]. Many of the melts resulting from mixing N,N  -dialkylimidazolium halides with AlCl3 even displayed lower melting points than the N-butylpyridiniumbased ionic liquids. In the case of the 1-ethyl-3-methyl-imidazolium chloride/AlCl3 mixtures the liquid range at room temperature extends from X(AlCl3 ) = 0.66 to 0.30 [11]. Further research on air- and water-stable anions and new cations has been carried out during the past years resulting in more than 1500 materials being described in the literature today [12]. The first part of this chapter focuses on the synthesis and properties of the socalled “first generation of ionic liquids”, the haloaluminate-based ionic liquids and in particular on those of chloroaluminate melts. 2.1.2 Synthesis of Room-temperature Chloroaluminate-based Ionic Liquids 2.1.2.1 Introduction The synthesis of haloaluminate-based ionic liquids from halide salts and aluminum Lewis acids (most commonly AlX3 ; X=Cl, Br) can generally be split into two steps: (i) fomation of the desired cation by the reaction of a trialkylamine, trialkylphosphine or dialkylsulfide with a haloalkane, and (ii) formation of the haloaluminate anion by addition of an appropriate aluminum halide to this salt (Scheme 2.1). 1) The composition of haloaluminate ionic liquids is often described by the mole fraction of AlCl3 X(AlCl3 ) present in the mixture.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.1 Synthesis of Chloroaluminate Ionic Liquids 17

Scheme 2.1 General synthesis route to haloaluminate-based ionic liquids.

Nowadays, as many halide salts are commercially available at reasonable prices, often only the second step is required. The most commonly used groups of cations are presented in Figure 2.1. The following section focuses on the quaternization reaction of 1-alkylimidazoles since these are the most commonly used starting materials for ionic liquids and have dominated ionic liquids research over the last twenty years. However, the general method for the quaternization reaction is similar for pyridines [13], isoquinolines [14], 1-methylpyrrolidine [15], trialkylamines [16], phosphines [17] and sulfides [18]. 2.1.2.2 The Quaternization Reaction From a practical point of view ionic liquids have no significant vapor pressure. As a consequence, their purification using conventional methods is extremely difficult. Thus, it is recommended to remove as many impurities as possible from the starting materials and to use synthetic procedures that produce as few side products as possible, or allow their easy separation from the final product. In addition, all starting materials should be dried prior to use considering the water-sensitive nature of many of the products. All reagents used for the synthesis of cations should be purified according to literature procedures before use [19]. Amines such as 1-alkylimidazoles or pyridines are typically distilled from sodium hydroxide or calcium hydride if dry amines are required and stored under dry nitrogen or argon at 0 ◦ C. Haloalkanes are washed with sulfuric acid until no further color is extracted into the acid layer and then neutralized with NaHCO3 and deionized water prior to distillation from CaCl2 . All solvents used in the syntheses should be dried and distilled prior to use. In order to obtain colorless halide salts it is recommended to perform all reactions under a protective atmosphere of a dry inert gas in order to exclude moisture and oxygen from the reaction. In order to obtain colorless chloroaluminate liquids it is recommended to sublime the AlCl3 several times prior to use after the addition of sodium chloride and aluminum wire [8]. The synthesis of the cation is typically performed by alkylation of an amine, phosphine or sulfide, most commonly using an alkyl halide [ ]. In most cases the reaction is carried out with chloro-, bromo- and iodoalkanes as readily available alkylating reagents, with the reaction conditions becoming more gentle changing from chloride to bromide to iodide, as can be expected for nucleophilic substitution

Fig. 2.1 Examples of cations commonly used for the synthesis of ionic liquids.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

18 2 Synthesis of Ionic Liquids

Scheme 2.2 Quaternization reaction of 1-alkylimidazoles.

reactions. Onium fluorides cannot be synthesized in this manner due to the poor leaving-group qualities of fluoride anions (Scheme 2.2). A typical lab scale alkylating reaction is performed in a round-bottomed flask equipped with a reflux condenser and a dropping funnel with a nitrogen or argon inlet. The alkylating reagent is dissolved in the solvent and the amine is added dropwise. After complete addition, the reaction mixture is heated until all of the amine has been consumed. The reaction conditions for the quaternization are strongly dependent on the haloalkanes used, with the chloroalkanes being the least reactive and the iodoalkanes the most. In general chloroalkanes have to be heated to 80 ◦ C for several days to ensure complete reaction, whereas reactions employing bromoalkanes are usually complete after 24 hours at lower temperatures, between 50 and 60 ◦ C. Alkylation reactions with iodoalkanes can often be performed at room temperature with exclusion of light, since iodoalkanes and the resulting iodide salts are light-sensitive. Taking safety aspects into account, care has to be taken with large-scale reactions employing bromoalkanes, as such reactions are strongly exothermic with increased reaction rates. Besides safety considerations, high thermal stress can also result in discoloration of the final product. The reactivity of haloalkanes in alkylation reactions also decreases with increasing chain length. In general, syntheses of salts with short alkyl substituents are more complex due to the low boiling points of the haloalkanes. The most frequently used halide salt in this field, 1-ethyl-3-methylimidazolium chloride ([EMIM] Cl), is typically synthesized in an autoclave with the chloroethane cooled to below its boiling point (12 ◦ C) before addition. In general, the use of solvents is not inevitably necessary as the reagents are liquid and mutually miscible, while the halide salts are usually immiscible with the starting materials. Nevertheless, solvents are often used to keep the reaction homogeneous and thus to ensure better heat transfer within the reaction mixture. Examples of solvents include the haloalkane itself [10], dichloromethane, acetonitrile, 1,1,1-trichloroethane [20], ethyl acetate [21] and toluene [22]. These solvents can be divided into two classes: those that are miscible with the product salt (dichloromethane, acetonitrile) and those that are immiscible with the halide salt product (1,1,1-trichloroethane, toluene, ethyl acetate). Reactions performed in the former solvents result in homogenous reaction mixtures from which the product can be precipitated, in many cases, by addition of an immiscible co-solvent. For reactions in the latter solvents, removal of the solvent and unreacted starting materials can be achieved by simple decantation and washing of the product with an immiscible solvent, as the product is generally denser than the solvents and starting materials. Purification of the halide salts is in all cases dependent on their state of aggregation. In many cases the halide salts are solids at room temperature and

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.1 Synthesis of Chloroaluminate Ionic Liquids 19

can be recrystallized from mixtures of dry acetonitrile and ethyl acetate. However, if the product does not crystallize, it is advisable to wash the oily product with an immiscible solvent to remove excess starting materials. In all cases it is necessary to remove all excess starting materials, solvents and moisture by heating the product salt under vacuum. Care has to be taken at this stage, as overheating can result in the decomposition of the product via retro-alkylation. It is recommended not to heat the halide salt at temperatures higher than 80 ◦ C. An alternative approach to the reaction conditions described above employs microwave irradiation for the quaternization reaction of 1-methylimidazole with various haloalkanes and 1,ω-dihaloalkanes [23]. High yields and acceptable purities can be obtained in short reaction times (minutes instead of hours) and scaling up this technology to an industrial scale can easily be achieved. As a new class of materials, ionic liquids require special analytical methods. In the case of imidazolium halides and similar compounds the most common impurities are amines, alkyl halides and of course water. Seddon et al. described a method for the detection of residual amines using the strong UV absorbance of copper tetramine complexes. These complexes are readily formed by the addition of Cu2+ ions [24]. The detection of both amines and alkyl halides is possible by NMR spectroscopy but with limited resolution [25]. By far the most powerful analytical method is liquid chromatography combined with UV detection. This sensitive method allows the detection of traces of amines and halides [26]. Unreacted amines can be also detected by ion chromatography combined with a suppressor module. In this case detection is achieved using a continuous flow conductivity cell since amines are protonated and thus detectable. For traces of other ionic impurities ion chromatography is also the most powerful analytical tool [27]. Finally, residual water can be quantified using Karl Fischer titration or coulometry [28]. 2.1.2.3 Chloroaluminate Synthesis Treatment of a quaternary halide salt Q+ X− with a Lewis acid MXn results in the formation of a salt with the composition Q+ MXn+1 − . In general, more than just one anion species is formed, depending on the relative proportions of MXn and the halide salt Q+ X− . A representative example is the reaction of 1-ethyl-3methylimidazolium chloride [EMIM]Cl with AlCl3 (Scheme 2.3). If the mole fraction X(AlCl3 ) is less than 0.5 in the final product, the ionic liquids are basic, as chloride ions are present which are not bound to aluminum and which act as Lewis bases. For mole fractions X(AlCl3 ) > 0.5 an excess of Lewis acid AlCl3 is present and the melts are acidic. If the mole fraction X(AlCl3 ) = 0.5 the salts are neutral as all of the chloride ions are bound to aluminum and the only species present is the [AlCl4 ]− ion. However, as a consequence of the autosolvolysis of

Scheme 2.3 Reaction between [EMIM]Cl and AlCl3 .

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

20 2 Synthesis of Ionic Liquids

Scheme 2.4 Autosolvolysis of AlCl4 melts.

AlCl4 − , Cl− and [Al2 Cl7 ]− species are always present in neutral liquids (Scheme 2.4) [ ]. A detailed description of the analysis of the chloroaluminate species in these ionic liquids is given by Welton et al. [29]. It has to be mentioned that chloroaluminate melts are not the only Lewis acidbased ionic liquids produced in this manner. Other examples include for example AlEtCl2 [30], AlBr3 [31], BCl3 [32], CuCl [33], SnCl2 [34], FeCl3 [35], ZnCl2 [36]. The preparation of these salts is similar to that described for the [AlCl4 ]− salts. Even the treatment of halide salts with metal halides or metal oxides that are not typical Lewis acids has been used to synthesize ionic liquids. Examples of these salts are [EMIM]2 [MCl4 ] (M=Co, Ni) [37], [EMIM]2 [VOCl4 ] [38], [BMIM][CrO3 Cl] [39]. The most common method for the synthesis of chloroaluminate-based ionic liquids is a solid-phase synthesis by mixing AlCl3 and a quaternary halide Q+ X− salt under vigorous stirring. This type of reaction should be carried out using Schlenk techniques or, preferably, in a glove box. Since the ionic liquid is formed directly in an exothermic reaction on contact of the two starting materials, care should be taken upon mixing the reagents. Although the starting materials as well as the products are relatively thermally stable in general, local overheating can result in decomposition and darkening of the ionic liquid. To prevent this, the reaction vessel should be cooled. An important point to avoid overheating is to add one starting material to the other in small portions in order to allow the reaction heat to dissipate. In addition the resulting ionic liquids should be stored under argon in a Schlenk-type flask or in a glove box until use. However, if a glove box is not available for the synthesis, the reaction can also be performed in a dry, inert solvent which covers the reaction mixture and protects it from hydrolysis. An advantage of this procedure is that the solvent, which is typically an alkane, can also react as a heat carrier in the exothermic reaction. After completion of the reaction the ionic liquid forms a second layer below the solvent. The solvent can be removed by simple distillation before use of the ionic liquid. However, the ionic liquid will be contaminated with the organic solvent, which has to be removed under vacuum. Another method involves microwave irradiation. It has been described for the synthesis of 1,3-dialkylimidazolium tetrachloroaluminates [40]. This method precludes the use of volatile organic solvents and is faster, more efficient and also ecofriendly, affording high yields of the desired products. As mentioned above, purification of the resulting ionic liquids cannot be achieved by distillation of the products since these materials show no significant vapor pressure. In most cases AlCl3 -based ionic liquids contain traces of oxo ion impurities such as [AlOCl2 ]− as major impurities, especially if water and oxygen are not totally excluded during synthesis. As shown by 17 O NMR experiments a complex set of equilibria is then present [41]. These impurities can easily be removed by

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.2 Air- and Water-stable Ionic Liquids 21

bubbling phosgene [42] or, considering the high toxicity of phosgene, triphosgene [43] through the ionic liquid. The by-product formed in this reaction is CO2 , which can be easily removed under vacuum. Further purification of room temperature haloaluminate-based ionic liquids is not recommended, as these materials are extremely sensitive towards moisture and must be handled either in vacuum or under an inert gas atmosphere. Although classic Schlenk techniques can be used to handle these materials, working in a glove box is recommended. Analysis of haloaluminate ionic liquids is much more limited than that of other ionic liquids. The most important analytical technique is surely NMR spectroscopy. The determination of residual water is difficult because of the instability of these materials. Hence it is crucial to work accurately to achieve the best results. To summarize the most important points for the synthesis of pure and colorless ionic liquids it is recommended to: Ĺ Purify all starting materials before use Ĺ Exclude oxygen and moisture from the reactions by working in a dry inert atmosphere to prevent darkening of the ionic liquids Ĺ Keep the reaction temperatures as low as possible, as overheating often results in discoloration of the products Ĺ Use Schlenk techniques or work in a glove box, as the chloride and bromide salts are highly hygroscopic and the chloroaluminate melts are highly moisture sensitive.

The importance of these first generation ionic liquids for metal deposition is summarized in Chapter 4.1. 2.1.3 Physical Data of Haloaluminate-based Ionic Liquids

A selection of physical data of selected haloaluminate-based ionic liquids is given in Table 2.1. 2.2 Air- and Water-stable Ionic Liquids 2.2.1 Introduction

For forty years following the introduction of haloaluminate-based ionic liquids by Hurley and Wier, [44, 45] the majority of research in this field was carried out on systems which were reactive with air and, more specifically, with water. The difficulty of working with these materials, using elaborate Schlenk-line airless techniques or expensive and difficult-to-maintain controlled-atmosphere glove boxes, had the effect of limiting the research to four American-based research groups, mostly funded by the US Air Force [46]. Well aware of this limitation, John Wilkes and coworkers made the decision to substitute the reactive haloaluminate anion

(mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%) (mol%)

34.0/66.0 60.0/40.0

40.0/60.0

50.0/50.0 60.0/40.0

34.0/66.0 50.0/50.0 34.0/66.0 50.0/50.0 33.3/66.7 33.3/66.7 33.3/66.7 33.3/66.7

[BMIM]Cl/AlCl3 [BMIM]Cl/AlCl3 [BBIM]Cl/AlCl3 [BBIM]Cl/AlCl3 [MP]Cl/AlCl3 [EP]Cl/AlCl3 [PP]Cl/AlCl3 [BP]Cl/AlCl3

[PMIM]Cl/AlCl3 [PMIM]Cl/AlCl3

[PMIM]Cl/AlCl3

[BMIM]+ [BMIM]+ [BBIM]+ [BBIM]+ [MP]+ [EP]+ [PP]+ [BP]+

[PMIM]+ [PMIM]+

[PMIM]+

[EMIM]+ [EMIM]+

[Al2 Cl7 ]− [Al2 Cl7 ]− [AlCl4 ]− Cl− , [AlCl4 ]− [Al2 Br7 ]− Br− , [AlBr4 ]− [AlCl4 ]− , [Al2 Cl7 ]− [AlCl4 ]− Cl− , [AlCl4 ]− [Al2 Cl7 ]− [AlCl4 ]− [Al2 Cl7 ]− [AlCl4 ]− [Al2 Cl7 ]− [Al2 Cl7 ]− [Al2 Cl7 ]− [Al2 Cl7 ]−

[MMIM]+ [EMIM]+ [EMIM]+ [EMIM]+

19 27 24 38 21 18 18 21

27

18

32 67

17 14 18 47

Viscosity (mPa s)

9.2 10.0 6.0 54.0 8.1 10.0 8.0 6.7

12.0 3.3

11.0

5.8 5.7

15.0 15.0 23.0 6.5

Conductivity (mS cm−1 )

3.04 2.49 2.32 1.50 2.23 2.91 2.47 2.18

2.79

2.94

1.89 1.15

4.26 4.46 4.98 1.22

Molar conductivity

1.334 1.238 1.252 1.164 1.441 1.408 1.375 1.346

1.262

1.351

2.219 1.828

1.404 1.389 1.294 1.256

Density (g cm−3 )

[10] [10] [10] [10] [36] [36] [36] [36]

[10] [10]

[10]

[31] [31]

[10] [10] [10] [10]

Ref.

13:16

[EMIM]Br/AlBr3 [EMIM]Br/AlBr3

[MMIM]Cl/AlCl3 [EMIM]Cl/AlCl3 [EMIM]Cl/AlCl3 [EMIM]Cl/AlCl3

Anion

Cation

December 25, 2007

34.0/66.0 34.0/66.0 50.0/50.0 60.0/40.0

Composition of the IL System

Table 2.1 Physical data of selected ionic liquids.

c02 (JWBG008-Endres) Char Count=

22 2 Synthesis of Ionic Liquids

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.2 Air- and Water-stable Ionic Liquids 23

systems with less reactive anions, under the belief that other anions could also produce low-melting organic salts [47]. Perhaps they were aware of the singular work by Walden to generate low melting, high conductivity salts [48], work which is often used to mark him as the discoverer or inventor of ionic liquids. However, Walden based his “molten salts” upon protonated primary amines such as ethylamine, and such mixtures of organic bases with mineral and organic acids will always exist as mixtures of the acid, the base and salt formed through their neutralization. Such equilibrium mixtures are well known to be thermally unstable, due to the vapor pressure of the two neutral components [49]. The ease with which the protonated cation can be reduced to yield hydrogen gas also limits the usefulness of these materials in electrochemical applications. The non-chloroaluminate ionic liquids introduced by Wilkes have the advantages of high thermal and electrochemical stability (with respect to Walden’s acid–base equilibrium mixtures) and ease of handling under ambient, humid conditions (as compared to Hurley and Wier’s haloaluminate ionic liquids). The methathesis route employed by Wilkes in the production of these materials has generally been followed for the majority of air- and moisture-stable ionic liquids. In brief (Scheme 2.5), an organic base (such as N-methylimidazole, pyridine or Nmethylpyrrolidine) is alkylated using a haloalkane to generate an organic halide salt. Anion exchange is carried out, generally in water, with the appropriate acid or metal salt. The ionic liquid is extracted from the aqueous salt into an organic phase, and the halide impurities removed through repeated washings with water. The more hydrophilic the ionic liquid, the more difficult it is to purify, as extraction of halides with water is complicated by loss of the ionic liquid to the aqueous phase.

Scheme 2.5 General synthetic route to producing air- and moisture-stable ionic liquids.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

24 2 Synthesis of Ionic Liquids

Fig. 2.2 Structure, full name and abbreviations for the anions discussed in this section.

The discovery made by Wilkes in the early nineties was completely dependent upon a change in the anionic systems: the cation components of the haloaluminate systems previously in use were not the cause of their reactivity with water and the pyridinium and imidazolium cations remain key components of air- and moisturestable ionic liquids under investigation today. Therefore the focus of this section will be air- and moisture-stable anionic systems (Figure 2.2), with the cation relegated to the rˆole of junior partner in an ionic couple. Due to their ease of handling, this report will also focus on those anion–cation combinations which yield roomtemperature ionic liquids (RTILs), even though operation at room-temperature is not a prerequisite for a commercially-viable electroplating bath.

2.2.2 Tetrafluoroborate and Hexafluorophosphate-based Ionic Liquids

Wilkes launched the field of air- and moisture-stable ionic liquids by introducing five new materials, each containing the 1-ethyl-3-methylimidazolium cation [EMIM]+ with one of five anions: nitrate [NO3 ]− , nitrite [NO2 ]− , sulfate [SO4 ]2− , methyl carbonate [CH3 CO2 ]− and tetrafluoroborate [BF4 ]− [47]. Only the last two materials had melting points lower than room temperature, and the reactive nature of the methyl carbonate would make it unsuitable for many applications. This led to the early adoption of [EMIM][BF4 ] as a favored ionic liquid, which has since been the subject of over 350 scientific publications. One of the first appeared in 1997 [50], reporting the investigation of [EMIM][BF4 ] as the electrolyte system for a number of processes, including the electrodeposition of lithium (intended for use in lithium ion batteries).

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.2 Air- and Water-stable Ionic Liquids 25

The [BF4 ]− anion is frequently used in battery electrolyte formulations, and it was not long before other anions from this branch were investigated for their capacity to make ionic liquids. The combination of hexafluorophosphate [PF6 ]− with [EMIM]+ produced a salt with a reported melting point of 58–60 ◦ C [51]. This was too high a melting point for most researchers to bother working with, but in 1995 Chauvin, Mussmann and Olivier reported the use of the analogous [BMIM][PF6 ], which is liquid at room temperature, as well as [BMIM][BF4 ] [52]. These two RTILs would dominate ionic liquid publications for the next decade. The preference for [BMIM][BF4 ] over [EMIM][BF4 ] can probably be explained by the fact that the synthesis of [BMIM]Cl is an easier process than the synthesis of [EMIM]Cl, which requires a pressurised reaction vessel, and by the high water-solubility of [EMIM][BF4 ], which makes it much more difficult to purify than [BMIM][BF4 ]. Another reason was the experimental symmetry afforded by switching from an ionic liquid which is completely miscible with water ([BMIM][BF4 ]) to one which formed biphasic aqueous mixtures ([BMIM][PF6 ]). Other cation combinations with [BF4 ]− and [PF6 ]− have proved uninteresting in the study of electrochemical systems. Although N-butylpyridinium tetrafluoroborate [bpyr][BF4 ]− is known to be a RTIL [53], the lower electrochemical stability of pyridinium-based cations relative to imidazolium limits their electrochemical applicability. On the other hand, pyrrolidinium-based cations are known to be more electrochemically stable than imidazolium salts, N-alkyl-N-methylpyrrolidinium salts of [BF4 ]− and [PF6 ]− are made less attractive to researchers by the fact that they are solids at room temperature [54, 55]. Therefore, most of the electrochemical investigations of ionic liquids containing [BF4 ]− and [PF6 ]− have focused on [BMIM][PF6 ], [BMIM][BF4 ] and, to a lesser extent, [EMIM][BF4 ]. Concerns about the stability of [BF4 ]− and [PF6 ]− and ionic liquids which contain these anions have led many researchers to turn their backs on these materials. Anecdotal evidence of glassware shattering during heating and vacuum drying is commonplace, but more rigorous investigations confirm the rumours that [BF4 ]− and [PF6 ]− -based ionic liquids hydrolyse to generate HF, a corrosive and toxic material [56]. Experiments performed by Merck KGaA demonstrate this instability with respect to ionic liquids based on other fluorinated anions (Figure 2.3). Despite such warnings, however, research continues on these materials for a number of reasons: the large amount of baseline data on [BMIM][PF6 ], [BMIM][BF4 ] and [EMIM][BF4 ] which is available from prior experimentation and publications; the ease with which these materials can be produced; and their low cost relative to other more complex and stable anion systems. In addition, [BF4 ]− and [PF6 ]− -based ionic liquids can possess properties which, for a given application, provide a superior performance to other ionic liquids [57]. 2.2.3 Triflate- and Trifluoroacetate-based Ionic Liquids

Small, fluorinated organic anions, such as trifluoromethanesulfonate (or triflate) and trifluoroacetate, were quickly considered as alternatives to inorganic fluorinated phosphates and borates. Carlin and coworkers were the first to report

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

26 2 Synthesis of Ionic Liquids

Fig. 2.3 A comparison of the hydrolytic stability of four 1-hexyl-3-methylimidazolium ionic liquids [HMIM]X ionic liquids. To 25 g of each ionic liquid, 7.1 mol% water was added. These solutions were heated to

60 ◦ C and the fluoride content measured once an hour for eight hours. All measurements performed at Merck KGaA, Darmstadt, Germany.

an investigation using 1-ethyl-3-methylimidazolium triflate [EMIM][OTF], looking at different RTILs for use in an ongoing battery project [58]. The 1-butyl-3methylimidazolium triflate [BMIM][OTF] was suggested as an electrolyte component for dye-sensitised solar cells (DSSCs) [59], as was 1-ethyl-3-methylimidazolium trifluoroacetate [EMIM][ATF] [60]. These materials in general form ionic liquids with relatively low viscosities, and are characterized by reasonably large electrochemical windows (though not comparable with the inorganic fluorinated anions, see Table 2.2) [61]. The sulfate and carboxylate functional groups make them strongly coordinating anions, although the electron-withdrawing trifluoromethane Table 2.2 Dependence of selected physicochemical properties (at 20 ◦ C)

of ionic liquids [EMIM]X on the anion X− .

Ionic liquid

[EMIM][BF4 ] [EMIM][FAP] [EMIM][ATF] [EMIM][OTF] [EMIM][NTF] [EMIM][SCN] [EMIM][DCA] [EMIM][TCM] [EMIM][TCB] a

Density/g cm−3

Dynamic viscosity/mPa s

Specific conductivity/mS cm−1

ERed–Ox /V

1.30 1.72 1.30 1.39 1.52 1.15 1.08a,b 1.11b 1.04

60 75 41 52 40 44 16a (17)b 18b 20

11 4 5 7 8 14 28a (27)b 18b 13

5.2 6.5 3.4 4.1 6.3 3.2 3.5a,b 3.5b 4.5

At 25 ◦ C [J. Phys. Chem. B, 111(18), 2007]. At 22 ◦ C [Inorganic Chemistry, 43(4), 2004]. All other measurements performed at Merck KGaA, Darmstadt, Germany.

b

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.2 Air- and Water-stable Ionic Liquids 27

component makes them much less basic than their methanesulfonate and acetate analogs. While the decreased electrochemical stability may be viewed as a negative when considering such ionic liquids as potential electroplating baths, their lower viscosities are an attractive feature while their ability to coordinate may also work in their favor by increasing the solubility of metal salts. 2.2.4 Bistriflamide-based Ionic Liquids

Lithium bis(trifluoromethylsulfonyl)amide Li[NTF] has been widely recognised as a possible component in battery electrolyte compositions since 1990 [62]. A readily available material, Li[NTF] can be converted into [NTF]− -based ionic liquids through a very simple ion exchange step in an aqueous mixture because most [NTF]− based ionic liquids form biphasic aqueous mixtures, as reported by Bonhˆote and coworkers in 1996 [59, 60]. The ionic-liquid-rich phase is easily separated and can be purified to a high level through simple washing with water. In the same year, Watanabe and Mizumura reported ionic liquids based upon Li[NTF] in combination with lithium acetate and triethyl methyl ammonium benzoate. In addition to their ease of preparation, [NTF]− -based ionic liquids are generally characterised by higher electrochemical and thermal stability, lower viscosity and higher conductivity than ionic liquids based on [BF4 ]− and [PF6 ]− (Table 2.2). This collection of favorable properties is one reason why there is currently great interest in this class of ionic liquids. While initial interest in [NTF]− -based ionic liquids focused on imidazolium salts, it soon became clear that a broader range of cations could be paired with [NTF]− to generate RTILs [63–65]. For electrochemists, the higher electrochemical stability of tetraalkylphosphonium, tetraalkylammonium, N,N-dialkylpyrrolidinium and N,N-dialkylpiperidinium [NTF]− -based RTILs make them attractive alternatives to 1,3-dialkylimidazolium and N-alkylpyridinium salts. This is especially true for applications involving the electrodeposition of active metals, where reactions between the electrodeposited metal and the ionic liquid plating bath should be avoided. More recently, concerns have arisen with respect to the assumed stability of [NTF]− -based ionic liquids. In an investigation of the electrochemical behavior of lithium in such ionic liquids, MacFarlane and coworkers reported the decomposition of [NTF]− , due either to unwanted reactions between the active metal surface of the electrode or the electrochemical reduction of the anion at the negative potentials required for lithium reduction [66]. This potential instability of [NTF]− to reduction had been predicted by Makato and coworkers two years earlier, using ab initio molecular orbital calculations [67]. In addition, their possession of a negligible vapor pressure, which was previously assumed to be a general property of all ionic liquids, has been called into question with reports of the distillation of [NTF]− -based ionic liquids [68]. Moreover, while [NTF]− has been demonstrated to be much more stable than [BF4 ]− and [PF6 ]− to hydrolysis (Figure 2.3), ab initio calculations suggest that [NTF]− -based ionic liquids may be much more volatile than those based on [BF4 ]− and [PF6 ]− [69]. In addition, [NTF]− -based ionic liquids, though often classified as “hydrophobic” due to their formation of biphasic

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

28 2 Synthesis of Ionic Liquids Table 2.3 A comparison of the relative hydrophobicity of eight ionic

liquids. Ionic liquid

[HMIM][BF4 ] [HMIM][TCB] [HMIM][PF6 ] [Ph3 t][NTF] [HPYR][NTF] [HMIM][NTF] [HMPL][NTF] [HMIM][FAP] [Ph3 t][FAP] [HMPL][FAP]

Water content of IL/wt%

IL content of water/ppm

18.1 5.39 1.84 1.58 1.13 1.12 0.900 0.195 0.180 0.114

7.27 0.44 — — 0.25 0.13 — 0.02 — —

aqueous mixtures, are demonstrably soluble in water (Table 2.3); even a a small loss of [NTF]− to the water-rich phase during synthesis could result in circumstances where the economics of commercial applications are called into question by the loss of this high-cost component. 2.2.5 Trispentafluoroethyltrifluorophosphate-based Ionic Liquids

The anion trispentafluoroethyltrifluorophosphate [FAP]− belongs to the broader class of perfluoroalkylphosphate-based anions first reported in the 1960s [70]. Merck KGaA began investigating the use of Li[FAP] as a component in battery electrolytes in 2001, as a replacement for Li[PF6 ] [71]. The hydrolytic instability of the hexafluorophosphate anions is due to the facial protonation of the fluorine atom, followed by HF elimination and further reaction with water. This problem is addressed by replacement of some of the fluorine atoms by hydrophobic perfluoroalkyl groups, reducing the rate of hydrolysis through steric hindrance of attacks on the phosphate center. Recently, Merck KGaA developed a convenient method for the synthesis of [FAP]− -based ionic liquids as replacement for [PF6 ]− -based ionic liquids [72]. Like their [PF6 ]− analogs, [FAP]− -based ionic liquids form biphasic aqueous mixtures and can be separated and recovered easily from aqueous reaction mixtures. They can be easily obtained with very low water and chloride content by washing with water followed by heating under reduced pressure. The hydrolytic stability (Figure 2.3) and electrochemical stability (Table 2.2) of [FAP]− and its ionic liquids are superior to [PF6 ]− and [BF4 ]− and comparable with [NTF]− . 2.2.6 Cyano-based Ionic Liquids

A family of ionic liquids has developed around anions containing a central element coordinated by one or more cyano groups. The stability of the carbon–nitrogen triple

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.2 Air- and Water-stable Ionic Liquids 29

bond of the cyano group, its high electronegativity and ability to increase charge delocalization combine to give this family a unique set of chemical properties. While generally electrochemically stable, they are also capable of strong coordination and solvation of polar hydrogen-bond donating materials such as cellulose and sugars [73, 74]. Cyano ligands are prone to polymerization during decomposition, and preliminary investigations have indicated that the main concern with this type of anion, the formation of HCN during their thermal decomposition, is negligible [75, 76]. They are generally much more hydrophilic than fluorinated anions. The simplest example is the thiocyanate anion [SCN]− . Thiocyanate-based ionic liquids, such as [EMIM][SCN] [77] show good thermal stability, low melting points and electrochemical stability sufficient for a wide range of electrochemical applications, most especially for dye-sensitised solar cells [78–80]. Their ability to dissolve metal thiocyanates in high quantities is significant [81]. Nitrogen can coordinate two cyano ligands, to form the dicyanamide anion [DCA]− , which can form ionic liquids [82]. Certainly the popularity of [EMIM][DCA] is its extremely low viscosity of 17 cP (extremely low for an ionic liquid, that is). This anion can form RTILs with a broad range of electrochemically stable cations, including imidazoliums, ammoniums and pyrrolidiniums [83]. Because of these properties, dicyanamide-based ionic liquids have been considered for a wide variety of electrochemical applications [84], most especially for dye-sensitized solar cells (DSSC) [85, 86]. While the tricyanomethide anion [TCM]− can also be used to make RTILs [87] and has also been investigated for photovoltaic applications [88], this anion system has been much less investigated than thiocyanates and dicyanamides. The synthesis of the tetracyanoborate anion [TCB]− was first described by Bessler [89, 90], but only the improvement of the sinter process of the key intermediate potassium [TCB] [91] has made this material available in reasonable amounts and thus allowed the synthesis of [TCB]− -based ionic liquids [92]. The low viscosity (20 cP at 20 ◦ C) and thermal and chemical robustness led to the use of [EMIM][TCB] as an electrolyte for DSSC [93]. 2.2.7 Effect of Anion on Ionic Liquid Physicochemical Properties

The choice of anion will have a known effect on the physicochemical properties of the ionic liquid. To demonstrate the anion effect, selected data on properties of general interest to electrochemists (density, viscosity, conductivity and electrochemical window) have been gathered in Table 2.2. In each case, the anion is paired with the same cation: 1-ethyl-3-methylimidazolium. Certain trends from this data can be generalized, as well as in other collections of such data (for example, see Ref. [61] and references therein), that hold true regardless of the identity of the cation. For example, the effect of the anion on density follows the trend: [TCB]− < [DCA]− < [TCM]− < [SCN]−  [BF4 ]− < [ATF]− < [OTF]−  [PF6 ]− < [NTF]− [TCB]− ∼ [SCN]−  [BF4 ]− > [NTF]− ∼ [ATF]− ∼ [OTF]− > [FAP]− > [PF6 ]− likewise appearing to be related to anion size and coordination strength, as well as the amount of charge delocalization. Ionic liquids which form biphasic aqueous mixtures are often classified as hydrophobic, regardless of the fact that this phase behavior is temperature dependent and that ionic liquids are, in general, hygroscopic. The hydrophobicity of an ionic liquid used as an electroplating bath is an important factor if exclusion of water from the bath is important: the more hydrophobic the ionic liquids, the lower the water content will be upon saturation. The relative hydrophobicity of an ionic liquid is a factor of both the anion and cation, as has been demonstrated by research carried out by Merck KGaA (Table 2.3). To make this comparison, equal volumes of an ionic liquid and water were mixed for 2 h at room temperature, then allowed to separate into two phases. The water content of the ionic-liquid-rich phase was then determined with Karl–Fischer titration, and the ionic liquid content of the water-rich phase was determined using high performance liquid chromatography (HPLC) (1-hexyl-3methylimidazolium [HMIM]+ and N-hexylpyridinium [HPYR]+ ) or ion chromatography (IC) (1-hexyl-1-methylpyrrolidinium [HMPL]+ ). No satisfactory method was found for quantifying the trihexyl-tetradecylphosphonium bistriflamide [Ph3 t][NTF] content of the water-rich phase. The clear trend of increasing hydrophobicity for the four cations evaluated is [Ph3 t]+ < [HMIM]+ ∼ [HPYR]+ < [HMPL]+ while the trend for the five anions is [BF4 ]− < [TCB]− < [PF6 ]− < [NTF]− < [FAP]− In summary, there are many anion types which offer useful properties for the creation of an electroplating medium. Choices must be made regarding electrochemical stability, relative hydrophobicity, the ability to coordinate metal salts and the mass transport properties of viscosity and conductivity. 2.2.8 Concluding Remarks

In the years following Wilkes introduction of air- and moisture-stable ionic liquids, these materials have been transformed from laboratory curiosities which each

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.3 Eutectic-based Ionic Liquids Table 2.4 Standard purity grades of ionic liquids available from Merck

KGaA. Merck purity grade “for synthesis”

“high purity”

“ultra pure”

> 98 < 1000 99 < 100 < 100 ppm

> 99 < 10 < 10 ppm

overall purity (%) halide content (ppm) water content

researcher had to prepare in-house, to commercially available materials. Over 200 individual ionic liquids in a variety of purity grades can be ordered from a range of manufacturing companies including Merck KGaA (EMD Chemicals), BASF and Cytec and chemical supply companies such as Fluka, Simga-Aldrich, VWR and Kanto Chemical Co. Although the technical grades of ionic liquids which are available from most companies are unsuitable for electrochemical applications, a number of suppliers do offer higher purity grades. For example, Merck KGaA offers its products in three purity grades (Table 2.4): “for synthesis” grade, which is suitable for most non-electrochemical applications; “high purity” grade, which is suitable for many catalytic and electrochemical applications; and “ultra pure” grade, which was specified with electrochemical applications in mind. As ionic liquids are adopted for industrial applications, questions are arising concerning their toxicity, their impact on the environment and the registration of these new chemicals with regulatory bodies. Several academic groups have led the way in exploring the relationship between ionic liquids structures and their (eco) toxicological effects [94–96] and their biodegradability [96, 97]. Information from these studies will be useful in the design of more benign ionic liquids. Although thorough studies of ionic liquids are rare, the fact that 1-ethyl-3-methylimidazolium ethylsulfate has been classified as a non-toxic material gives a good indication that other benign ionic liquids are highly likely. Although this ionic liquid is unlikely to find use in electrochemical processes, due to the instability of the anion, there is so far no reason to doubt that one or more environmentally friendly ionic liquids exist with the physicochemical properties suitable to make a key component in an electroplating system.

2.3 Eutectic-based Ionic Liquids

The melting point of two component mixtures is dependent upon the interaction between the components. For non-interacting components the freezing point can vary linearly with mole fraction whereas large negative deviations can occur when the components interact strongly with each other. This is shown schematically in Figure 2.4. The composition at which the minimum freezing point occurs is

31

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

32 2 Synthesis of Ionic Liquids

Fig. 2.4 Schematic representation of a eutectic point on a two-component phase diagram.

known as the eutectic point and this is also the temperature where the phases simultaneously crystallise from molten solution. The word eutectic comes from eutektos which is Greek for easily melted. Eutectic mixtures have been used extensively for applications of molten salts to reduce the operating temperature and this is where the significant area of ionic liquids developed from i.e. the quest to find aluminum-based salt mixtures. While the development of aluminum-containing ionic liquids is technologically very important for the field of metal deposition it is clear that there are many other issues that also need to be addressed and hence methods need to be developed to incorporate a wide range of other metals into ionic liquid formulations. While the first aluminum-based ionic liquids were reported in the 1950s [94], it was not until the late 1990s that other metal salts were used to form ionic liquids. Work by Abbott et al. [95, 96] and Sun et al. [97, 98] showed that eutectic mixtures of zinc halides and quaternary ammonium halides also have melting points close to ambient conditions. This has been further extended to a wide range of other salts and organic compounds that form eutectic mixtures with quaternary ammonium salts. This area has received comparatively little attention compared with the chloroaluminate and discrete anions but the principle is simple in that the complexing agent just needs to be able to complex the simple anion to effectively delocalize the charge and decrease the interaction with the cation. This is shown schematically in Figure 2.5. The systems so far described can be expressed in terms of the general formula Cat+ X− · z Y, where Cat+ is in principle any ammonium, phosphonium or sulfonium cation, X is generally a halide anion (usually Cl− ). They are based on equilibria set up between X− and a Lewis or Brønsted acid Y, z refers to the number of Y molecules which complex X− . The ionic liquids described can be subdivided into three types depending on the nature of the complexing agent used.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.3 Eutectic-based Ionic Liquids

Fig. 2.5 Schematic representation of the complexation occurring when a Lewis acid or a Brønsted acid interacts with a quaternary ammonium salt. Eutectic Type 1 Y = MClx, M = Zn [95–98], Sn [96], Fe [96], Al [94], Ga [99], In [100] Eutectic Type 2 Y = MClx·yH2 O, M = Cr [101], Co, Cu, Ni, Fe Eutectic Type 3 Y = RZ, Z = CONH2 [102], COOH [103], OH [104]

To date the only Cat+ species studied have been based on pyridinium, imidazolium and quaternary ammonium moieties. In general, as with the chloroaluminate and discrete anion systems, the imidazolium-based liquids have the lowest freezing points and viscosities and higher conductivities. The depression of freezing point is related to the strength of interaction between the anion and complexing agent although this has not really been quantified as yet due primarily to a lack of thermodynamic data for the individual components. One of the key advantages of these types of ionic liquids is the ease of manufacture. The liquid formation is generally mildly endothermic and requires simply mixing the two components with gentle heating. Another key advantage is that they are water insensitive which is very important for practical electroplating systems. As will be shown in Chapter 6.3, the electrochemistry of metals is relatively unaffected by relatively large concentrations of water either naturally absorbed or deliberately added to the ionic liquids. The final key advantage of eutectic-based systems is that because they are simple mixtures of known chemicals they do not have to be registered as new entities as they revert to their constituent components upon excessive dilution in water. 2.3.1 Type I Eutectics

An extensive range of metal salts [96] have been studied but the only ones which produce ionic liquids (i.e. liquid below 100 ◦ C) with pyridinium, imidazolium and quaternary ammonium halides are FeCl3 , ZnCl2 , SnCl2 , CuCl [105], InCl3 [100] and AuCl3 [106, 107]. It is thought that the ability of a metal salt to form a low melting point ionic liquid will be related to its own melting point. The reason for this is apparent from Figure 2.4. Hence aluminum chloride (mp 190 ◦ C) has been shown to be useful with a wide range of quaternary ammonium salts. Table 2.5 shows that relatively low melting points are also possessed by ZnCl2 , SnCl2 and FeCl3 . Metal salts that do not form ionic liquids with ammonium salts tend to have high melting points resulting from large lattice energies. It is generally true that the metals have linear or tetrahedral

33

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

34 2 Synthesis of Ionic Liquids Table 2.5 Freezing temperature data for a variety of metal salts and

amides when mixed with choline chloride in 2:1 ratio.

Type I

Type II

Type III

a

ZnCl2 SnCl2 FeCl3 CrCl3 .6H2 O MgCl2 .6H2 O CoCl2 .6H2 O LaCl3 .6H2 O CuCl2 .2H2 O Urea 1, Methyl urea, 1,3 Dimethyl urea 1,1 Dimethyl urea Thiourea Acetamide Benzamide

T f /◦ C

T f * /◦ C

T f /◦ C

24 37 65 4 10 16 6 48 12 29 70 149 69 51 92

283 246 306 83 117a 86 91 100a 134 93 102 180 175 80 129

259 209 241 79 107 70 85 52 122 64 32 31 106 29 37

Denotes decomposition temperatures.

geometries and tend to form predominantly univalent anionic complexes. In the cases of FeCl3 , ZnCl2 and SnCl2 a variety of complex anions are known to form whereas for CuCl, InCl3 , AuCl3 and TeCl4 only monometallate anions are known to form i.e. CuCl2 − , InCl4 − , AuCl4 − and TeCl6 2− . For the zinc chloride: choline chloride mixtures the eutectic is observed at a 2:1 composition, whereas for the tin chloride: choline chloride mixtures it is observed at 2.5:1. This is presumably because SnCl2 is less Lewis acidic than ZnCl2 and hence more SnCl2 is require to push the equilibrium for the reaction SnCl2 + SnCl3 −  Sn2 Cl5 − to the optimum Sn2 Cl5 − composition. The ZnCl2 system has probably been studied in the most detail. Fast atom bombardment mass spectrometry (FAB MS) has been used to identify the species present. It was found that ZnCl3 − , Zn2 Cl5 − and Zn3 Cl7 − species are all present in the liquids. The relative proportions of anionic species depend on the ionic liquid composition. Lecocq et al. [108] used electrospray ionization to look at the various species present and found that in Lewis basic liquids x(ZnCl2 ) < 0.5 ZnCl3 − whereas the di- and tri-metallate species were more prevalent in Lewis acidic liquids. Presumably, small changes in concentration of each of the complex anions change the ion–ion interactions markedly and this in turn changes the freezing point. For example, ZnCl3 − ions are smaller and have a higher charge density than Zn2 Cl5 − anions so are likely to have stronger electrostatic interactions with the cation thus increasing the freezing point. Hence, as the mole fraction of ZnCl2 increases from 50% the amount of Zn2 Cl5 − relative to ZnCl3 − should increase and the freezing point decreases. Above a mole fraction of 66% ZnCl2 the freezing point increases again.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.3 Eutectic-based Ionic Liquids

The composition of the various chlorozincate anions in the Lewis acid ionic liquids was determined using potentiometry in an analogous manner to that used by Heerman and D’Olislager [109] who measured the potential of the cell Al|BuPyCl, AlCl3 (ref )||AlCl3 (x), BuPyCl(1 − x)|Al They found that the equilibrium constant for the process 2Al2 Cl7 − DAlCl4 − + Al3 Cl10 − was 2.93 × 10−3 at 60 ◦ C i.e. Al2 Cl7 − is the most abundant species in solution. The cell Zn|ZnCl2 (0.667) ChCl (0.333)||ZnCl2 (x) ChCl (1 − x)|Zn was used to determine an equilibrium constant of 2.0 × 10−5 for the reaction → ZnCl3 − + Zn3 Cl7 − 2Zn2 Cl5 −← The value is lower than that for the analogous aluminum case, which would be expected because of the difference in Lewis acidity. Hence the main species at the eutectic composition was found to be Zn2 Cl5 − [96]. Liu et al. [110] studied the crystal structures of chlorozincate–choline chloride complexes and identified that in an equimolar ratio the liquid is made up of two species. It was shown that two types of crystal could be grown from the super-cooled liquid, one rod-like and the other sheet-like, and these were thought to be due to the Zn2 Cl5 − and ZnCl3 − salts, respectively. 13 C and 35 Cl NMR spectra of [BMIM]Cl and [BMIM]ZnCl3 showed that at 25 ◦ C there is a significant difference between the two systems whereas at 110 ◦ C the systems are similar; this showed that the zinc-containing liquid is highly associated at lower temperatures. A more dissociated structure is favored at high temperatures. This is significant for metal deposition studies as the coordination geometry will affect the way in which the metal is reduced. The phase behavior of ionic liquids will depend upon the potential energy between the ions but this is difficult to model for a eutectic-based ionic liquid because of the complex nature of the anion and the non-centrosymmetric charge distribution on the cation. However, if the difference in freezing point between that of the quaternary ammonium salt and that of the complex with the metal salt is considered then the issue becomes significantly easier. The change in interionic potential energy, E p , and the resulting change in freezing temperature, T f will be related to the expansion of the ionic lattice resulting from the formation of a complex anion. Since Ep =

q1q2 4πεo r

(2.1)

where q is the charge on the ions, ε o is the vacuum permittivity and r is the separation between the two charges. Therefore E p ∝

rc − rs rs

(2.2)

35

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

36 2 Synthesis of Ionic Liquids

where r c and r s are the charge separation in the complex mixture and simple quaternary ammonium halide salt, respectively (assuming that at the eutectic composition the complex anion is predominantly of the form M2 X5 − ). The depression of freezing point, T f , is taken as the difference between the measured freezing point at the eutectic composition, T f and the freezing temperature of the pure quaternary ammonium halide. It was shown [96] that by plotting the freezing point depression as a function of the normalized change in charge separation (Eq. (2.2)) produces a good correlation for ZnCl2 -based ionic liquids. This is a significant result as it allows phase behavior to be predicted from simple ionic size considerations and it shows that symmetry has a negligible effect on the depression of freezing point, but it does change the absolute freezing point. Angell [111] recently used a similar approach to show that the glass transition temperature of a range of ionic liquids is related to the molar volume of the ions. The effect of the quaternary ammonium cations is quite complex because the smaller cations depress the freezing point more because the halide salts of the smaller cations also have a higher freezing point; the net result is that all of the eutectic mixtures will have reasonably similar freezing points. Hence the cation is observed to have little effect on the absolute freezing point of the eutectic-based ionic liquids. Lecocq et al. [108] studied ionic liquids formed between zinc chloride and 1-butyl-2,3-dimethylimidazolium chloride [BMMIM]Cl with the amount of ZnCl2 between 0 and 0.75 mol%. Analysis using NMR, and mass spectrometry showed Cl− and [ZnCl3 ]− in Lewis basic liquids and [ZnCl3 ]− and [Zn3 Cl7 ]− in Lewis acidic liquids. Infrared spectra with pyridine were used to quantify the Lewis acidity and high temperature (110 ◦ C) NMR experiments showed that the structure varies with time from [BMMIM][ZnCl3 ] to [BMMIM. . .Cl. . .ZnCl2 ]. The iron-based systems have two eutectic points in an analogous manner to the chloroaluminate systems. The eutectic points occur at 33 and 67 mol% FeCl3 [96]. We were only able to identify the species FeCl4 − by FAB MS in the choline chloride–FeCl3 system but this could be because other species are too weak to be observed by this technique. Other groups have prepared iron-containing liquids with FeCl2 and FeCl3 . Sitze et al. [112] found that [BMIM]Cl formed liquids with FeCl2 in the molar ratio 0.3 FeCl2 : 1 [BMIM]Cl whereas the ferric chloride formed in the molar ratio 0.53 to 1.7. Raman scattering and ab initio calculations showed that FeCl4 2− was the prevalent anion present with ferrous chloride, whereas FeCl4 − and Fe2 Cl7 − were present in the ferric chloride system. The relative concentrations were dependent upon the Lewis acidity in an analogous manner to the zinc and aluminum systems. Zhang et al. [113] also studied FeCl3 and 1-methyl-3butylimidazolium chloride ([BMIM]Cl) with a molar ratio of 1:1 and characterized the physical properties of the liquid. Hayashi et al. [114] also studied the [BMIM] FeCl4 system and found that the liquid is ferromagnetic. The ability to vary the composition of Lewis or Brønsted acid adds an additional dimension to the tuneability of the eutectic-based ionic liquids. It has been shown that the Lewis acidity of the liquid affects not only the physical properties of the liquids but also the electrochemical behavior. Type I ionic liquids are also clearly

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.3 Eutectic-based Ionic Liquids

Scheme 2.6

useful for electroplating if the metal of interest falls in the category defined above as the metal ion concentration can be as high as 10 mol dm−3 . Eutectic mixtures of imidazolium chloride with GaCl3 and InCl3 have also been reported [99, 100]. These metals will be of limited interest for electrodeposition although some studies have been made on the deposition of semiconductors [115, 116]. Other metal halides that have been used include AuCl3 , NiCl2 and CoCl2 [106, 107]. These tend to have higher melting points than other metal salts for the reasons explained above. They have been used for synthetic applications and while, in principle, they could be used for electrodeposition there are better alternatives that would be more suitable. Seddon et al. [117] have produced ionic liquids of the type [EMIM]2 [UCl6 ] from the reaction of UCl4 with [EMIM]Cl. The uranium was isolated using electrochemical reduction and it was proposed that this was a potential method for recycling spent nuclear fuel. Hagiwara [118] produced ionic liquids containing niobium and tantalum from the reaction of [EMIM]F.2.3HF with TaF5 and NbF5 to produce [EMIM] TaF6 and [EMIM] NbF6 . While the majority of studies in this area have concentrated on halide salts some intriguing work has been carried out using metal oxides. Noguera et al. [119] showed that CrO3 and Na2 MoO4 could be incorporated into ionic liquids. Scheme 2.6 shows the synthesis of two ionic liquids and although the electrodeposition of the metal was not reported it could, in principle, be used for such applications. These liquids have been shown to be good oxidants for organic reactions. A number of other strategies have been published for the production of metal-containing ionic liquids and while most of these are very exotic and have been used for catalysis some of the generic methodologies may eventually find application in electrodeposition. This area has recently been reviewed by Lin and Vasam [120]. The conductivities of Type I ionic liquids based on anhydrous zinc and iron salts tend to be lower than those of the corresponding aluminum ionic liquids. This is due largely to the higher viscosity of the former, primarily because of the large size of the ions and the availability of suitably sized holes in the ionic liquids for the ions to move into. This has been quantified by the application of hole theory as is explained in Section 2.3.4. In general imidazolium-based liquids have lower viscosities and higher conductivities than the corresponding pyridinium or quaternary ammonium eutectics formed under the same conditions.

37

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

38 2 Synthesis of Ionic Liquids

2.3.2 Type II Eutectics

These were developed in an endeavor to expand the range of metals that could be incorporated into an ionic liquid. The presence of waters of hydration decreases the melting point of metal salts because it decreases the lattice energy. Hence, as Figure 2.4 shows, hydrated salts should be more likely to form mixtures with quaternary ammonium salts that are liquid at ambient temperature than anhydrous salts. Table 2.5 shows a list of some of the metal salts that have been made into ionic liquids with choline chloride and the freezing point of a 1ChCl:2metal salt mixture. Electrospray MS of the eutectic mixture showed two primary signals M+ 104 [Choline]+ and M− 192/194* /196 [CrCl4 ]− (the waters of hydration are bound too weakly to be observed and Cr(H2 O)3 Cl3 is neutral and therefore not detected) [101]. The UV–vis spectrum of the eutectic mixture showed the presence of predominantly Cr(H2 O)3 Cl3 with some evidence of [CrCl4 ·2H2 O]− . It was concluded that the main charge carrying species were [Choline]+ and [Cl·3H2 O]− . This would account for the high conductivities of these liquids compared to the anhydrous salt mixtures. The addition of LiCl to the ionic liquid was found to have only a small effect upon the conductivity of the liquid, but it did affect the speciation [121], producing more of the [CrCl4 ·2H2 O]− . It was anticipated that the small Li+ ion would have a high mobility in the liquid but the conductivity is less than expected, suggesting that the ion must be strongly solvated or highly associated with the anion. Unlike the anhydrous metal salts, these mixtures are very sensitive to temperature fluctuations. At ambient temperatures they are extremely hygroscopic and rapidly absorb up to 10 wt% water from the atmosphere. Above 70 ◦ C the liquids lose water and this is characterized by a change in color of the chromium-based liquid from dark green to purple. At about 50 to 60 ◦ C the water concentration in the liquid remains constant and can be used in an open atmosphere without significant alteration in the liquid composition. Thermogravimetry shows that the waters of hydration are released in two steps; the first starts at about 85 ◦ C, which equates to approximately 3 waters, and the second at about 180 ◦ C, corresponding to the other 3 water molecules [101]. To date the only concerted study has been carried out using chromium chloride, but it has been reported that a number of other metals form this type of eutectic mixture and Table 2.5 lists just some of the metal salts that have been studied, together with their freezing points in eutectic mixtures with choline chloride. Potentially there are some very interesting systems but to date only Cr and Co have been deposited from these liquids. The deposition of metals such as Al and Ca is not possible due to the limited potential window of these liquids. 2.3.3 Type III Eutectics

It has recently been shown that the principle of creating an ionic fluid by complexing a halide salt can be applied to mixtures of quaternary ammonium salts with a

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.3 Eutectic-based Ionic Liquids

range of amides [102, 103]. The charge delocalization is achieved through hydrogen bonding between the halide anion with an amide, carboxylic acid or alcohol moiety. Table 2.5 lists the freezing points of a number of hydrogen bond donor (HBD) mixtures with choline chloride. These liquids have interesting solvent properties that are similar to other eutectic ionic liquids and a wide variety of solutes were found to exhibit high solubilities [102, 103]. The depression of freezing point (with respect to an ideal mixture of the two components) for a number of these eutectic systems is extremely large e.g. the oxalic acid–choline chloride system, is 212 ◦ C and the choline chloride–urea system is 178 ◦ C [103]. The freezing point depressions are not as large as the choline chloride–zinc chloride system (272 ◦ C) [97] due to the covalent bonds formed in the metal chloride case. To differentiate these liquids from ionic liquids the term Deep Eutectic Solvents (DES) has been adopted. Unlike the room-temperature ionic liquids, these eutectic mixtures are easy to prepare in a pure state. They are non-reactive with water, many are biodegradable and the toxicological properties of the components are well characterized. It is thought that the chloride complexes with 2 HBDs and this accounts for the varying eutectic composition. For monofunctional HBDs e.g. urea, phenylpropionic acid, the eutectic point occurs at 67 mol% HBD, for difunctional HBDs, e.g. oxalic acid and malonic acid, the eutectic point occurs at 50 mol% HBD and for citric acid the eutectic occurs at 33 mol% HBD. The tris-carboxylic acids exhibit the rheology of gels and presumably have extensive bridging of the acids between neighboring chloride ions. The existence of hydrogen bonding in ChCl/urea eutectic mixtures can be observed using NMR spectroscopy [102]. Heteronuclear Overhauser effect spectroscopy (HOESY) of HOCH2 CH2 N + (CH3 )3 F · 2(NH2 )2 CO shows intense cross-correlation between the fluoride ion and the NH2 protons on the urea molecule. Some anion complexes have been identified using FAB MS and it is evident that the HBD is sufficiently strongly coordinated to the chloride anion to be detected by this technique. In a 1 choline chloride: 2 urea mixture the presence of Cl− with two ureas (M− = 155) and Cl− with one urea (M− = 95) was observed. As with the chlorometallate eutectics a model for the effect of HBD on the freezing point depression of the mixture would be beneficial for the design of new liquids. No correlations were observed between the freezing point of the mixtures and the enthalpy of formation or fusion of the pure acids but Table 2.5 shows qualitatively that the larger depressions of freezing point occur with the lower molecular weight HBDs. The freezing point of the HBD–salt mixtures will be dependent upon the lattice energies of the salt and HBD and how these are counteracted by the anion–HBD interaction and the entropy changes arising from forming a liquid. For a given quaternary ammonium salt, the lattice energy of the HBD will be related to the anion–HBD interaction and hence, to a first approximation, the depression of freezing point will be a measure of the entropy change. It has been shown [103] that the depression of freezing point correlates well with the mass fraction of HBD in the mixture. The lowest viscosities and highest conductivities are obtained with diol-based HBDs. It is thought that the comparatively weak interactions between the alcohol

39

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

40 2 Synthesis of Ionic Liquids Table 2.6 Viscosity and conductivity of a variety of ionic liquids at

298 K. Cation

Anion

EMIM EMIM BMIM BuMePy choline choline choline choline choline acetylcholine choline choline

BF4 − N(CF3 SO2 )2 BF4 − N(CF3 SO2 )2 Zn2 Cl5 − CrCl4 ·6H2 O CoCl3 ·6H2 O Cl·2urea Cl·2propanediol Cl·2propanediol Cl·malonic acid Cl·2ethylene glycol

j/mS cm−1

14 8.4 3.5 2.2 0.02 0.37 1.7 0.75 2.2 0.51 0.36 7.6

g/cP

32 28 180 85 76 000 2346 392 632 89 117 3340 36

and the chloride mean that some ‘free’ glycol is able to move, decreasing the viscosity of the liquid. The glycol-based liquids tend also to have comparatively large potential windows. Hence the Abbott group has carried out a number of studies using ethylene glycol with choline chloride. This mixture has been shown to be useful for the deposition of zinc and zinc alloys [122] as well as the electropolishing of stainless steel [104] (see Chapter 11.1). The liquid is inexpensive, non-toxic, non-viscous and highly conducting compared to other ionic liquids. 2.3.4 Modelling Viscosity and Conductivity

One of the main differences between ionic liquids and aqueous solutions is the comparatively high viscosity of the former. Table 2.6 shows that viscosities are typically in the range 10–500 cP (0.01–0.50 Pa s) and this affects the diffusion coefficients of species in solution. Most new liquids have viscosity that varies as a function of temperature and the majority vary in an Arrhenius manner with temperature [123]: ln η = ln η0 +

Eη RT

(2.3)

where Eη is the activation energy viscous flow and η0 is a constant. Other researchers have found that the viscosity obeys a Vogel–Tamman–Fulcher relationship [124]. A comprehensive study of viscosity is that of VanderNoot [124] and there are several collections of viscosity data in recent reviews [125–127]. We have fitted the viscosity of ionic liquids using hole theory [123]. The theory was developed for molten salts but has been shown to be very useful for ionic liquids. It was shown that the value of E η is related to the size of the ions and the size of the voids present in the liquid [103]. The viscosity of ionic liquids is

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

2.3 Eutectic-based Ionic Liquids

several orders of magnitude higher than that of high-temperature molten salts due partially to the difference in size of the ions, but also to the increased void volume in the latter. It has been shown [128] that hole theory can be applied to both ionic and molecular fluids to account for viscosity. The viscosity of a fluid, η, can be modeled by assuming it behaves like an ideal gas, but its motion is restricted by the availability of sites for the ions/molecules to move into. Hence it was shown that η=

m¯c /2.12σ P(r > R)

(2.4)

where m is the molecular mass (for ionic fluids this was taken as the geometric mean), c¯ is the average speed of the molecule (=(8kT/πm)1/2 ) and σ is the collision diameter of the molecule (4πR2 ). The probability of finding a hole of radius, r, greater than the radius of the solvent molecule, R, in a given liquid, (P(r > R)) is given by integration of the following expression [123]: Pdr =

16 7/2 6 −ar 2 dr √ a r e 15 π

(2.5)

where a = 4πγ /kT and γ is the surface tension. The good correlation obtained between the calculated and measured viscosities shows that it is valid to think of the viscosity of fluids as being limited by the availability of holes. It is evident from Eqs. (2.4) and (2.5) that decreased viscosity can be obtained by decreasing the surface tension of the liquid, i.e. increasing the free volume, or by decreasing the ionic radius. Hence the ionic liquids with the lowest viscosity tend to have highly fluorinated anions as these shield the charge density and result in low surface tensions. The cation also affects the viscosity of ionic liquids. For imidazolium cations, the viscosity initially decreases as the length of the R group increases, as the ion–ion interactions decrease and hence the surface tension decreases. However, as the alkyl group increases in size its mobility will decrease due to a lack of suitably sized voids for the cations to move into. This can be seen in the data presented by Tokuda et al. who showed a minimum in viscosity for ethyl methyl imidazolium salts [129]. The conductivity of ionic liquids can be modeled in the same manner as the viscosity, i.e. despite the high ionic strength of the liquid, ionic migration is limited by the availability of suitably sized voids [130]. Since the fraction of suitably sized holes in ambient temperature ionic liquids is effectively at infinite dilution, migration should be described by a combination of the Stokes–Einstein and Nernst–Einstein equations. This is explained in greater detail in Chapter 11.3 on process scale-up but it is sufficient to say that an expression can be derived for the conductivity, κ κ=

z2 F e 6πη



1 1 + R+ R



ρ Mw

(2.6)

where ρ is the density and Mw is the molar mass of the ionic fluid. Hence the molar conductivity ( = κ/c) is, in effect, independent of the number of charge carriers and this is the reason why the empirical Walden rule [123, 126] ( η = constant)

41

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

42 2 Synthesis of Ionic Liquids

is applicable to ionic liquids. The Walden rule is normally only valid for ions at infinite dilution where ion–ion interactions can be ignored, which is clearly not the case in ionic liquids. It is apparent from the above discussion that ionic mobility is controlled by the free volume of a liquid and the size of the ions. The size of the voids in the liquid and their effect on liquid density can be changed by decreasing the ion–ion interactions. This will manifest itself by a decrease in surface tension and, in general, the liquids with lower surface tensions are more fluid and have higher conductivities. This is the reason why ionic liquids with discrete, highly fluorinated anions such as PF6 and (F3 CSO2 )2 N have become popular. It has recently been shown that the same principle can be applied to deep eutectic solvents by using small quaternary ammonium cations such as ethylammonium and fluorinated hydrogen bond donors such as trifluoroacetamide. However, there is only a limited benefit that can be achieved using this approach as the physical parameters cannot be varied totally independently of one another. For example there will be an optimum ion size; too small and the lattice energy will increase the surface tension, too large and the ionic mobility will be impeded. 2.3.5 Conclusions

This chapter shows that eutectic-based ionic liquids can be made in a variety of ways. The above description of liquids falling into three types is by no means exclusive and will certainly expand over the coming years. While there are disadvantages in terms of viscosity and conductivity these are outweighed for many metal deposition processes by issues such as cost, ease of manufacture, decreased toxicity and insensitivity to moisture. The high viscosity of some of these liquids could be ameliorated in many circumstances by the addition of inert diluents. The physical principles underlying eutectic-based ionic liquids are now relatively well understood, however, the liquids described above have tended to be less academically fashionable and have received comparatively little attention. Concerted effort with these types of liquids could lead to optimization of their properties such that they would be suitable for commercial deposition processes. References 1 Walden, P. (1914) Bull. Acad. Imper. Sci. (St. Petersburg), 1800. 2 Hurley, F.H. and Wier, T.P. (1951) J. Electrochem. Soc., 98, 203. 3 Hurley, F.H. and Wier, T.P. (1951) J. Electrochem. Soc., 98, 207. 4 Wier, T.P. and Hurley, F.H., US Patent, 2446349. 5 Wier, T.P., US Patent, 2446350. 6 Hurley, F.H., US Patent, 2446331.

7 Gale, R.J., Gilbert, B., and Osteryoung, R.A. (1978) Inorg. Chem., 17, 2728. 8 Robinson, J. and Osteryoung, R.A. (1979) J. Am. Chem. Soc., 101, 323. 9 Gale, R.J. and Osteryoung, R.A. (1979) Inorg. Chem., 17, 1603. 10 Wilkes, J.S., Levisky, J.A., Wilson, R.A., and Hussey, C.L. (1982) Inorg. Chem., 21, 1263.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

References 11 Fannin, A.A. Jr., Floreani, D.A., King, L.A., Landers, J.S., Piersma, B.J., Stech, D.J., Vaughn, R.L., Wilkes, J.S., and Williams, J.L. (1984) J. Phys. Chem., 88, 2614. 12 Wasserscheid, P. and Welton, T. (2003) Ionic Liquids in Synthesis, Wiley-VCH, Verlag GmbH. 13 Gordon, C.M., Holbrey, J.D., Kennedy, A.R., and Seddon, K.R. (1998) J. Mater. Chem., 8, 2627. 14 Visser, A.E., Holbrey, J.D., and Rogers, R.D. (2001) Chem. Comm., 2484. 15 MacFarlane, D.R., Meakin, P., Sun, J., Amini, N., and Forsyth, M. (1999) J. Phys. Chem. B, 103, 4164. 16 Sun, J., Forsyth, M., and MacFarlane, D.R. (1998) J. Phys. Chem. B, 102, 8858. 17 (a) Johnson, W.A. (1993) Ylids and Imines of Phosphorus, John Wiley and Sons, New York; (b) Hartley, F.R. (ed.) (1994) The Chemistry of Organophosphorus Compounds Vol. 3: Phosphonium Salts, Ylids and Phosphoranes, John Wiley and Sons, New York. 18 Paulsson, H., Hagfeldt, A., and Kloo, L. (2003) J. Phys. Chem. B, 107, 13665. 19 Armarego, W.L.F. and Perrin, D.D. (1996) Purification of Laboratory Chemicals, 4th edn, Butterworth-Heinemann. 20 Bonhˆote, P., Dias, A.P., Papageorgiou, N., Kalyanasundaram, K., and Gr¨atzel, M. (1996) Inorg. Chem., 35, 1168. 21 Huddleston, J.G., Willauer, H.D., Swatloski, R.P., Visser, A.E., and Rogers, R.D. (1998) Chem. Comm., 1765. 22 Lucas, P., El Mehdi, N., Ho, H.A., B´elanger, D., and Breau, L. (2000) Synthesis, 9, 1253. 23 Varma, R.S. and Namboodiri, V.V. (2001) Chem. Comm., 643. 24 Holbrey, J.D., Seddon, K.R., and Wareing, R. (2001) Green Chem., 3, 33–36. 25 Beyersdorff, T. and Schubert, T.J.S., IOLITEC, unpublished results. 26 Beyersdorff, T. and Schubert, T.J.S., IOLITEC; unpublished results. 27 Beyersdorff, T. and Schubert, T.J.S., IOLITEC, unpublished results. 28 Beyersdorff, T. and Schubert, T.J.S., IOLITEC, unpublished results.

29 Welton, T. (1999) Chem. Rev., 99, 2071. And references therein. 30 (a) Chauvin, Y., Eiloft, S., and Olivier, H. (1995) Ind. Eng. Chem. Res., 34, 1149; (b) Gilbert, B., Chauvin, Y., Olivier, H., DiMarco-van Tiggelen, F. (1995) J. Chem. Soc., Dalton Trans., 3867. 31 Sanders, J.R., Ward, E.H., and Hussey, C.L. (1986) J. Electrochem. Soc., 133, 325. 32 Williams, S.D., Schoebrechts, J.P., Selkirk, J.C., and Mamantov, G. (1987) J. Am. Chem. Soc., 109, 2218. 33 Chauvin, Y. and Olivier-Bourbigou, H. (1995) CHEMTECH, 25, 26. 34 Parshall, G.W. (1972) J. Am. Chem. Soc., 94, 8716. 35 (a) Sitze, M.S., Schreiter, E.R., Patterson, E.V., and Freeman, R.G. (2001) Inorg. Chem., 40, 2298; (b) Hayashi, S. and Hamaguchi, H. (2004) Chem. Lett., 33, 1590. 36 Carpio, R.A., King, L.A., Lindstrom, R.E., Nardi, J.C., and Hussey, C.L. (1979) J. Electrochem. Soc., 126, 1644. 37 Hitchcock, P.B., Seddon, K.R., and Welton, T. (1993) J. Chem. Soc., Dalton Trans., 2639. 38 Hitchcock, P.B., Lewis, R.J., and Welton, T. (1993) Polyhedron, 12, 2039. 39 Noguera, G., Mostany, J., Agrifoglio, G., and Dorta, R. (2005) Adv. Synth. Catal., 347, 231. 40 Namboodiri, V.V. and Varma, R.S. (2002) Chem. Comm., 342. 41 Zawodzinski, T.A. and Osteryoung, R.A. (1990) Inorg. Chem., 29, 2842. 42 (a) Abdul-Sada, A.K., Avent, A.G., Parkington, M.J., Ryan, T.A., Seddon, K.R., and Welton, T. (1987) Chem. Commun., 1643; (b) Abdul-Sada, A.K., Avent, A.G., Parkington, M.J., Ryan, T.A., Seddon, K.R., and Welton, T. (1993) J. Chem. Soc., Dalton Trans., 3283. 43 Dent, A.J., Lees, A., Lewis, R.J., and Welton, T. (1996) J. Chem. Soc., Dalton Trans., 2787. 44 Hurley, F.H. and Wier, T.P. (1951) J. Electrochem. Soc., 98, 203–206. 45 Hurley, F.H. and Wier, T.P. (1951) J. Electrochem. Soc., 98, 207–212. 46 Wilkes, J.S. (2002) Green Chem., 4, 73–80.

43

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

44 2 Synthesis of Ionic Liquids

47 Wilkes, J.S. and Zaworotko, M.J. (1992) Chem. Commun., 965–967. 48 Walden, P. (1914) Bull. Acad. Imper. Sci. St. Petersburg, 405–427. 49 Hirao, M., Sugimoto, H., and Ohno, H. (2000) J. Electrochem. Soc., 147, 4168–4172. 50 Fuller, J., Carlin, R.T., and Osteryoung, R.A. (1997) J. Electrochem. Soc., 144, 3881–3886. 51 Fuller, J., Carlin, R.T., De Long, H.C., and Haworth, D. (1994) J. Chem. Soc., Chem. Commun., 299–300. 52 Chauvin, Y., Mussmann, L., and Olivier, H. (1995) Angew. Chem. Int. Ed. Engl., 34, 2698–2700. 53 Noda, A., Hayamizu, K., and Watanabe, M. (2001) J. Phys. Chem. B, 105, 4603–4610. 54 Golding, J., Hamid, N., MacFarlane, D.R., Forsyth, M., Forsyth, C., Collins, C., and Huang, J. (2001) Chem. Mater., 13, 558–564. 55 Forsyth, S., Golding, J., MacFarlane, D.R., and Forsyth, M. (2001) Electrochim. Acta, 46, 1753–1757. 56 Villagr´an, C., Deetlefs, M., Pitner, W.R., and Hardacre, C. (2004) Anal. Chem., 76, 2118–2123. 57 Meindersma, G.W., Podt, A., and de Haan, A.B. (2006) J. Chem. Eng. Data, 51, 1814–1819. 58 Carlin, R.T., De Long, H.C., Fuller, J., and Trulove, P.C. (1994) J. Electrochem. Soc., 141, L73–L76. 59 Papageorgiou, N., Athanassov, Y., Armand, M., Bonhˆote, P., Pettersson, H., Azam, A., and Gr¨atzel, M. (1996) J. Electrochem. Soc., 143, 3099–3108. 60 Bonhˆote, P., Dias, A.-P., Papageorgiou, N., Kalyanasundaram, K., and Gr¨atzel, M. (1996) Inorg. Chem., 35, 1168– 1178. 61 Hagiwara, R. and Ito, Y. (2000) J. Fluorine Chem., 105, 221–227. 62 Dahn, J.R., Fong, R., and Spoon, M.J. (1990) Phys. Rev. B, 42, 6424–6432. 63 Sun, J., Forsyth, M., and MacFarlane, D.R. (1998) J. Phys. Chem. B, 102, 8858–8864. 64 MacFarlane, D.R., Sun, J., Golding, J., Meakin, P., and Forsyth, M. (2000) Electrochim. Acta, 45, 1271–1278.

65 Xu, W., Cooper, E.I., and Angell, C.A. (2003) J. Phys. Chem. B, 107, 6170– 6178. 66 Howlett, P.C., MacFarlane, D.R., and Hollenkamp, A.F. (2004) Electrochem. Solid–State Lett., 7, A97–A101. 67 Ue, M., Murakami, A., and Nakamura, S. (2002) J. Electrochem. Soc., 149, A1572–A1577. 68 Earle, M.J., Esperanc¸a, J.M.S.S., Gilea, M.A., Lopes, J.N.C., Rebelo, L.P.N., Magee, J.W., Seddon, K.R., and Widegren, J.A. (2006) Nature, 439, 831–834. 69 Rebelo, L.P.N., Lopes, J.N.C., Esperanc¸a, J.M.S.S., and Filipe, E. (2005) J. Phys. Chem. B, 109, 6040–6043. 70 Chan, S.S. and Willis, C.J. (1968) Can. J. Chem., 46, 1237–1248. 71 Schmidt, M., Heider, U., Kuehner, A., Oesten, R., Jungnitz, M., Ignat’ev, N., and Sartori, P. (2001) J. Power Sources, 97–98, 557–560. 72 Ignat’ev, N.V., Welz–Biermann, U., Kucheryna, A., Bissky, G., and Willner, H. (2005) J. Fluorine Chem., 126, 1150–1159. 73 Liu, Q., Janssen, M.H.A., van Rantwijk, F., and Sheldon, R.A. (2005) Green Chem., 7, 39–42. 74 Swatloski, R.P., Spear, S.K., Holbrey, J.D., and Rogers, R.D. (2002) J. Am. Chem. Soc., 124, 4974–4975. 75 Vijayaraghavan, R. and MacFarlane, D.R. (2004) Aust. J. Chem., 57, 129–133. 76 Wooster, T.J., Johanson, K.M., Fraser, K.J., MacFarlane, D.R., and Scott, J.L. (2006) Green Chem., 8, 691–696. 77 Pringle, J.M., Golding, J., Forsyth, C.M., Deacon, G.B., Forsyth, M., and MacFarlane, D.R. (2002) J. Mater. Chem., 12, 3475–3480. 78 Kumara, G.R.A., Konno, A., Shiratsuchi, K., Tsukahara, J., and Tennakone, K. (2002) Chem. Mater., 14, 954–955. 79 Meng, Q.-B., Takahashi, K., Zhang, X.-T., Sutanto, I., Rao, T.N., Sato, O., Fujishima, A., Watanabe, H., Nakamori, T., and Uragami, M. (2003) Langmuir, 19, 3572–3574. 80 Wang, P., Zakeeruddin, S.M., Humphry-Baker, R., and Gr¨atzel, M. (2004) Chem. Mater., 16, 2694–2696.

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

References 81 Nockemann, P., Thijs, B., Postelmans, N., Van Hecke, K., Van Meervelt, L., and Binnemans, K. (2006) J. Am. Chem. Soc., 128, 13658–13659. 82 MacFarlane, D.R., Golding, J., Forsyth, S., Forsyth, M., and Deacon, G.B. (2001) Chem. Commun., 1430–1431. 83 MacFarlane, D.R., Forsyth, S.A., Golding, J., and Deacon, G.B. (2002) Green Chem., 4, 444–448. 84 Barisci, J.N., Wallace, G.G., MacFarlane, D.R., and Baughman, R.H. (2004) Electrochem. Commun., 6, 22–27. 85 Wang, P., Zakeeruddin, S.M., Moser, J.-E., and Gr¨atzel, M. (2003) J. Phys. Chem. B, 107, 13280–13285. 86 Kawano, R., Matsui, H., Matsuyama, C., Sato, A., Susan, Md.A.B.H., Tanabe, N., and Watanabe, M. (2004) J. Photochem. Photobiol. A, 164, 87–92. 87 Yoshida, Y., Muroi, K., Otsuka, A., Saito, G., Takahashi, M., and Yoko, T. (2004) Inorg. Chem., 43, 1458–1462. 88 Wang, P., Wenger, B., Humphry-Baker, R., Moser, J.-E., Teuscher, J., Kantlehner, W., Mezger, J., Stoyanov, E.V., Zakeeruddin, S.M., and Gr¨atzel, M. (2005) J. Am. Chem. Soc., 127, 6850–6856. 89 Bessler, E. and Goubeau, J. (1967) Z. Anorg. Allg. Chem., 352, 67–76. 90 Bessler, E. (1977) Z. Anorg. Allg. Chem., 430, 38–42. 91 Bernhardt, E., Finze, M., and Willner, H. (2003) Z. Anorg. Allg. Chem., 629, 1229–1234. 92 Welz-Biermann, U., Ignat’ev, N., Bernhardt, E., Finze, M., and Willner, H. (2004) German Pat. DE 10306617/A1. 93 Kuang, D., Wang, P., Ito, S., Zakeeruddin, S.M., and Gr¨atzel, M. (2006) J. Am. Chem. Soc., 128, 7732–7733. 94 Hurley, F.H. and Weir, T.P. (1951) J. Electrochem. Soc., 98, 207. 95 Abbott, A.P., Capper, G., Davies, D.L., Munro, H., Rasheed, R., and Tambyrajah, V. (2001) Chem. Commun., 2010. 96 Abbott, A.P., Capper, G., Davies, D.L., Munro, H., and Rasheed, R. (2004) Inorg. Chem., 43, 3447. 97 Hsiu, S.-I., Huang, J.-F., Sun, I.-W.,

98 99

100

101

102

103

104

105 106

107

108

109 110 111 112

113

114 115 116

Yuan, C.-H., and Shiea, J. (2002) Electrochim. Acta, 47, 4367–4372. Lin, Y.-F. and Sun, I.-W. (1999) Electrochim. Acta, 44, 2771. Xu, W.G., Lu, X.-M., Yang, J.Z., Gui, J.-S., and Yang, J.Z. (2006) Chinese J. Chem., 24, 331–335. Yang, J.Z., Xu, W.G., Tian, P., and He, L.-L. (2003) Fluid Phase Equilib., 204, 295–302. Abbott, A.P., Capper, G., Davies, D.L., and Rasheed, R. (2004) Chem. Eur. J., 10, 3769. Abbott, A.P., Capper, G., Davies, D.L., Rasheed, R., and Tambyrajah, V. (2003) Chem. Commun, 70. Abbott, A.P., Boothby, D., Capper, G., Davies, D.L., and Rasheed, R. (2004) J. Am. Chem. Soc., 126, 9142. Abbott, A.P., Capper, G., Swain, B.G., and Wheeler, D.A. (2005) T. Inst. Met. Fin. 83, 51. Bolkan, S.A. and Yoke, J.T. (1986) Inorg. Chem., 25, 3587. Schreiter, E.R., Stevens, J.E., Ortwerth, M.F., and Freeman, R.G. (1999) Inorg. Chem., 38, 3935. Hasan, M., Kozhevnikov, I.V., Siddiqui, M.R.H., Femoni, C., Steiner, A., and Winterton, N. (1999) Inorg. Chem., 38, 5637. Lecocq, V., Graille, A., Santini, C.C., Baudouin, A., Chauvin, Y., Basset, J.M., Arzel, L., Bouchu, D., and Fenet, B. (2005) New J. Chem., 29, 700–706. Heerman, L. and D’Olislager, W. (1985) Inorg. Chem., 24, 4704. Liu, Y., Wu, G., and Qi, M. (2005) J. Cryst. Growth, 281, 616. Xu, W. and Angell, C.A. (2003) Science, 203, 422. Sitze, M.S., Schreiter, E.R., Patterson, E.V., and Freeman, R.G. (2001) Inorg. Chem., 40, 2298. Zhang, Q.-G., Yang, J.Z., Lu, X.-M., Gui, J.-S., and Huang, M. (2004) Fluid Phase Equil., 226, 207–211. Hayashi, S. and Hamaguchi, H. (2004) Chem. Lett., 33, 1590–1591. Verbrugge, M.W. and Carpenter, M.K. (1990) AIChE J., 36, 1097. Carpenter, M.K. and Verbrugge, M.W. (1987) J. Electrochem. Soc., 87, 591.

45

c02 (JWBG008-Endres)

December 25, 2007

13:16

Char Count=

46 2 Synthesis of Ionic Liquids

117 Hitchcock, P.B., Mohammed, T.J., Seddon, K.R., and Zora, J.A. (1986) Inorg. Chim. Acta, 113, L25. 118 Matsumoto, K., Hagiwara, R., and Ito, Y. (2002) J. Fluorine Chem., 115, 133. 119 Noguera, G., Mostany, J., Agrifoglio, G., and Dorta, R. (2005) Adv. Synth. Catal., 347, 231–234. 120 Lin, I.J.B. and Vasam, C.S. (2005) J. Organomet. Chem., 690, 3498–3512. 121 Abbott, A.P., Capper, G., Davies, D.L., Rasheed, R.K., Archer, J., and John, C. (2004) Trans. Inst. Metal Finish, 82, 14. 122 Abbott, A.P., Capper, G., McKenzie, K.J., and Ryder, K.S. (2007) J. Electroanal. Chem., 599, 288. 123 O’Bockris, J.M. and Reddy, A.K.N. (1970) Modern Electrochemistry, Vol. 1, Plenum Press, New York, Chapter 6.

124 Okoturo, O.O. and Van der Noot, T.J. (2004) J. Electroanal. Chem., 568, 167. 125 Wasserscheid, P. and Welton, T. (2003) Ionic Liquids in Synthesis, Wiley-VCH Verlag GmbH. 126 Galinski, M., Lewandowski, A., and Stepniak, I. (2006) Electrochim Acta, 51 5567–5580. 127 Crosthwaite, J.M., Muldoon, M.J., Dixon, J.K., Anderson, J.L., and Brennecke, J.F. (2005) J. Chem. Thermodyn., 35, 559– 568. 128 Abbott, A.P. (2004) Chem. Phys. Chem., 5, 1242. 129 Tokuda, H., Hayamizu, K., Ishii, K., Abu Bin Hasan Susan, M., and Watanabe, M. (2005) J. Phys. Chem. B, 109, 6103. 130 Abbott, A.P. (2005) Chem. Phys. Chem., 6, 2404.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

47

3 Physical Properties of Ionic Liquids for Electrochemical Applications Hiroyuki Ohno

3.1 Introduction

In spite of the explosion in studies on ionic liquids (ILs), there is only a small number of studies of their basic characteristics. There are limitless possibilities for the design of ILs by changing their component ion structures. However, the chance of success is not very great without accurate information on the structure–properties relationship. Physico-chemical property data for ILs are therefore very important for the present and future of the field of ILs. In this chapter, some basic properties of airstable ILs have been summarized. Some are not directly related to electrochemistry but are very important and useful for a wide range of science and technology related to ILs. 3.2 Thermal Properties

ILs are defined as organic salts having a melting point (T m ) below 100 ◦ C [1–5]. In order to use these ILs as non-volatile electrolyte solutions, it is necessary to maintain the liquid phase over a wide temperature range. Consequently, T m and the thermal degradation temperature (T d ) of ILs are important properties for ILs as electrochemical media. In this section, the thermal properties of ILs, especially of imidazolium salts, are summarized. The difference between ILs and general electrolyte solutions based on molecular solvents is clarified. Recent results on the correlation between the structure and properties of ILs will also be mentioned. 3.2.1 Melting Point

ILs are differentiated from typical inorganic salts by their low T m . Typical inorganic salts have a high T m , around 1000 ◦ C reflecting high lattice energies, i.e., the high T m is attributable to a strong electrostatic attractive force between the ions. Since ILs are organic compounds, van der Waals interaction, hydrogen bonding, and π–π Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

48 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.1 Melting point (◦ C) of several salts

Cation

Na+ (1.02 Å) K+ (1.38 Å) Cs+ (1.67 Å) [N1111 ]+ [N2222 ]+ (3.35 Å) [emim]+ (3.04 Å)

Anion Cl− (1.81 Å)

Br− (1.96 Å)

I− (2.20 Å)

BF4 − (2.29 Å)

808 772 645 420 >300 87

747 734 636 230a 285 77

662 685 621 230[a] 300 78

384 530

72 11

(CF3 SO2 )2 N− (3.25 Å)

104 −15

Ion radius is given in parentheses. a Decomposition temperature, [N1111 ]+ , tetramethylammonium cation; [N2222 ]+ , tetraethylammonium cation; [EMIM]+ , 1-ethyl-3-methylimidazolium cation.

interaction are additionally present among the component ions. These interactions affect the T m of ILs. Accordingly, structural design of component ions to weaken the electrostatic interaction and other interactions is directly effective in lowering the T m of the salts, as discussed in detail below. However, it is still difficult to predict the T m of any given salt from its structure. 3.2.1.1 Effect of Ion Radius When ions have equivalent charges, the electrostatic interaction decreases with increasing ion radius because the surface charge density decreases with increasing ion radius and the separation between the ions also increases. The electrostatic interaction of larger ions is therefore weaker, and accordingly the salts show lower T m . Table 3.1 shows the T m of typical salts and their ion radii [6–9]. In general, organic salts have lower T m than inorganic salts because of their larger ion size, as shown in Table 3.1. However, the tetraethylammonium cation ([N2222 ]+ ) is larger than 1-ethyl-3-methylimidazolium cation ([EMIM]+ ), yet the [N2222 ]+ salts show higher T m than [EMIM]+ salts. This is a typical example that illustrates the case where T m depends not only on the electrostatic interaction. Since the delocalization of charge also contributes to lowering the electrostatic interaction, the existence of π electron orbitals is important in lowering the T m . 3.2.1.2 Effect of Cation Structure on the Melting Point Onium Cations. Major families of ILs are composed of quaternary onium cations such as imidazolium, pyridinium, ammonium, phosphonium, sulfonium cations and so on. As described above, the fact that most ILs are composed of organic cations is attributed to weaker electrostatic interaction among component ions. There have been several reports on the effect of cation structure on the T m of ILs. The relationship between T m and the basic structure of onium cations is important in developing a protocol to prepare low melting ILs. The cation structure and T m of bis(trifluoromethanesulfonyl)imide type ILs are shown in Table 3.2. These cations, having similar length of alkyl chain, are chosen

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.2 Thermal Properties Table 3.2 Melting point (T m ) of a series of

bis(trifluoromethanesulfonyl)imide type ionic liquids Cation

T m /◦ C

Ref.

−15

7

−4

10

20

11

86

12

−18

12

44

13

29.2

14

28.7

14

104

15

90

15

for comparison to exclude the effect of the alkyl chain on their T m s. Among the salts in Table 3.2, the imidazolium and pyridinium cations are aromatic and the others are aliphatic. These aromatic salts show relatively low T m because of delocalized positive charge, as mentioned previously. Effect of Side Chain Length. The side chain bound to the cation also affects the T m due to flexibility and excluded volume effects. To discuss the relation between side chain structure and T m , the onium cation structure was fixed. The relation between the side chain structure and the thermal properties of imidazolium salts has already been reported by Seddon et al. [16, 17]. The effect of alkyl chain length of 1-alkyl3-methylimidazolium tetrafluoroborate on the phase transition temperatures is

49

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

50 3 Physical Properties of Ionic Liquids for Electrochemical Applications

Fig. 3.1 Phase diagram for 1-alkyl-3-methylimidazolium tetrafluoroborate showing the melting (•), glass (◦) and clearing () transitions measured by differential scanning calorimetry. Data from Ref. [16].

shown in Figure 3.1 [16]. With a carbon number from 2 to 9 on the imidazolium ring, the salts are liquid at room temperature. On the other hand, when the carbon number of the alkyl chain of the imidazolium ring was 0, 1, or larger than 9, the salts showed a clear T m . A liquid crystalline phase appeared for imidazolium salts with a carbon chain longer than the dodecyl group. This liquid crystalline phase arises due to the orientational effect of the long alkyl chains. A similar tendency has also been observed in the case of PF6 imidazolium salts [17]. Symmetry is another factor to affect T m . The salts with symmetric ions generally show higher T m than those with asymmetric ones. For example, 1,3dimethylimidazolium tetrafluoroborate showed higher T m than 1-methylimidazolium or 1-ethyl-3-methylimidazolium salts, as shown in Figure 3.1. In the case of tetraalkylammonium salts, their T m also increased with increasing symmetry of the cation structure [18]. This tendency is understood to relate to the structural effect on crystallinity [19], i.e., highly symmetric ions are more efficiently packed into the crystalline structure than unsymmetric ones. Other kinds of chain structures such as polyether [20], perfluorocarbon [21], etc. [22] are obviously also effective in influencing thermal properties. 3.2.1.3 Anion Species There are many choices of anion species for IL synthesis. In particular, halogencontaining anions such as BF4 − , PF6 − , and TFSI− are often used to prepare ILs. Room-temperature ILs are obtained with ions having weaker electrostatic interaction originating from negative charge delocalization and stabilization by the electron-withdrawing effect of halogen atoms. Non-halogenated anion-containing ILs with low T m have also been prepared after suitable structural design to lower the anionic charge density [23]. Thermal properties of such 1-ethyl-3methylimidazolium and 1-butyl-3-methylimidazolium salts are summarized in Table 3.3. Larger anions generally form ILs with lower T m . Lowering the surface

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.2 Thermal Properties Table 3.3 Thermal properties of imidazolium-type ionic liquids com-

posed of various anions Salt Tm / ◦C

Tg / ◦C

Td / ◦C

[Cl]− [Br]− [I]− [BF4 ]− [PF6 ]− [NO3 ]− [CH3 COO]− [CF3 COO]− [CH3 SO3 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [(C2 F5 SO2 )2 N]− [(CN)2 N]− [(CF3 SO2 )3 C]− [(CN)3 C]−

8915 7915 7915 1115 , 1524 6215 , 5825 117 , 3826 4515 –146 3927 –96 –157 –115 –2123a) 3915 –1128

–– –– –– –8624 –– –– ––

28515

–– –987

–9528

44010 4557 42315 –– 45015 ––

[Cl]− [Br]− [I]− [BF4 ]− [PF6 ]− [NO3 ]− [CH3 COO]− [CF3 COO]− [CH3 SO3 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [(C2 F5 SO2 )2 N]− [(CN)2 N]− [(CF3 SO2 )3 C]− [(CN)3 C]−

6510 –– –– –8129 –810 ,1027

–– –5031 –– –9731 –8031

25031 27332 26515 40331 34931

––

–7832

22035 17632

1610 –410

–– –8733

–628 ––

–9031 –6531

Cation

Anion

[EMIM]+

[BMIM]+

–10423a)

30315 42034 –– –– –– 15010

40932 43931 40232 30031 41331

[EMIM]+ , 1-ethyl-3-methylimidazolium cation; [BMIM]+ , 1-butyl-3-methylimiadzolium cation. Reference numbers are shown as superscripts to the data.

charge densities of the anion also lowers the T m of the ILs. In spite of the fact that most anions used are symmetric, there are a few approaches where T m has been lowered by using asymmetric anions [36]. 3.2.2 Glass Transition Temperature

The glass transition temperature (T g ) is generally understood to be the temperature where segmental motion begins on heating from the quenched amorphous solid.

51

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

52 3 Physical Properties of Ionic Liquids for Electrochemical Applications

Accordingly, the ionic conductivity and viscosity of many ILs are a function of T g . The T g is thus not so important for electrodeposition from ILs, but is important for ion conduction of and in the ILs (see ionic conductivity). In the case of ILs, there are many examples that show both T g and T m . Detailed phase studies have been reported elsewhere [16]. The relation between T g and T m has already been discussed by Angell et al. [37] who observed that T g is almost equal to two-thirds of T m in Kelvin. 3.2.3 Thermal Decomposition Temperature

ILs are thermally stable but certainly decompose at high temperature. The decomposition temperature (T d ) of general imidazolium type ILs is summarized in Table 3.3. The T d of ILs depends on the component ion structure, similarly to other thermal properties [30]. ILs having excellent thermal stability up to 400 ◦ C have been reported [15,31,34]. However, this does not mean that these ILs can be used at any temperature below T d , because most T d values are determined by using temperature sweeping thermogravimetric measurements. ILs gradually decompose even below T d . It is therefore important to analyze the thermal behavior at constant temperature. Previously reported T d values of several imidazolium-type ILs are plotted against side chain length (n) in Figure 3.2. T d is affected by water content, impurities, type of flow gases, and vessel material for thermal gravimetric measurements [15]; hence it should be noted here that the T d shown in Figure 3.2 is not the absolute value for each IL. As shown in Figure 3.2, the ILs composed of BF4 − , PF6 − and TFSI− have T d 100 ◦ C higher than ILs composed of halogen anions such as Cl− or I− . Chan et al. reported that an alkyl chain at the N-position of the imidazolium cation suffers nucleophilic attack by the halide anion in the manner shown in Scheme 3.1 [38]. While remarkable differences in T d are observed by changing the anion species, T d only depends slightly on the alkyl chain length on the imidazolium cation.

Fig. 3.2 Relation between thermal decomposition temperature (T d ) of 1-alkyl-3-methylimidazolium-type ILs with alkyl chain length (n). Anion species are Cl− : (), I− : (•), BF4 − : (), PF6 − : () and TFSI− : (♦).

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.2 Thermal Properties

Scheme 3.1 Pyrolysis mechanism of ionic liquids containinig halide anions.

There is a report that the thermal stability of imidazolium salts is largely the same, despite a great difference in alkyl chain length between the butyl and octadecyl groups, suggesting simple extension of alkyl chain hardly affects T d [30]. 3.2.4 Liquid Crystallinity and Solid–Solid Transitions

Some compounds show meso-phases between the solid and liquid phases. These phases are classified into two kinds, namely liquid crystals in which the molecules have orientational order and disorganized position in one or more dimensions, and plastic crystal in which the molecules have organized positions and orientational disorder. Although the component ions in ILs are largely disordered, the appearance of liquid crystalline or plastic crystal phases could be the function of ion structures, when component ions have a tendency towards orientational or positional ordering by alignment of the ions and/or interaction among ions. Onium salt-type plastic crystals have been reported by MacFarlane [12,39]. A series of 1-alkyl-3-methylimidazolium salts show liquid crystalline phases by elongation of the alkyl side chain as shown in Figure 3.1. Seddon et al. reported the liquid crystalline phase for pyridinium salts with longer alkyl chains, as well as imidazolium salts [16, 17,40]. Introduction of a hydrophobic moiety into the anion is also effective in yielding liquid crystalline properties in imidazolium salts. However, it should be noted here that not all hydrophobic anions show liquid crystal phases. The salts composed of an imidazolium cation having multiple methyl groups and long chain alkylsulfonate anions showed liquid crystalline properties, depending on the position and quantity of substituent groups on the imidazolium ring [41]. These materials have been discussed as anisotropic ion conductors [42] and anomalous reaction media [43]. It might be interesting to examine electrodeoposition in these materials for super-fine surface designs such as parallel nanowires. 3.2.5 Thermal Conductivity

The thermal conductivity of ILs is an important property when using ILs for electrochemical synthesis or thermal storage. The thermal conductivity of ILs was reported, together with heat capacity, by Wilkes et al., as summarized in Table 3.4 [44]. The heat capacities of ILs are 3 or 4 times larger than that of copper, but smaller than that of water. The thermal conductivity of general ILs is lower than that of cop R per or water. Therminol VP-1, diphenyl oxide/biphenyl type thermal conductor, is commercially available as a heat transport fluid. The thermal conductivity and heat capacity of ILs are, in general, similar to those of VP-1.

53

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

54 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.4 Heat capacity and thermal conductivity of ionic liquids

 R Therminol VP-1 [EMIM][BF4 ] [BMIM][BF4 ] [p-DMIM][(CF3 SO2 )2 N] H2 O at 30 ◦ C H2 O at 100 ◦ C copper

Heat capacity at 100 ◦ C J g−1 K−1

Thermal conductivity at 25 ◦ C W m−1 K−1

Ref.

1.78 1.28 1.66±0.08 1.20±0.05 4.18 4.22 0.385

0.127 0.200±0.003 0.186±0.001 0.131±0.001 0.615 0.679 398

44 44 44 44 45 45 46

[p-DIMIM][(CF3 SO2 )2 N]: 1-propyl-2,3-dimethylimidazolium bis(trifluoromethanesulfonyl)imide.

3.2.6 Vapor Pressure

The vapor pressure of ILs is substantially zero under ambient conditions. Therefore ILs have been recognized as non-volatile liquids at normal pressures. However, it is known experimentally that some ILs, synthesized by the neutralization of protic acid with organic base, easily evaporate on heating. Angell et al. pointed out that the acid–base equilibrium of those ILs becomes imbalanced on heating and then generates the volatile acid and base [47]. Based on this, MacFarlane et al. recently reported that the ILs prepared by neutralization, N-methylpyrrolidinium formate, could be distilled 100% at 70 ◦ C under 0.9 mmHg [48]. Distillable ILs can, therefore, be prepared by neutralization of volatile bases with volatile acids. On the other hand, in general, ILs composed of quaternized onium cations and anions do not show such an equilibrium. These ILs are generally decomposed on heating without evaporation. Recently, Seddon et al. reported that many known ILs including [EMIM][TFSI] can be evaporated at 300 ◦ C under high vacuum (less than 0.1 mbar) [49]. Details of the evaporation mechanism are not yet clear; a cluster ion model is proposed because it is hardly conceivable that individual anions and cations are vaporized, even under high vacuum.

3.3 Viscosity

Viscosity is an important property of ILs used as electrolyte solutions. There are some basic studies on the viscosity of ILs in the literature [50, 51]. The reported viscosities of imidazolium type ILs composed of commercially available anions are relatively low, as summarized in Table 3.5. The reported viscosity values are not always the same for any given IL owing to water content, impurities, synthetic route, starting materials, and measurement method.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.4 Density 55 Table 3.5 Viscosity of several liquids at room temperature (25 ◦ C ± 1)

[EMIM]+

[BMIM]+

[HMIM]+ [OMIM]+

[BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]− [(CF3 CF2 SO2 )2 N]− [CF3 CO2 ]− [CF3 SO3 ]− [BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]− [(CF3 CF2 SO2 )2 N]− [CF3 CO2 ]− [CF3 SO3 ]− [BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]− [BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]− [(CF3 CF2 SO2 )2 N]−

g / cP

Ref.

43 15 (80 ◦ C) 28 61 35 45 (30 ◦ C) 219 450 69 77 70 93 314 (20 ◦ C) 585 68 439 682 93 492

7 7 7 7 6 54 30 30 30 54 35 55 56 30 35 57 30 58 59

g / cP

water methanol acetic acid acetone acetonitrile N,N-dimethylformamide ethylene glycol propylene glycol glycerol

0.89 0.54 1.13 0.30 0.34 0.80 16.1 40.4 934

[HMIM]+ : 1-hexyl-3-methylimidazolium, [OMIM]+ : 1-octyl-3-methylimidazolium.

The viscosity of ILs is typically 10 to 100 times higher than that of water or organic solvents [50–52] as a result of the strong electrostatic and other interaction forces. The fluorohydrogenate type ILs reported by Hagiwara et al. have some of the lowest viscosities known [53]. Low viscosity ILs are obviously preferred in electrolyte or other reaction solvent applications, but it is quite difficult to design low viscosity ILs. The imidazolium ILs tend to show decreasing viscosity in the following order of anion species; PF6 − , BF4 − , and TFSI− , depending on the alkyl side chain length. In addition, CF3 CO2 − and CF3 SO3 − anions tend to form relatively low viscosity ILs. There are only a few studies of ILs containing Cl− or Br− anion [22], because these ILs are not in the liquid state at room temperature. Since viscosity is directly affected by electrostatic interaction, it is expected that ILs composed of larger ions or charge delocalized ions should show lower viscosity. The degree of dissociation of salts is another important factor.

3.4 Density

Table 3.6 summarizes the densities of various ILs. Since ILs are composed only of ions, almost all ILs are denser than water, from 1.0 to 1.6 g cm−3 depending on their ion structure. The densities of some complex salts are even higher than ordinary ILs. Details will be given elsewhere.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

56 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.6 Density of several ionic liquids

Cation

Anion

[EMIM]+

[NO3 ]− [BF4 ]− [PF6 ]− [CF3 COO]− [C3 F7 COO]− [CH3 SO3 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [(C2 F5 SO2 )2 N]− [(CN)2 N]− [(CN)3 C]− [PF6 ]− [BF4 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [Cl]− [PF6 ]− [(CF3 SO2 )2 N]−

[BMIM]+

[HMIM]+

q /g cm−3 Ref.

1.28 1.28 1.56 1.29 1.45 1.25 1.38 1.46 1.52 1.08 1.11 1.37 1.21 1.30 1.43 1.03 1.31 1.37

Cation

60 [OMIM]+ 24a 24a 6 6 [b-diMIM]+ 60 6 [bpyr]+ 6 32 [N3111 ]+ 60 [N4111 ]+ 28 [N6222 ]+ 60 [N8222 ]+ 60 [P14 ]+ 6 [P13 ]+ 6 [S111 ]+ 60 [S222 ]+ 60 [S444 ]+ 60

Anion

q /g cm−3 Ref.

[BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]−

1.12 1.23 1.31

60 60 60

[BF4 ]− [PF6 ]− [BF4 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]−

1.20 1.35 1.22 1.45 1.44 1.39 1.27 1.25 1.41 1.44 1.58 1.46 1.29

60 60 24a 24a 61 62 18 18 12 12 63 63 63

[b-diMIM]+ : 1-butyl-2,3-dimethylimidazolium, [bpyr]+ : butylpyridinium.

The density has been found to decrease with increasing alkyl chain length on the imidazolium cation [33]. Similarly, in the ammonium and sulfonium salts, the density decreases with increasing alkyl chain length. This clearly shows that the charged ion unit is heavier than the hydrocarbon chain. Accordingly, the density of ILs is tunable to some extent. The density of aromatic onium salts is higher than that of aliphatic ammonium salts. Generally, density decreases in the order of pyridinium salts > imidazolium salts > aliphatic ammonium salts and piperidinium salts. The densities of ILs are also affected by the anion species. Similarly to the trends for cations, the density of ILs decreases with increasing alkyl chain length of the anion. The density of ILs is increased on the introduction of a “heavy” chain such as fluoroalkyl chains. For example, 1-ethyl-3-methylimidazolium (EMIM) salts became heavier with the following anion species; CH3 SO3 − < BF4 − and CF3 COO− < CF3 SO3 − < (CF3 SO2 )2 N− < (C2 F5 SO2 )2 N− . It is easy to understand this order as an effect of formula weight of the ions. However, these tendencies are still empirical, and a perfect correlation between ion structure and density is not yet available.

3.5 Refractive Index

In the field of processing or engineering, there is a potential requirement for materials with high refractive index. However, these materials are typically all solid and those liquids that are known are poisonous. Accordingly liquids having

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.5 Refractive Index 57 Table 3.7 Refractive index of ionic liquids

n

T/◦ C

Ref.

[CH3 CO2 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [Br]− [I]− [CH3 CO2 ]− [BF4 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [Cl]− [Cl]− [BF4 ]− [BF4 ]−

1.4405 1.4332 1.4231 1.54 1.572 1.4887 1.42 1.4380 1.4271 1.515 1.505 1.4322 1.4367

20 20 20 25 25 20 25 20 20 25 25 25 25

6 6 6 30 64 6 64 6 6 30 30 65 65

[IBr2 ]− [BrI2 ]− [I5 ]− [Br3 ]− [IBr2 ]− [BrI2 ]− [I5 ]− [I7 ]− [I9 ]− [IBr2 ]− [I3 ]−

1.715 1.833 2.23 1.699 1.701 1.805 2.16 2.3 2.4 1.685 1.88

–– –– –– –– –– –– –– –– –– –– ––

67 67 67 67 67 67 67 67 67 67 67

Cation

Anion

[EMIM]+ [BMIM]+

[HMIM]+ [OMIM]+ [decyl-MIM]+ [EMIM]+ [BMIM]+

[HMIM]+

[decyl-MIM]+ : 1-decyl-3-methylimidazolium cation.

high refractive index with low toxicity are highly desirable. The refractive index of water or methanol is ca. 1.33 at room temperature, whereas that of ILs is 1.4 or more. Higher values (ca. 1.5) have been found in the ILs composed of halogenated anions. Table 3.7 summarizes the refractive index for imidazolium salts composed of various anions. From this table, it is clear that refractive index of ILs increases with increase in the alkyl chain length of the imidazolium cation. Moreover, the refractive index is strongly affected by the anion; the refractive index of ILs with larger anions, such as [(CF3 SO2 )2 N]− (TFSI anion) is lower than that of ILs having smaller anions such as acetate or halide anions. As mentioned before, the densities of ILs depend on the ion structure. There is also a tendency for ILs with lower density to show higher refractive index. However, it was reported that the ILs containing complex metal anions ([BMIM][Ln(NCS)x (H2 O)y ] show relatively high refractive index (ca. 1.57) [66]. In the case of these ILs, a clear correlation between refractive index and density is observed; higher refractive index is found for the ILs with higher density (Figure 3.3). This tendency is the opposite of what occurs with common ILs.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

58 3 Physical Properties of Ionic Liquids for Electrochemical Applications

Fig. 3.3 Plot of refractive indices (n) vs. density (D) for a series of ionic liquids.

Furthermore, Seddon et al. reported that the poly-halide salts, such as [EMIM][IBr2 ] or [EMIM][I5 ], have a high refractive index of 1.6 or more, as shown in Table 3.7 [67]. The high refractive indices of the lanthanide salts and the heavy halogens and their trihalide salts are well predictable from their polarizabilities, which in turn are well understood on the basis of periodic table trends: atoms/ions with partly filled 4f, 5d etc. shells tend to be quite polarizable and hence have high refractive indices.

3.6 Polarity

Polarity is one of the most important parameters of ILs for its effect on electrochemical reactions. It is important when we characterize ILs to measure not only thermal properties such as melting point but also solvent properties such as polarity [68–70]. The most common method of polarity measurement is a dielectric constant measurement. Weingartner et al. and Hefter et al. have shown, by applying appropriately high frequency methods, that the dielectric constants are uniformly around 10–15. Accordingly, the polarity of ILs should be estimated by other methods. Solvatochromism is heavily applied for this purpose due to its simplicity. 3.6.1 Solvatochromism

Solvatochromism is the shift of the maximum absorption wavelength of dye molecules depending on the polarity of the solvents. The advantage of this method

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.6 Polarity 59

is the small amount of sample required for spectroscopic measurement. A review by Reichardt introduced over 80 probe molecules [71]. 3.6.2 Reichardt’s Betaine Dye

There are many kinds of probe molecules for estimation of polarity. Among them, the most widely used dye is Reichardt’s dye (2,4,6-triphenylpyridinium-N-4-(2,6diphenylphenoxide) betaine) [72]. Both empirical scales for polarity; E T (30) and E T N are frequently used for polarity studies in ILs. The solvatochromism of the Reichardt’s dye is based on the interaction of the solvent with the ground state of the dye. Kamlet and Taft proposed that about two-thirds of the shift of the maximum absorption wavelength of Reichardt’s Dye could be assigned to the interactions involving the phenoxide oxygen with the solvent [73]. E T (30) is estimated by Eq. (3.1). Reichardt and Harbusch-Gornert have defined an E T N value according to Eq. (3.2) as a dimensionless figure, using water and tetramethylsilane (TMS) as references of extreme polar and non-polar solvents, respectively. Hence, the E T N scale ranges from 0 to 1 [74]. E T (30) = 28591.5/λmax E TN

= [E T (solvent) − 30.7]/32.4

(3.1) (3.2)

The E T (30) (in kcal mol−1 ) and E T N scales of ILs are summarized in Table 3.8. The E T (30) for 1-methyl-3-alkylimidazolium type ILs is about 51–53, similar to that of methanol and ethanol. The alkyl chain length and the nature of the anion have no influence on the E T (30). For the E T (30) values of alkylimidazoliumtype ILs, substitution of C-2 proton with a methyl group lowered the E T (30) to 48–49, which is similar to that of octanol or isopropanol. On the other hand, [HO(CH2 )2 MIM][(CF3 SO2 )2 N], which contains a hydroxy group, has a high E T (30) value (61.4). This suggests that the C-2 proton on the imidazolium ring shows high acidity/hydrogen bond donor capability; furthermore the hydroxy group on the side chain shows even higher acidity/ hydrogen bond donor capability than the C-2 proton. For the aliphatic cations, the E T (30) decreased in the following order: primary > tertiary > quaternary. Additionally, for quaternary ammonium cations, the values decrease with increasing cation size. Generally, E T (30) values are dominated by the nature of the cations. 3.6.3 Kamlet–Taft Parameter

Kamlet–Taft parameters are known to express three distinct measures of the solvent polarity such as dipolarity/polarizability (π * ), hydrogen-bond acidity (α) and hydrogen-bond basicity (β). These parameters have been determined by absorption measurements for individual or pairs of the following dye molecules; N,N-diethyl4-nitroaniline, 4-netroaniline and Reichardt’s dye, as seen in Figure 3.4 [81–83].

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

60 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.8 The value of ET (30) and ET N for some ionic liquids

Cation [EMIM]+ [PMIM]+ [BMIM]+

[HMIM]+ [OMIM]+ [DMIM]+ [b-diMIM]+ [BZMIM]+ [OH(CH2 )2 -MIM]+ [P14 ]+ n-butylammonium sec-butylammonium dipropylammonium ethylammonium n-propylammonium tributylammonium tetrabutylammonium tetrapropylammonium tetrapentylammonium water methanol ethanol acetonitrile acetone dichloromethane toluene hexane dimethyl sulfoxide

Anion

ET (30)

ET N

Ref.

[BF4 ]− [(CF3 SO2 )2 N]− [BF4 ]− [(CF3 SO2 )2 N]− [BF4 ]− [PF6 ]− [TfO]− [(CF3 SO2 )2 N]− [SbF6 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [BF4 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [SCN]− [SCN]− [SCN]− [NO3 ]− [NO3 ]− [NO3 ]− [CHES]− [CHES]− [(CF3 SO2 )2 N]−

53.7 52.9 53.1 52.0 52.5 52.3 51.2 51.5 52.4 52.3 51.9 48.3 51.2 51.1 51 49.4 48.6 52.5 61.4 48.3 61.4 61.6 63.3 61.6 60.6 56.7 48 51 44 63.1 55.5 51.9 46 42.2 40.7 33.9

0.71 0.69 0.69 0.65 0.67 0.67 0.66 0.64 0.67 0.67 0.65 0.54 0.63 0.63 0.63 0.58 0.54 0.67 0.95 0.54 0.95 0.95 1.01 0.95 0.92 0.8 0.62 0.53 0.41 1.00 0.77 0.65 0.47 0.36 0.309 0.1 (0.009 0.44

75 76 75 58 77 77 77 77 77 77 77 79 77 77 58 78 77 58 58 78 79 79 79 79 58 79 80 80 69

45

[OMIM]+ : 1-octyl-3-methylimidazolium, [DMIM]+ : 1-decyl-3-methylimidazolium, [b-diMIM]+ : 1-butyl-2,3-dimethylimidazolium, [BZMIM]+ : 1-benzyl-3-methylimidazolium, [OH-EMIM]+ : 1-hydroxyethyl-3-methylimidazolium, [CHES]− : 2-(cyclohexylamino)ethanesulfate

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.6 Polarity 61

Fig. 3.4 Solvatochromic probe molecules.

N,N-Diethyl-4-nitroaniline, has an aromatic ring but no hydrogen bond donor substituent, shows a π–π * transition based on a non-specific interaction between ions. Dipolarity/polarizability, π * , is estimated by the solvatochromic shift of N,Ndiethyl-4-nitroaniline using Eq. (3.3), where λmax is the absorption maximum for N,N-diethyl-4-nitroaniline. π ∗ = 8.649 − 0.314ν1 (ν1 = 1/(λmax × 10−4 ))

(3.3)

4-Nitroaniline can interact with solvent molecules with an amino group at the C-1 position as a proton donor. The λmax shows a red shift when it interacts with a solvent having a hydrogen bond acceptor group. The β value (hydrogen bond basicity) is estimated with Eq. (3.4).using spectral data of both N,N-diethyl-4-nitroaniline and 4-nitroaniline. B = (1.035ν2 − ν1 + 2.64)/2.80

(3.4)

Reichardt’s dye has thus been used to estimate hydrogen-bond acidity of solvents. The absorption maximum of Reichardt’s dye shows a blue shift when the solvent molecule interacts with the dye through a hydrogen bond. The α value (hydrogenbond acidity) is estimated using E T (30) and π * with Eq. (3.5). α = 0.0649E T (30) − 0.72π ∗ − 2.03

(3.5)

Welton reported the effect of cations and anions on the Kamlet–Taft Parameters [78]. The Kamlet–Taft parameters for some ILs are summarized in Table 3.9. As seen, π * values for these ILs are high, 0.9–1.3, in comparison with those for protic molecular solvents as shown in the same table. Both cation and anion affect the π * value. For anions, the π * value for ILs having TFSI anion is low due to weakened coulombic interaction caused by delocalized anionic charge. The β values of ILs are mainly governed by the nature of the anions. They decrease in the order [Cl]− > [RSO3 ]− > [BF4 ]− > [PF6 ]− . On the other hand, α values of ILs are largely affected by the nature of the component cations, especially the presence of hydrogen-bond donor groups. The nature of the anion seldom affects the α value [78].

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

62 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.9 Kamlet–Taft parameters for typical ionic liquids

Cation

[EMIM]+ [BMIM]+

[HMIM]+ [b-diMIM]+ [P14 ]+ n-butylammonium sec-butylammonium dipropylammonium ethylammonium n-propylammonium tributylammonium water methanol ethanol acetonitrile acetone dichloromethane toluene hexane dimethyl sulfoxide

Anion

[(CF3 SO2 )2 N]− [BF4 ]− [Cl]− [PF6 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [SbF6 ]− [(CF3 SO2 )2 N]− [BF4 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [SCN]− [SCN]− [SCN]− [NO3 ]− [NO3 ]− [NO3 ]−

Kamlet-Taft parameters

Ref.

p*

α

β

0.980 1.047 1.17 1.032 1.006 0.984 1.04 0.971 1.083 1.01 0.954 1.23 1.28 1.16 1.24 1.17 0.97

0.705 0.627 0.41 0.634 0.625 0.617 0.64 0.259 0.402 0.381 0.427 0.92 0.91 0.97 0.85 0.88 0.84

0.233 0.376 0.95 0.207 0.464 0.243 0.15 0.650 0.363 0.239 0.252

1.09 0.6 0.54 0.75 0.71 0.791 0.532 (–0.12) 1

1.17 0.93 0.83 0.19 0.08 0.042 –0.213 (0.07) 0

0.18 0.62 0.77 0.31 0.48 –0.014 0.077 (0.04) 0.76

0.39 0.46 0.52

76 58 68 77 77 77 78 76 78 77 77 79 79 79 79 58 79

3.6.4 Acetylacetonatotetramethylethyldiaminecopper (II)

[Cu(acac)(tmen)]BPh4 , is known to provide a good correlation between the donor number (DN) of the solvent and the λmax corresponding to the lowest energy of d–d transition [84]. In spite of the small number of experiments, there is a certain relation between anion species and λmax , as shown in Table 3.10. 3.6.5 Pyrene

Pyrene is one of the most widely studied neutral fluorescence probes, and accordingly, this was sometimes used to determine the polarity of some ILs. The polarity scale of the IL analyzed with pyrene is defined as the emission intensity

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.6 Polarity 63 Table 3.10 Polarity of ionic liquids determined by Nile Red and

[Cu(acac)(tmen)]+ BPh4 − Cation

Anion

[Cu(acac)(tmen)] [BPh4 ]

Nile Red kmax [EMIM]+ [PMIM]+ [BMIM]+

[HMIM]+ [OMIM]+

[DMIM]+ [BZMIM]+ [HMIM]+ [HEIM]+ [HBIM]+ water methanol ethanol acetonitrile acetone hexane DMSO

[BF4 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [BF4 ]− [PF6 ]− [CF3 SO3 ]− [(CF3 SO2 )2 N]− [NO3 ]− [BF4 ]− [PF6 ]− [NO3 ]− [BF4 ]− [PF6 ]− [(CF3 SO2 )2 N]− [NO3 ]− [(CF3 SO2 )2 N]− [BF4 ]− [(CF3 SO2 )2 N]− [BF4 ]− [BF4 ]− [BF4 ]−

ENR

Ref.

550.3 550.8 547.5

52.3 51.9 52.2

88 88 88

548.7 555.7 551.9 551.7 552.9 549.5 549.8

52.1 51.5 51.8 51.8 51.7 52 52

88 88 88 88 88 88 88

550.1 560.5 545.7 546 562.3 562.9 562.8 584.5 542.9 539.8 520.7

217.4 52.2 52.4 51.8 50.9 50.8 50.8 48.2 52 52.2 53.8

88 88 88 88 88 89 89

544.8

59 52

kmax

Ref.

541 547

69 69

517 516.5 546 602

77

517 549

77 77

77

573 569

ratio “II /IIII ”, where band I corresponds to an S1 (ν = 0)→S0 (ν = 0) transition (at 373 nm), and band III is an S1 (ν = 0)→S0 (ν = 1) transition (at 384 nm). The “II /IIII ” emission intensity ratio is known to increase with increasing solvent polarity [85–87]. The II /IIII ratio for monoalkylammonium thiocyanates is 1.01–1.23. In the case of [EMIM][(CF3 SO2 )2 N], it is 0.85, and [BMIM][PF6 ] shows a particularly high ratio: 2.08 (cf. water = 1.87, acetonitrile = 1.79, methanol = 1.35) [68]. Additionally, estimation of the dielectric constant for some ILs has been carried out from these measurements. Since [EMIM][(CF3 SO2 )2 N] shows a λmax at 431 nm, the dielectric constant is estimated to be lower than 10 since λmax is shorter than that in hexanol (dielectric constant: 13.5) [69]. This correlates well with the direct measurements of dielectric constant by Weingartner.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

64 3 Physical Properties of Ionic Liquids for Electrochemical Applications

3.6.6 Nile Red

Nile Red shows positive solvatochromism. The degree of λmax shift is known to depend on the dipolarity/polarizability of the medium. The λmax and the calculated E NR values of Nile Red are summarized in Table 3.10. For ILs of the [BMIM]+ cation, the E NR value decreased with anions in the following order: [NO3 ]− > [BF4 ]− > [(CF3 SO2 )2 N]− > [PF6 ]− [88, 89]. Additionally, Nile Red was applied to estimate the polarity of protic ILs, while most other dyes were bleached in the presence of protons [89]. It should be noted here that comparison of ILs via only one polarity parameter is dangerous in discussion of the polarity of ILs.

3.7 Solubility of Metal Salts

Solubility of metal salts in ILs is extremely important in electrodeposition. In this section, the solubility of metal salts in air stable ILs is summarized. The solubility of metal salts in halometalate type ILs has been summarized in previous reports [90, 91]. In addition, many IL systems have been reported as electrolytes for lithium-ion secondary batteries. Some metal salts were reported to be soluble above 50 mol%. However, these systems were obtained by mixing ILs with metal salts in organic solvent or water followed by removal of the solvent; this may produce supersaturated solutions. In this section, these systems are omitted due to space limitations. In ILs, anions and/or cations have weakly coordinating properties, and this solvation energy is not large enough to break the electrostatic interactions between ions or metal atoms in metal salts. Consequently, it is generally expected that common ILs have very low solubilizing ability for metals or metal salts. Rogers et al. reported the evaluation of distribution ratios of Cs+ , Na+ , Sr2+ , Cl− in [Cn MIM][PF6 ] (n = 4, 6, 8)/water mixtures [40b]. The distribution ratio is defined as the concentration ratio of solute in the IL phase to that in the aqueous phase. As shown in Table 3.11, all distribution ratios are very low, such as 10−3 –10−2 . Although the solubility of these ions in [Cn MIM][PF6 ] is unknown, it is expected from these results to be very low. Alfonso et al. evaluated the solubility of LiCl, HgCl2 , and LaCl3 in [Cn MIM][BF4 ] (n = 4, 8, 10) and [Cm MIM][PF6 ] (m = 4, 8) (Table 3.12 entries 1–5) [92]. The ILs containing the BF4 anion solubilized these salts more than those containing the PF6 anion. However, the highest solubility was around 10−4 wt%, still very low. In addition, they prepared ILs from ions having ether or hydroxy groups, expecting further interaction with ions. In these ILs the solubility of HgCl2 and LaCl3 was certainly improved (Table 3.12 entries 6–12). MacFarlane et al. evaluated the solubility of CoCl2 ·6H2 O and CuCl2 ·2H2 O in [C2 MIM] dicyanamide [93]. Compared to traditional ILs containing the Tf2 N anion, the IL containing DCA anion dissolved CoCl2 ·6H2 O and CuCl2 ·2H2 O to a greater

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.7 Solubility of Metal Salts 65 Table 3.11 Distribution ratios between RTIL/aqueous phases. Data

from Ref. [40d] ion

[C4 MIM][PF6 ]

[C6 MIM][PF6 ]

[C8 MIM][PF6 ]

Na+ Cs+ Sr2+ Cl−

0.023 0.067 0.048 0.0017

0.011 0.068 0.029 0.0014

0.011 0.072 0.026 0.00041

Aqueous phase, pH 7

extent. In order to elevate the solubility, it is important to strengthen the interaction between the ILs and the metal ion. As an approach to enhancing the interaction, functional groups were incorporated in the cation or anion to prepare so-called taskspecific ILs (TSIL). Davis et al. synthesized TSIL with thioether or thiourea groups introduced into a side chain of the imidazolium cation. These were effective in extracting Hg2+ or Cd2+ ions from an aqueous phase [94]. Table 3.13 shows the distribution ratios of Hg2+ and Cd2+ in TSIL/water mixtures. Although the ion species are different, these distribution ratios were significantly improved. This is attributed to the interaction between the sulfur atom in TSIL and Hg2+ or Cd2+ ions. As another approach to enhancing the interaction between ILs and metal salts, an extractant highly compatible with both ILs and the metal salt was added. There are many reports on the extraction of metal ions from the aqueous phase into an IL phase with such extractants. Typical examples include crown ethers [40b,95], Table 3.12 Observed solubility constants (K s ) of inorganic salts in sev-

eral ionic liquids. Data from Ref. [92] Entry

1 2 3 4 5 6 7 8 9 10 11 12 a

K s [a]

Ionic liquids Cation

Anion

LiCl

HgCl2

LaCl3

[C4 MIM]+ [C4 MIM]+ [C8 MIM]+ [C8 MIM]+ [C10 MIM]+ [C2 OHMIM]+ [C2 OHMIM]+ [C3 OMIM]+ [C3 OMIM]+ [C5 O2 MIM]+ [C5 O2 MIM]+ [C5 O2 MIM]+

[PF6 ]− [BF4 ]− [PF6 ]− [BF4 ]− [BF4 ]− [PF6 ]− [BF4 ]− [PF6 ]− [BF4 ]− [Cl]− [PF6 ]− [BF4 ]−

12.08 15.54 35.32 56.02 12.64 144.47 18.46 12.44 14.43 9.98 35.52 21.36

4.06 41.41 32.98 35.92 2.12 44.64 84.73 50.13 220.86 295.34 147.48 174.17

6.58 10.92 8.49 53.25 47.12 32.47 54.01 37.61 180.27 379.23 97.22 292.46

Observed K s (10−6 g of salt g−1 of ILs)

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

66 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.13 Distribution ratio for Hg2+ and Cd2+ in the mixed systems

of water and TSIL 1 or TSIL 2. Data from Ref. [94] TSIL

Cation

pH

Distribution ratio

Hg2+

1

198

Hg2+

7 1 7 1 7 1 7

208 330 376 346 343 20 23

Cd2+ Cd2+ Hg2+ Hg2+ Cd2+ Cd2+

molecules containing phosphine oxide groups [96], and calixarenes [97]. The development of ILs having improved coordinating properties for metal salts will be an important area of study in the future. 3.8 Electrochemical Properties 3.8.1 Potential Window

For electrochemical applications, the potential window of the electrolyte solution is one of the important properties. The potential window is governed not only by the chemical structure of the materials used but also by the electrode materials, sweep rate of the potential, temperature, atmosphere, solvent, impurity and so on. Since values of potential windows in the literature have been evaluated under various conditions, it is not easy to compare the values. Use of various reference electrodes (RE) for the determination of cathodic and anodic limits of ILs makes the situation even more complicated. At least, the potential of REs should be confirmed with common redox potentials for non-aqueous systems. For example, the ferrocene(Fc)/ferrocenium(Fc+ ) redox couple is helpful as a standard for many ILs. Although Ag/AgCl(aq), Ag/Ag+ (organic solvents) and pseudo-metal electrodes such as Ag wire and Pt wire are often used as REs, these are not stable enough due to the generation of unstable membrane potentials, chemical reactions on the metal surface, and so on. As a stable RE, Katayama et al. reported an Ag/Ag+ (IL) reference consisting of a silver wire inserted in a silver salt/IL solution as the inner solution [98]. The Ag/Ag+ (IL) reference is stable also in the measurement under specific conditions (under reduced pressure, high temperature, dry atmosphere, etc.). The potential windows are usually evaluated by cyclic voltammetry (CV) or linear sweep voltammetry (LSV). In the CV method, it must be noted that the electrochemically oxidized (or reduced) products of the first sweep must affect the voltammograms of the reverse sweeps. Such effects do not appear in the LSV method, since fresh

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.8 Electrochemical Properties 67

test solution and electrodes are employed for anodic sweep and cathodic sweep, respectively. Instead, reproducibility must be checked for LSV measurements. In both cases, the anodic and cathodic limits are defined as the voltage where the current density reaches a certain value. The cut-off current density is generally 1.0 mA cm−1 with a sweeping rate of 50 mV s−1 . Generally, cathodic and anodic limits of pure ILs are attributed to the oxidative decomposition of the anion and the reductive decomposition of the cation, respectively. Impurities, especially water and halide anions, must be removed carefully, otherwise these drastically narrow the potential window. Table 3.14 shows a series of potential windows and the conditions of measurement for a series of ILs. The potential windows of imidazolium salts are around 4 V. Imidazolium salts having active protons on the 2-carbon sometimes decompose easily. In fact, the potential windows of imidazolium salts are wider when

Table 3.14 Electrochemical windows for a variety of ionic liquids

Cation

[EMIM]+

[e-diMIM]+ [N1113 ]+ [N1114 ]+ [N1114 ]+ [N111,2O1 ]+ [N1224 ]+ [N122,2O1 ]+ [N2226 ]+ [P14 ]+ [P14 ]+ [PP13 ]+ [C2 -dabco]+ [S222 ]+ [S444 ]+

Anion

Working electrode

Reference electrode

Potential window[a]

Ref.

[BF4 ]− [BF4 ]− [BF4 ]− [BF4 ]− [(CN)2 N]− [CF3 CO2 ]− [CF3 SO3 ]− [CF3 BF3 ]− [(CF3 SO2 )(CF3 CO)N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 CF2 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [CF3 BF3 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [C2 F5 BF3 ]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]−

Pt GC Pt Pt Pt Pt Pt GC GC Pt GC GC Pt GC GC GC GC GC Pt GC GC GC GC Pt GC GC

Al/Al3+ Al/Al3+ I− /I3 − Ag wire Ag wire I− /I3 − I− /I3 − Fc/Fc+ Fc/Fc+ I− /I3 − Ag wire Ag wire I− /I3 − I− /I3 − Fc/Fc+ Ag wire I− /I3 − Fc/Fc+ Ag/AgCl aq Ag wire Ag wire Fc/Fc+ I− /I3 − Fc/Fc+ I− /I3 − I− /I3 −

4.4 > 2.1[b] 4.4 4.4 3.0 3.2 3.8 4.6 3.1 4.2 4.1 4.1 4.4 5.8 5.9 5.6 5.4 5.8 5.8 5.6 5.5 5.4 5.8 5.0 5.2 5.2

105 105 106 107 108 6 6 109 101 6 7 7 6 61 110 57 61 110 102 57 57 111 103 112 63 63

a Many of the potential windows were estimated from the voltammograms shown in the reference papers; cut off current density ∼ 1 mA cm−2 . b Anodic limit was not given. [e-diMIM]+ : 1-ethyl-2,3-dimethylimidazolium, [PP13 ]+ : N-methyl-N-(n-propyl)piperidinium [C2 -dabco]+ : N-ethyl-1,4-diazabicyclo[2.2.2]octane.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

68 3 Physical Properties of Ionic Liquids for Electrochemical Applications

the 2-position is substituted by an alkyl chain. However, 2-substituted imidazolium salts generally have higher melting temperatures or higher viscosity than unsubstituted ones. Aliphatic cations such as ammonium cations and piperidinium cations are relatively strong against both oxidation and reduction. Therefore, their potential windows are usually around or wider than 5 V. Thus far, ionic liquids having a potential window over 7 V have also been reported [99]. Generally, TFSI anion-based ILs have relatively wide electrochemical windows on a wide variety of electrodes. Also, BF4 − -based ILs have good properties, but it must be noted here that this anion is not stable against carbon electrodes [100]. Since some ILs have excellent electrochemical stability, as shown in Table 3.14, they are favorable for application as electrolyte materials. Recently, ionic liquids have been investigated as conductive and redox media for lithium ions. Stable electrochemical deposition and dissolution of Li metal (Li/Li+ ) was observed for the lithium salt solution of [N1113 ][TFSI], [N122,2O1 ][TFSI], and [PP14 ][TFSI] [101–103]. In order to observe the redox couple of lithium metal, Ni should be used as working electrode because it does not form alloys with lithium metal. In addition to this, the atmosphere must be pure Ar, because Li metal reacts rapidly with N2 to form conductive LiN. The potential window is one of the most important physical properties for the selection of a solvent for electrolysis. However, it should also be noted that the surface layer on the electrode, which is formed by chemical or electrochemical deposition, often stabilizes the system. For example, Katayama reported that a lithium ion conductive passivated layer, which is formed on the tungsten electrode as a result of reductive decomposition of the cation of the IL during the first sweeps, enables the reversible deposition and dissociation of Li metal [104]. Howlett et al. have also discussed this extensively [112a–c]. This surface film formation is being proposed to protect corrosion of some reactive metals [112d]. 3.8.2 Ionic Conductivity

The ionic conductivity (σ i ) can be described by the following equation.  ni e i µi i =

(3.6)

where ni is the number of ith ions, e is the charge of an electron and µi is the mobility of the ith ion. The net ionic conductivity is the sum of the product for each effective carrier ion species in the system. In order to compare the ionic conductivity of some ILs, one has to note that every IL has a different ion concentration (n). Therefore, the molar conductivity () is usually helpful to know the contribution of the ion mobility (µ) for the ionic conductivity.  = σi /d where d is the salt concentration in mol L−1 .

(3.7)

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.8 Electrochemical Properties 69

For classical dilute aqueous electrolyte solutions, where the salts are perfectly dissociated, the molar conductivity is governed by the viscosity of the system. σ η = constant

(3.8)

This equation is known as Walden’s rule. The constant is called the Walden product. Although the salt contents of bulk ILs are very high (about 3–7 mol L−1 ), the Walden plots for a variety of ILs are similar to that of a conventional diluted system [113]. This observation indicates that ILs are ionized effectively, even in the bulk. However, ILs also contain ion aggregates which do not contribute to the ionic conductivity. Recent research shows more specifically how much ILs are ionized [114]. The Arrhenius plot of the viscosity of the ILs is not a straight line but a Vogel–Fulcher–Tamman (VFT) type curve. Since ionic conductivity is the inverse of the viscosity (Eq. (3.8)), it also obeys the VFT equation.  σi = σ0 exp

−B T − T0

 (3.9)

where σ 0 and B are constants and T 0 is the ideal T g . Equation (3.9) clearly indicates that ionic conductivity could be improved by lowering the T g of the system. The difference in the temperature dependences of ionic conductivity (and viscosity) for ion-conductive glass-forming materials has been discussed by Angell et al. using “fragility” parameters [115]. So far, the ionic conductivity of most ILs has been measured by the complex impedance method [116]. In this method, charge transfer between carrier ions and electrode is not necessary. Therefore platinum and stainless steel are frequently used as “blocking” electrodes. However, it is often difficult to distinguish the resistance and dielectric properties from Nyquist plots obtained by the impedance measurement. In order to clarify this, additional measurements using non-blocking electrodes or DC polarization measurement are needed. The ionic conductivities of ILs are lower than those of conventional aqueous electrolytes, since the viscosity of ILs is generally high (> 30 cP, except for some systems). However, comparing with salt solutions having similar viscosity such as oligo(ethylene oxide)/lithium salt solutions [117], ILs show even higher ionic conductivity because of the much larger number of carrier ions. The ionic conductivity and related properties of [EMIM] salts are summarized in Table 3.15. Among them, [EMIM][TFSI] and [EMIM][BF4 ] show both relatively high ionic conductivity and low viscosity. Imidazolium salts are known to show higher ionic conductivity than those of ammonium ones having similar formula weight. The effect of the alkyl chain length on the ions is also obvious. [EMIM][TFSI] shows the maximum ionic conductivity among the [TFSI]-based imidazolium salts, but further elongation of the alkyl chain causes a decrease in conductivity. We have also investigated the ion conductive properties of a series of “neutralized” ILs, prepared by neutralization of amines with equimolar amounts of Brønsted

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

70 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.15 Specific ionic conductivity and related properties of imida-

zolium salts at 25 ◦ C

Cation

[EMIM]+

Anion

Conductivity r i /mS cm−1

Molar conductivity / S cm2 mol−1

Ref.

[BF]− [BF4 ]− [BF4 ]− [PF6 ]− [CF3 SO3 ]− [CF3 CO2 ]− [C3 F7 CO2 ]− [CH3 COO]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]− [(CF3 CF2 SO2 )2 N]− [C(CN)3 ]− [(CN)2 N]− [NbF6 ]− [TaF6 ]− [CH3 BF3 ]− [C2 H5 BF3 ]− [n-C3 H7 BF3 ]− [n-C4 H9 BF3 ]− [n-C5 H11 BF3 ]− [CH2 CHBF3 ]− [CF3 BF3 ]− [C2 F5 BF3 ]− [n-C3 F7 BF3 ]− [n-C4 F9 BF3 ]−

13.6 13.6 13 (26) 5.2 (26) 8.6 (20) 9.6 (20) 2.7 (20) 2.8 (20) 8.8 (20) 5.7 8.4 (26) 9.0 3.4 (26) 180 (20) 270 (20) 8.5 7.1 9.0 6.3 5.7 3.2 2.7 10.5 14.8 12.0 8.6 5.2

2.1 2.1 2.0 (26)

24a) 34 7 7 6 6 6 6 6 24a) 7 121 7 28 28 120 120 34 34 34 34 34 34 109 109 109 109

1.5 2.1 (26) 2.3 1.1 (26) 32.8 (20) 44.3 (20) 1.6 1.3 1.5 1.2 1.1 0.7 0.6 1.9 2.7 2.5 2.0 1.3

The number in parentheses is the measurement temperature.

acids [118]. Some conductivity values are shown in Table 3.16. Physical properties of the neutralized ILs showed similar trends to those of the quaternary ones. Since neutralized ILs are easy to prepare, these are useful models to find candidate ions for new ILs. They are also expected to be proton conductors. Some ion conductive properties of lithium salt/ IL solutions are summarized in Table 3.17. Generally, the ionic conductivity of ILs containing lithium salts are lower than those of pure ILs, even though the addition of lithium salt increases the net number of ions in the system, due to smaller formula weight of lithium. According to the literature, the major reason for these phenomena may be the increase in viscosity and T g (some specific values are shown in Table 3.17). The aggregation of lithium ions in ILs, which causes the decrease in the effective carrier ion number, might be another reason for the decrease in ionic conductivity.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.8 Electrochemical Properties 71 Table 3.16 Ionic conductivity (25◦ C) of amines neutralized by HBF4 .

Data from Ref. [118] Conductivity r i /mS cm−1

Tg/ ◦C

T m /◦ C

1-methylpyrazole

19

−109.3

−5.9

2-methyl-1-pyrroline

16

−94.3

17.1

1-methylpyrrolidine

16

––

−31.9

1-ethylcarbazole

2.2

−68.0

––

2,3-lutidine

5.9 × 10−3

––

59.4

2,6-lutidine

< 10−4

−10.9

104.6

pyrrole

< 10−5

0.1

––

1-methylpyrrole

< 10−5

−15.9

––

Amine

Structure

Although the lithium ion transference numbers in lithium salt/ IL solutions are important, especially for battery applications, few literature reports refer to the specific values. Since the component ions of the ILs themselves have high mobility, the lithium ion transference number should be low for most cases. In order to suppress the mobility of the component ions, we proposed zwitterion compounds having an imidazolium cation structure [119]. Normally, such zwitterions are solid

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

72 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.17 Ionic conductivity of ionic liquids containing lithium salts at

25 ◦ C

Cation

Anion

Added salt, amount

Conductivity r i /mS cm−1

Viscosity/ cP

Ref.

[EMIM]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 0.32 mol kg−1

10.6 (30) 6.6 (30)

122

[DMPIM]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 0.32 mol kg−1

3.41 (30) 2.1 (30)

122

[EMIM]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 1m

10 2

30 ∼200

123

[EMIM]+

[BF4 ]−

–– LiTf2 N,1 m

15 7

36 ∼100

123

[N111,1-CN ]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 0.2 mol L−1

∼10−4 slight increase

124

[N1114 ]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 0.2 mol L−1

∼10−3 slight decrease

124

[N112, 2O1 ]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 0.9 mol L−1

4.0 (30) 0.37 (30)

∼100 (20) 300 (20)

102

[N112, 2O1 ]+

[BF4 ]−

–– LiTf2 N, 0.9 mol L−1

∼10−2 (30) 0.34 (30)

∼1000 (20) 3450 (20)

102

[DEDMIM]+

[(CF3 SO2 )2 N]− [(CF3 SO2 )2 N]−

2.7 (20) 1.4 (20) 0.8 (20) < 0.0001 > 0.001

125

[C2dabco]+

–– LiTf2 N, 0.4 mol L−1 LiTf2 N, 0.8 mol L−1 –– LiTf2 N, 33 mol %

[DEMPZ]+

[(CF3 SO2 )2 N]−

–– LiTf2 N, 10 mol %

2.6 (20) 1.7 (20)

126

112

The numbers in parentheses are the temperature of measurement, [DMPIM]+ : 1,2-dimethyl-3-(n-propyl)imidazolium, [DEDMIM]+ : 1,2-diethyl-3,4-dimethylimidazolium, [DEMPZ]+ : N,N-diethyl-3-methylpyrazolium.

at room temperature and obviously show almost no ionic conductivity. However, the zwitterions were readily changed to liquid by mixing with suitable lithium salts having soft anions. The ionic conductivity drastically increased after adding lithium salts, as shown in Table 3.18. The lithium ion transference number of such lithium salt/zwitterion mixtures was estimated to be higher than 0.5. The usefulness of zwitterions will be mentioned in Section 3.8.4.2. Recently, proton conductive ILs and iodide ion conductive ILs have also been investigated separately. These ILs for specific ion transport are quite important for the development of energy devices such as lithium batteries, fuel cells, and solar cells. This will be discussed further in the next section.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.8 Electrochemical Properties 73 Table 3.18 Ionic conductivity of zwitterions containing equimolar

lithium salts

+ LiTFSI + LiBETI + LiCF3 SO3 + LiBF4 + LiClO4

r i /mS cm−1 at 100◦ C

Tm/ ◦C

< 10 −5 0.89 6.1 × 10−2 7.5 × 10−3 < 10−3 < 10−3

175 –– –– –– –– ––

Tg/ ◦C

18 −37 −5 19 4 24

3.8.3 Diffusion Coefficients of Component Ions

In spite of the high ionic conductivity, there is no guarantee that the IL can transport the desired ions such as metal ions or protons. It is therefore important to analyze the ion transport properties in ILs. The ion conduction mechanism in ILs is different from that in molecular solvents. The ionic conductivity is generally coupled to carrier ion migration and ionic conductivity (σ ) correlates to diffusion coefficient (D) according to the Nernst–Einstein equation (see Eq. (3.10)) where n and q imply the number of carrier ions and electric charge, respectively. R, T, and F stand for the gas constant, the temperature in K, and the Faraday constant, respectively. σ =

Dnq 2 F 2 RT

(3.10)

Compared with electrochemical measurements of ion mobility and diffusion in ion conductive materials [127], pulse-field-gradient NMR (PFG-NMR) is useful to directly measure the diffusion coefficient of ions containing measurable nuclei [128]. In general, diffusion coefficients of 1 H, 13 C, 19 F, and 7 Li are frequently measured as target nuclei and the values obtained are used to calculate the diffusivity parameters of the corresponding component ions. Therefore, ions having no measurable nuclei for NMR cannot be analyzed. Many low-viscosity ILs are composed of fluorinated anions such as BF4 − and TFSI− , and hence it is rather easy to distinguish the diffusion behavior of anions from that of the onium cations. From previous studies [24a,129], diffusion coefficients of relatively low viscosity ILs such as [EMIM][BF4 ] and [EMIM][TFSI] are reported as around 10−11 m2 s−1 at room temperature. The diffusivity of [EMIM][AlCl4 ] measured by electrochemical methods [127] is similar in value to that of [EMIM][BF4 ] and [EMIM][TFSI]. Compared with water, in which diffusion coefficients are around 10−9 m2 s−1 at room temperature [130], it is understandable that component ions in ILs find it hard to diffuse due to strong electrostatic interaction forces. It is readily seen that the difference in these diffusion coefficients arises from the higher viscosity of ILs

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

74 3 Physical Properties of Ionic Liquids for Electrochemical Applications Table 3.19 Physical properties and diffusion coefficients of ionic liquids

at 30 ◦ C. Data from Ref. [114]

Cation

[BMIM]+

[MMIM]+ [EMIM]+ [BMIM]+ [HMIM]+ [OMIM]+ [N4111 ]+ [bpy]+ [P14 ]+

Anion

[BF4 ]− [PF6 ]− [CF3 CO2 ]− [CF3 SO3 ]− [(CF3 SO2 ]2 N]− [(C2 F5 SO2 ]2 N]− [(CF3 SO2 )2 N]−

g/cP

75 182 58 64 40 87 31 27 40 56 71 77 49 60

r /mS cm−1

4.5 1.9 3.8 3.6 4.6 1.9 11 11 4.6 2.7 1.6 2.6 4 3.4

D/10−11 m2 s−1 Cation

Anion

t+

1.8 0.89 2.2 2.2 3.4 1.6 5.8 6.2 3.4 2.2 1.5 1.7 2.8 2.2

1.8 0.71 1.9 1.6 2.6 1.1 3.3 3.7 2.6 1.9 1.5 1.4 2.2 1.8

0.50 0.56 0.54 0.58 0.57 0.59 0.64 0.63 0.57 0.54 0.50 0.55 0.56 0.55

t+ = Dcation / (Dcation + Danion )

compared with that of water. Diffusion coefficients of ILs containing fluorinated (or fluorine-containing) anions are generally large due to weaker electrostatic forces [131]. Reported diffusion coefficients of relatively low viscosity ILs are summarized in Table 3.19. MacFarlane et al. [129] and Watanabe et al. [24a,114] discussed the difference in diffusivity of component ions. Reported diffusion coefficients of ILs are shown in Table 3.19 together with viscosity and ionic conductivity. From that table, it is easy to see that lower viscosity ILs show larger diffusion coefficients and higher ionic conductivity. Cations generally have larger diffusion coefficient values than do anions in ILs. This means that the cation diffuses more easily than the anion. However, the transference numbers of onium cation (t+ ) in ILs calculated from the results of PFG-NMR is in the range 0.5 to 0.6 and their contribution to the ionic conductivity is mostly the same, irrespective of the ion species. In the case of [bpy][BF4 ], the BF4 − shows a larger diffusion coefficient than that of bpy+ , and therefore t+ is below 0.5 [24a]. Thus, as well as thermal and electrochemical properties, the diffusion behavior of component ions is dependent on their structure. Diffusion coefficients (Dimp ) obtained from measurement are calculated via the Nernst–Einstein equation. Furthermore, electrochemical diffusion coefficient measurements are possible which directly measure the diffusion coefficient. The degree of dissociation of a component ion in the IL can be estimated from the relation (DNMR /Dimp ) between Dimp and the diffusion coefficient measure by PFG-NMR (DNMR ) [132]. This parameter is called the “Haven ratio” and should be unity

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.8 Electrochemical Properties 75

if all components completely dissociate into ions. In most cases, the diffusion coefficient of ILs measured by PFG-NMR is larger than that calculated from impedance measurements [62,129]. These results imply that part of the component ions of ILs do not contribute to ion conduction. The gap between the diffusion coefficient and ionic conductivity is attributed to the fact that PFG-NMR could not distinguish the dissociation state of ions, i.e., either ion or ion pair. As a result, measured diffusion coefficients are an average value obtained from the summation of diffusion coefficients of ions and ion pairs. The difference between DNMR and Dimp shows that a fraction of the component ions form ion pairs or an aggregated state. Therefore the carrier ion number is smaller than the calculated value based on the molar concentration of the ILs. Lithium cation transportation in ILs can be analyzed with PFG-NMR. The mixtures [P13 ][TFSI]/LiTFSI and [EMIM][BF4 ]/LiBF4 have been analyzed [133]. When inorganic salts were added to ILs, their viscosity increased and accordingly ionic conductivity decreased. In both reported mixture systems, the diffusion coefficient of the component ions became smaller with increasing inorganic salt concentration. The diffusion coefficient of the lithium cation is the smallest among the ions in the mixture. The lithium cation, which has a smaller ion radius than any of the other component ions, has the strongest electrostatic interactions. This low lithium cation transport number is one of the reasons why ILs are not currently applied as substituents for electrolyte solutions in secondary batteries. Design of specific ILs for target ion transport will be mentioned in the next section. 3.8.4 Ionic Liquids for Specific Ion Conduction

Physicochemical properties of ILs can be changed by variation of the component ions. There are important studies to achieve ILs having excellent properties such as low T m , low viscosity, high ionic conductivity and wide electrochemical potential windows. It is generally understood that ILs are difficult to apply as electrolyte solution substituents because they contain a large number of ions which cannot work as carrier ions for electrochemical devices such as secondary batteries. Therefore, structural design of ions for particular applications is important for ILs. Selective ion conduction is one of the attractive and challenging tasks for IL science. 3.8.4.1 Ionic Liquids Containing Specific Ions Significant differences between molecular solvents and ILs are based on the component species. This is not as a serious problem for the electrochemical devices that require non-specific ions. The unique properties of ILs, especially their high ionic conductivity, thermal stability and non-volatility are great advantages for electrolyte solutions in electrochemical devices These devices need long-range transportation of particular carrier cations such as lithium cations or protons between electrodes. These small cations interact with

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

76 3 Physical Properties of Ionic Liquids for Electrochemical Applications

Fig. 3.5. Ionic liquids with a multivalent anion.

a counter anion with a strong electrostatic interaction and, accordingly, they are difficult to migrate. Then, suitable carrier ions such as lithium cation and proton should be added to ILs for these purposes. It is easy to generate a target carrier ion for ILs if lithium salts or acids are added to the ILs [134]. On the other hand, preparation of ILs composed of the required cation is a better way to provide a higher concentration of the required cations in ILs [135]. A room-temperature molten lithium salt (lithium IL) has been designed by introducing the lithium salt structure into the tail of polar and flexible polymers, or by using anions having highly delocalized negative charge [69]. ILs inherently containing target ions have been prepared by the combination of an onium cation and a multivalent anion. It is believed that the salts composed of multivalent ions hardly melt at room temperature owing to their strong electrostatic interaction. However, some multivalent anions give room temperature ILs by coupling with specific onium cations like alkylimidazolium cations which have been known to form good ILs [131,136]. Multivalent anions can interact with multiple cations and form ILs containing the target cation as shown in Fig. 3.5. For this strategy, sulfate, phosphate, phosphite, and pyrophosphate have been used to couple with both imidazolium cation and protons [136b]. They showed excellent properties with moderate ionic conductivity of 10−5 to 3 × 10−3 S cm−1 at room temperature; with this strategy lithium ion-containing ILs can be prepared. 3.8.4.2 Selective Ion Conduction Since component ions of ILs are highly mobile, these ions potentially move together with target small ions such as the lithium cation and proton. Component ion migration should be inhibited in order to use the ILs for target ion transport. Zwitterionic salts have been proposed as IL derivatives to inhibit component ion migration along the potential gradient [119,137]. Zwitterionic salts in which both the onium cation and the counter anion are tethered covalently cannot migrate along with the potential gradient. Therefore, pure zwitterionic salts show no ionic conductivity. They become conductive when salts are added. The mixtures can be regarded as target ion transport materials. The cation and anion structures of zwitterionic salts are shown in Figure 3.6. Most of the prepared zwitterionic salts are solid at room temperature. However, the salt containing an equimolar mixture of LiTFSI and zwitterionic salt (having imidazolium cation and sulfonic acid anion) is liquid at room temperature and shows ionic conductivity of 1.0 × 10−5 S cm−1 at 50 ◦ C and a lithium ion transference number of 0.56. Liquidization was

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

3.9 Conclusion and Future Prospects 77

Fig. 3.6 Structure of a zwitterionic salt for selective ion conductive materials.

explained by the formation of an IL-like domain with low T g between the cationic part of the zwitterions and the anion of the added salt. Accordingly, the equimolar mixture of these can be regarded as an IL containing a negatively charged tail (and its counter cation). These zwitterions have much potential for electrochemical applications.

3.9 Conclusion and Future Prospects

In this chapter, the basic characteristics of ILs have been summarized. These basic data should be helpful for the use of known ILs for electrochemical purposes and for the design of ILs with better properties.. With the aid of chemistry, there are increasing numbers of ILs being discovered and studied; there may be some superILs that remain undiscovered or not yet synthesized. Through these studies we will be able to achieve better ILs. Some properties of ILs have been analyzed and collected to construct a database for future use as a guideline. A serious problem at present is the fluctuation of data. Even ILs having identical structure have different data reported. These differences are attributed mainly to trace amounts of contaminants and to different analytical methods. Accordingly, the construction of an accurate database of highly pure ILs is the burning issue. Their movement on this issue and a set of accurate data for ultra pure ILs will be supplied in the near future. After a few years, most physicochemical properties of many ILs will be corrected and then we will be able to obtain new strategies to accurately predict the properties from their component ion structures. Acknowledgement

The following coworkers in our laboratory should be acknowledged for their cooperation: Dr. Wataru Ogihara, Dr. Tomonobu Mizumo, Mr. Yukinobu Fukaya, Miss Junko Kagimoto and Mr. Masahiro Tamada. Especially, the author would like to thank Dr. Wataru Ogihara for his considerable contribution to the preparation of this chapter.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

78 3 Physical Properties of Ionic Liquids for Electrochemical Applications

References 1 Welton, T. (1999) Chem. Rev., 99, 2071–2083. 2 Merrigan, T.L., Bates, E.D., Dorman, S.C., and Davis, J.H., Jr. (2000) Chem. Commun., 2051–2052. 3 Wilkes, J.S. (2002) Green Chem., 4, 73–80. 4 Rogers, R.D. and Seddon, K.R. (2003) Science, 302, 792–793. 5 Seddon, K.R. (2003) Nat. Mater., 2, 363–365. 6 Bonhˆote, P., Dias, A-P., Armand, M., Papageorgiou, N., Kalyanasundaram, K., and Gr¨atzel, M. (1996) Inorg. Chem., 35, 1168–1178. 7 McEwen, A.B., Ngo, H.L., LeCompte, K., and Goldman, J.L. (1999) J. Electrochem. Soc., 146, 1687–1695. 8 (a) Janz, G.J. (1967) Molten Salts Handbook, Academic Press, New York and London; (b) Evans, H.T., Jr (1994) CRC Handbook of Chemistry and Physics, 75th edn, D.R. Lide (Editor-in-chief), CRC Press, Boston, Section 3 and Section 4. 9 (a) Ui, K., Ueda, M., Hagiwara, R., and Mizuhata, M. (2004) Yoyuen oyobi Koon Kagaku, 47, 114–123; (b) Ue, M. (1994) J. Electrochem. Soc., 141, 3336–3342; (c) Ue, M. (1996) J. Electrochem. Soc., 143, L270–L272, (d) Evans, H.T., Jr (1994) CRC Handbook of Chemistry and Physics, 75th, edn, D.R. Lide (Editor-in-chief), CRC Press, Boston, Section 12, p. 1. 10 Carda-Broch, S., Berthod, A., and Armstrong, A.W. (2003) Anal. Bioanal. Chem., 375, 191–199. 11 Baker, S.N., Baker, G.A., Kane, M.A., and Bright, F.V. (2001) J. Phys. Chem. B, 105, 9663–9668. 12 MacFarlane, D.R., Meakin, P., Sun, J., Amini, N., and Forsyth, M. (1999) J. Phys. Chem. B, 103, 4164–4170. 13 Kim, J.W., Singh, R.P., and Shreeve, J.M. (2004) Inorg. Chem., 43, 2960–2966. 14 (a) Kim, K.S., Choi, S., Demberelnyamba, D., Lee, H., Oh, J., Lee, B.B., and Mun, S.J. (2004) Chem. Commun., 828–829; (b) Klapotke, T.M., Mayer, P., Schulz, A., and Weigand, J.J.

15

16 17

18

19 20

21

22

23

24

25

26

27

(2005) J. Am. Chem. Soc., 127, 2032–2033. Ngo, H.L., LeCompte, K., Hargens, L., and McEwen, A.B. (2000) Thermochim. Acta, 357–358, 97–102. Holbrey, J.D. and Seddon, K.R. (1999) J. Chem. Soc., Dalton Trans., 2133–2140. Gordon, C.M., Holbrey, J.D., Kennedy, A.R. and Seddon, K.R. (1998) J. Mater. Chem., 8, 2627–2636. (a) Gordon, J.E. and SubbaRao, G.N. (1978) J. Am. Chem. Soc., 100, 7445–7454; (b) Jones, S.D., and Blomgren, G.E. (1989) J. Electrochem. Soc., 136, 424–427; (c) Sun, J., Forsyth, M., and MacFarlane, D.R. (1998) J. Phys. Chem. B, 102, 8858–8864. Dzyuba, S.V. and Bartsch, R.A. (2001) Chem. Commun., 16, 1466–1467. (a) Pernak, J., Sobaszkiewicsz, K., and Foksowics-Flaczyk, J. (2004) Chem.-Eur. J., 10, 3479–3485; (b) Pernak, J., Olszowka, A., and Olszewski, R. (2003) Pol. J. Chem., 77, 179–187. Merrigan, T.L., Bates, E.D., Dorman, S.C., and Davis, J.H., Jr. (2000) Chem. Commun., 2051–2052. Mizumo, T., Marwanta, E., Matsumi, N., and Ohno, H. (2004) Chem. Lett., 33, 1360–1361. (a) MacFarlane, D.R., Golding, J., Forsyth, S., Forsyth, M., and Deacon, G.B. (2001) Chem. Commun., 1430–1431; (b) Holbrey, J.D., Matthew, W., Swatloski, R.P., Broker, G.A., Pinter, W.R., Seddon, K.R., and Rogers, R.D. (2002) Green Chem., 4, 407–413. (a) Noda, A., Hayamizu, K., and Watanabe, M. (2001) J. Phys. Chem. B, 105, 4603–4610; (b) Noda, A. and Watanabe, M. (2000) Electrochim. Acta, 45, 1265–1270. Fuller, J., Carlin, R.T., DeLomg, H.C., and Haworth, D. (1994) J. Chem. Soc., Chem. Commun., 299–300. Wilkes, J.S. and Zaworotko, M. (1992) J. Chem. Soc., Chem. Commun., 965– 967. Cooper, E.I. and O’Sullivan, E.J.M. (1992) Proceedings of the 8th International Symposium on Molten salts, The

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

References

28

29

30

31

32

33 34

35

36

37

38

39

40

Electrochemical Society, Pennington, NJ, Vol. 92–16, p. 386. Yoshida, Y., Muroi, K., Otsuka, A., Saito, G., Takahashi, M., and Yoko, T. (2004) Inorg. Chem., 43, 1458–1462. Susan, P.A.Z., Einloft, S., Dullius, J.E.L., de Souza, R.F., and Dupont, J. (1998) J. Chim. Phys. Chim. Biol., 95, 1626–1639. Huddleston, J.G., Visser, A.E., Reichert, W.M., Willauer, H.D., Broker, G.A., and Rogers, R.D. (2001) Green Chem., 3, 156–164. Fredlake, C.P., Crosthwaite, J.M., Hert, D.G., Aki, S.N.V.K., and Brennecke, J.F. (2004) J. Chem. Eng. Data, 49, 954–964. Tokuda, H., Hayamizu, K., Ishii, K., Susan, M.A.B.H., and Watanabe, M. (2004) J. Phys. Chem. B, 108, 16593–16600. Dzyuba, S.V. and Bartsch, R.A. (2002) Chem. Phys. Chem., 3, 161–166. Zhou, Z., Matsumoto, H., and Tastumi, K. (2005) Chem. Phys. Chem., 6, 1324– 1332. Crosthwaite, J.M., Muldoon, M.J., Dixon, J.K., Anderson, J.L., and Brennecke, J.F. (2005) J. Chem. Thermodyn., 37, 559–568. (a) Matsumoto, H., Kageyama, H., and Miyazaki, Y. (2002) Proc. Electrochem. Soc. (Molten Salts XIII), 1057; (b) Matsumoto, H., Kageyama, H., and Miyazaki, Y. (2002) Chem. Commun., 1726–1727. Alba-Simionesco, C., Fan, J., and Angell, C.A. (1999) J. Chem. Phys., 110, 5262–5272. Chan, B.K.M., Chang, N.-H., and Grimmett, M.R. (1977) Aust. J. Chem., 30, 2005–2003. (a) Hill, A.J., Huang, J., Efthimiadis, J., Meakin, P., Forsyth, M., and MacFarlane, D.R. (2002) Solid State Ionics, 154–155, 119–124; (b) Golding, J., Hamid, N., MacFarlane, D.R., Forsyth, M., Forsyth, C., Collins, C., and Huang, J. (2001) Chem. Mater., 13, 558–564; (c) Forsyth, S.A., Batten, S.R., Dai, Q., and MacFarlane, D.R. (2004) Aust. J. Chem., 57, 121–124. (a) Bowlas, C.J., Bruce, D.W., and Seddon, K.R. (1996) Chem. Commun., 1625–1626; (b) Visser, A.E., Swatolski, R.P., Reichert, W.M., Griffin, S.T., and

41

42

43 44

45

46

47

48

49

50

51

52

53

54

55

Rogers, R.D. (2000) Ind. Eng. Chem. Res., 39, 3596–3604. Mukai, T., Yoshio, M., Kato, T., and Ohno, H. (2004) Chem. Lett., 33, 1630–1631. (a) Yoshio, M., Mukai, T., Kanie, K., Yoshizawa, M., Ohno, H., and Kato, T. (2002) Adv. Mater., 14, 351–354; (b) Yoshio, M., Mukai, T., Ohno, H., and Kato, T. (2004) J. Am. Chem. Soc., 126, 994–995. Lee, C.K., Huang, H.W., and Lin, I.J.B. (2000) Chem. Commun., 1911–1912. Valkenburg, M.E., Vaughn, R.L., Williams, M., and Wilkes, J.S. (2005) Thermochim. Acta, 425, 181–188. (1991) CRC Handbook of Chemistry and Physics, 71st edn (ed. D.R. Lide) CRC Press, Boston, Section 6, p. 8. (1991) CRC Handbook of Chemistry and Physics, 71st edn (ed. D.R. Lide) CRC Press, Boston, Section 12, p. 107. Yoshizawa, M., Xu, W., and Angell, C.A. (2003) J. Am. Chem. Soc., 125, 15411–15419. MacFarlane, D.R., Pringle, J.M., Johnsson, K.M., and Forsyth, S.A. (2006) Chem. Commun., 1905–1917. Earle, M.J., Esperanc¸a, J.M.S.S., Gilea, M.A., Lopes, J.N.C., Rebelo, L.P.N., Magee, J.W., Seddon, K.R., and Widegren, J.A. (2006) Nature, 439, 831–834. Trulove, P. and Mantz, R. (2003) in (eds P. Wasserscheid and T. Welton), Ionic Liquids in Synthesis, Wiley-VCH, Verlag GmbH, pp. 112–116. Seddon, K.R., Stark, A., and Torres, M.-J. (2002) Am. Chem. Soc., Symp. Ser., 819, 34–49. Evans, H.T., Jr (1994) CRC Handbook of Chemistry and Physics, 75th edn (ed. D.R. Lide), CRC Press, Boston, Section 6, p. 243. Ue, M., Takeda, M., Toriumi, A., Kominato, A., Hagiwara, R., and Ito, Y. (2003) J. Electrochem. Soc., 150, A499–A502. Morgan, D., Lee, F., and Scovazzo, P. (2005) Ind. Eng. Chem. Res., 44, 4815–4823. Kitaoka, S., Nobuoka, K., and Ishikawa, Y. (2005) Tetrahedron, 61, 7678–7685.

79

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

80 3 Physical Properties of Ionic Liquids for Electrochemical Applications

56 Seddon, K.R., Stark, A., and Torres, M.J. (2000) Pure. Appl. Chem., 72, 2275–2287. 57 MacFarlane, D.R., Sun, J., Golding, J., Meakin, P., and Forsyth, M. (2000) Electrochim. Acta, 45, 1271–1278. 58 Dzyuba, S.V. and Bartsch, R.A. (2002) Tetrahedron Lett., 43, 4657–4659. 59 Hasan, M., Ivan, V., Kozhevnikov, M., Siddiqui, R.H., Steiner, A., and Winterton, N. (1999) Inorg. Chem., 38, 5637–5641. 60 Suga, T. (2006) in Ionic liquid IIMarvelous Developments and Colorful Near Future (ed. H. Ohno), CMC, Tokyo, pp. 16–22. 61 Matsumoto, H., Ynagida, M., Tanimoto, K., Nomura, M., Kitagawa, Y., and Miyazaki, Y. (2000) Chem. Lett., 922–923. 62 Tokuda, H., Ishii, K., Susan, Md. Abu Bin Hasan, Tsuzuki, S., Hayamizu, K., and Watanabe, M. (2006) J. Phys. Chem. B, 110, 2833–2839. 63 Matsumoto, H., Matsuda, T., and Miyazaki, Y. (2000) Chem. Lett., 1430–1431. 64 Kim, K.S., Shin, B.K., Lee, H., and Ziegler, F. (2004) Fluid Phase Equilib., 218, 215–220. 65 Przybysz, K., Drzewinska, E., Stanislawska, A., Wysocka-Robak, A., Cieniecka-Roslonkiewicz, A., Foksowicz-Flaczyk, J., and Pernak, J. (2005) Ind. Eng. Chem. Res., 44, 4599–4604. 66 Nockemann, P., Thijs, B., Postelmans, N., Hecke, K.V., Meervelt, L.V., and Binnemans, K. (2006) J. Am. Chem. Soc., 128, 13658–13659. 67 Deetlefs, M., Shara, M., and Seddon, K.R. (2005) Refractive Indexes of ILs, in (eds R.D. Rogers and K.R. Seddon), Ionic Liquids IIIA: Fundamentals, Progress, Challenges, and Opportunities, , p. 219. 68 Wassercheid, P. and Welton, T. (2003) Ionic Liquids in Synthesis, Wiley-VCH, Verlag GmbH, pp. 94–103. 69 (a) H. Ohno (ed.) (2006) Ionic Liquid :The Front and Future of Material Development, CMC, Tokyo; (b) H. Ohno (ed.) (2005) Electrochemical Aspects of Ionic Liquids, John Wiley and Sons, Inc., New York. 70 Poole, C.F. (2004) J. Chromatogr. A, 1037, 49–82.

71 Reichardt, C. (1994) Chem. Rev., 94, 2319–2358. 72 Reichardt, C. (2005) Green Chem., 7, 339–351. 73 Taft, R.W. and Kamlet, M.J. (1976) J. Am. Chem. Soc., 98, 2886–2894. 74 Reichardt, C. and Harbusch-Gornert, E. (1983) Justus Liebigs Ann. Chem., 721–743. 75 Park, S. and Kazlauskas, R.J. (2001) J. Org. Chem., 66, 8395–8401. 76 Chiappe, C. and Pieraccini, D. (2006) J. Phys. Chem. A, 110, 4937–4941. 77 Muldoon, M.J., Gordon, C.M., and Dunkin, I.R. (2001) J. Chem. Soc., Perkin Trans. 2, 3, 433–435. 78 Crowhurst, L., Mawdsley, P.R., Perez-Arlandis, J.M., Salter, P.A., and Welton, T. (2003) Phys. Chem. Chem. Phys., 5, 2790–2794. 79 Abaraha, S.J. and Thomas, J. (1985) J. Anal. Appl. Pyrolysis, 9, 65–74. 80 Poole, S.K., Shetty, P.H., and Poole, C.F. (1989) Anal. Chim. Acta, 218, 241–264. 81 Taft, R.W., Abboud, J.L.M., Kamlet, M.J., and Abraham, M.H. (1985) J. Solution Chem., 14, 153–186. 82 Juan, A.D., Fonrodona, G., and Casassas, E. (1997) Trac-Trends Anal. Chem., 16, 52–62. 83 Kamlet, M.J., Abbound, J.L.M., and Taft, R.W. (1983) J. Org. Chem., 48, 2877– 2887. 84 Soukup, R.W. and Sone, K. (1987) Bull. Chem. Soc. Jpn., 60, 2286–2288. 85 Waris, R., Acree, W.E. Jr., and Street, K.W., Jr. (1988) Analyst, 113, 1465– 1467. 86 Dong, D.C. and Winnik, M.A. (1984) Can. J. Chem., 62, 2560–2565. 87 Karpovich, D.S. and Blanchard, G.J. (1995) J. Phys. Chem., 99, 3951–3958. 88 Carmichael, A.J. and Seddon, K. (2000) J. Phys. Org. Chem., 13, 591–595. 89 Ogihara, W., Aoyama, T., and Ohno, H. (2004) Chem. Lett., 33, 1414–1415. 90 Holbrey, J.D., Visser, A.E., and Rogers, R.D. (2003) in Ionic Liquids in Synthesis (eds P. Wasserscheid and T. Welton), Wiley-VCH: Weinheim, pp. 68–81. 91 Hussey, C.L. (1988) Pure Appl. Chem., 60, 1763–1772.

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

References 92 Branco, L.C., Rosa, J.N., Moura Ramos, J.J., and Alfonso, C.A.M. (2002) Chem. Eur. J., 8, 3671–3677. 93 MacFarlane, D.R., Forsyth, S.A., Golding, J., and Deacon, G.B. (2002) Green Chem., 4, 444–448. 94 Visser, A.E., Swatloski, R.P., Reichert, W.M., Mayton, R., Sheff, S., Wierzbicki, A., Davis, J.H., Jr., and Rogers, R.D. (2002) Environ. Sci. Technol., 36, 2523– 2529. 95 (a) Chun, S., Dzyuba, S.V., and Bartsch, R.A. (2001) Anal. Chem., 73, 3737–3741; (b) Luo, H.M., Dai, S., Bonnesen, P.V., and Buchanan, A.C. (2004) Anal. Chem., 76, 2773–2779; (c) Dai, S., Ju, Y.H., and Barnes, C.E. (1999) J. Chem. Soc., Dalton Trans., 1201–1202; (d) Dietz, M.L. and Dzielawa, J.A. (2001) Chem. Commun., 2124–2125. 96 (a) Visser, A.E., Jensen, M.P., Laszak, I., Nash, K.L., Choppin, G.R., and Rogers, R.D. (2003) Inorg. Chem., 42, 2197–2199; (b) Visser, A.E. and Rogers, R.D. (2003) J. Solid State Chem., 171, 109–113. 97 (a) Shinjo, K. and Goto, M. (2004) Anal. Chem., 76, 5039–5044; (b) Luo, H.M., Dai, S., Bonnesen, P.V., and Buchanan, A.C. (2004) Anal. Chem., 76, 3078–3083. 98 Katayama, Y. (2005) in Electrochemical Aspects of Ionic Liquids (ed. H. Ohno), Ch. 3, John Wiley and Sons, Inc., New York. 99 Ignat’et, N.V., Welz-Biermann, U., Kucheryna, A., Bissky, G., and Willner, H. (2005) J. Fluorine Chem., 126, 1150–1159. 100 Chandrasekaran, M., Noel, M., and Krishnan, V. (1990) Talanta, 37, 695–699. 101 Matsumoto, H., Sakaebe, H., and Tatsumi, K. (2005) J. Power Sources, 146, 45–50. 102 Sato, T., Maruo, T., Marukane, S., and Takagi, K. (2004) J. Power Sources, 138, 253–261. 103 (a) Sakaebe, H. and Matsumoto, H. (2003) Electrochem. Commun., 5, 594–598; (b) Howlett, P.C. et al. (2004) Electrochem. Solid State Lett., 7 (5), A97–A101. 104 Katayama, Y., Morita, T., Yamagata, M., and Miura, T. (2003) Electrochemistry, 71, 1033–1035.

105 Fuller, J., Carlin, R.T., and Osteryoung, R.A. (1997) J. Electrochem. Soc., 144, 3881–3886. 106 Katayama, Y., Dan, S., Miura, T., and Kishi, T. (2001) J. Electrochem. Soc., 148, C102–C105. 107 Nakagawa, H., Izuchi, S., Kuwana, K., Nukuda, T., and Aihara, Y. (2003) J. Electrochem. Soc., 150, A695–A700. 108 Barisci, J.N., Wallace, G.G., MacFarlane, D.R., and Baughman, R.H. (2004) Electrochem. Commun., 6, 22–27. 109 Zhou, Z-B., Matsumoto, H., and Tatsumi, K. (2004) Chem. Eur. J., 10, 6581–6591. 110 Zhou, Z-B., Matsumoto, H., and Tatsumi, K. (2005) Chem. Eur. J., 11, 752–766. 111 Zhou, Z-B., Matsumoto, H., and Tatsumi, K. (2004) Chem. Lett., 33, 1636–1637. 112 (a) Yoshizawa-Fujita, M., MacFarlane, D.R., Howlett, P.C., and Forsyth, M. (2006) Electrochem. Commun., 8, 445–449; (b) Howlett, P.C., Brack, N., Hollenkamp, A.F., Forsyth, M., and Macfarlane, D.R. (2006) J. Electrochem. Soc., 153 (3), A595–A606; (c) Howlett, P.C., Izgorodina, E.I., Forsyth, M., and MacFarlane, D.R. (2006) Z. Phys. Chem., 220 (10–11), 1483–1498; (d) Forsyth, M., Howlett, P.C., Tan, S.K., MacFarlane, D.R., and Birbilis, N. (2006) Electrochem. Solid State Lett., 9 (11), B52–B55. 113 Xu, W., Cooper, E.I., and Angell, C.A. (2003) J. Phys. Chem. B, 107, 6170–6178. 114 Tokuda, H., Tsuzuki, S., Susan, M.A.B.H., Hayamizu, K., and Watanabe, M. (2006) J. Phys. Chem. B, 110, 19593– 19600. 115 Angell, C.A., Imrie, C.T., and Ingram, M.D. (1998) Polym. Int., 47, 9–15. 116 Ohno, H., Yoshizawa, M., and Mizumo, T. (2005) Electrochemical Aspects of Ionic Liquids (ed. H. Ohno), Ch. 6, John Wiley and Sons, Inc., New York. 117 Baril, D., Michot, C., and Armand, M. (1997) Solid State Ionics, 94, 35–47. 118 Hirao, M., Sugimoto, H., and Ohno, H. (2000) J. Electrochem. Soc., 147, 4168– 4172. 119 Yoshizawa, M., Narita, A., and Ohno, H. (2004) Aust. J. Chem., 57, 139–144.

81

c03 (JWBG008-Endres)

January 11, 2008

9:31

Char Count=

82 3 Physical Properties of Ionic Liquids for Electrochemical Applications

120 Matsumoto, K., Hagiwara, R., and Ito, Y. (2002) J. Fluorine Chem., 115, 133–135. 121 Tokuda, H., Hayamizu, K., Ishii, K., Susan, M.A.B.H., and Watanabe, M. (2005) J. Phys. Chem. B, 109, 6103– 6110. 122 Seki, S., Kobayashi, Y., Miyashiro, H., Ohno, Y., Usami, A., Mita, Y., Kihira, N., Watanabe, M., and Terada, N. (2006) J. Phys. Chem. B, 110, 10228–10230. 123 Garcia, B., Lavallee, S., Perron, G., Michot, C., and Armand, M. (2004) Electrochim. Acta, 49, 4583–4588. 124 Egashira, M., Okada, S., Yamaki, J.-I., Dri, D.A., Bonadies, F., and Scrosati, B. (2004) J. Power Sources, 138, 240– 244. 125 Hayashi, K., Nemoto, Y., Akuto, K., and Sakurai, Y. (2005) J. Power Sources, 146, 689–692. 126 Abu-Lebdeh, Y., Abouimrane, A., Alarco, P-J., and Armand, M. (2006) J. Power Sources, 154, 255–261. 127 (a) Carlin, R.T. and Osteryoung, R.A. (1988) J. Electroanal. Chem., 252, 81; (b) Simonsen, L.R. and Donahue, F.M. (1990) Electrochim. Acta, 35, 89–92. 128 (a) Gorecki, W., Andreani, R., Berthier, C., Armand, M., Mali, M., Roos, J., and Brinkmann, D. (1986) Solid State Ionics, 18–19, 295–299; (b) Bhattacharja, S., Smoot, S.W., and Whitmore, D.H. (1986) Solid State Ionics, 18–19, 306– 314. 129 (a) Every, H., Bishop, A.G., Forsyth, M., and MacFarlane, D.R. (2000) Electrochim. Acta, 45, 1279–1284; (b) Every, H.A., Bishop, Andrea G., MacFarlane, D.R., Or¨add, G., and Forsyth, M. (2004) Phys. Chem. Chem. Phys., 6, 1758–1765. 130 Mills, R. (1973) J. Phys. Chem., 77, 685–688.

131 Ogihara, W., Sun, J., Forsyth, M., MacFarlane, D.R., Yoshizawa, M., and Ohno, H. (2004) Electrochim. Acta, 49, 1797–1801. 132 Reiche, A., Cramer, T., Fleischer, G., Sandner, R., Sandner, B., Kremer, F., and K¨arger, J. (1998) J. Phys. Chem. B, 102, 1861–1869. 133 (a) Hayamizu, K., Aihara, Y., Nakagawa, H., Nukuda, T., and Price, W.S. (2004) J. Phys. Chem. B, 108, 19527– 19532; (b) Nicotera, I., Oliviero, C., Henderson, W.A., Appetecchi, G.B., and Passerini, S. (2005) J. Phys. Chem. B, 109, 22814–22819. 134 (a) Cooper, E.I. and Angell, C.A. (1983) Solid State Ionics, 9–10, 617–622; (b) Henderson, W.A. and Passerini, S. (2004) Chem. Mater., 16, 2881–2885. 135 (a) Xu, K. and Angell, C.A. (1995) Electrochim. Acta, 40, 2401–2403; (b) Ito, K. and Ohno, H. (1998) Electrochim. Acta, 43, 1247–1252; (c) Tokuda, H., Muto, S., Hoshi, N., Minakata, T., Ikeda, M., Yamamoto, F., and Watanabe, M. (2002) Macromolecules, 35, 1403–1411; (d) Shobukawa, H., Tokuda, H., Susan, M.A.B.H., and Watanabe, M. (2005) Electrochim. Acta, 50, 3872–3877. 136 (a) Ogihara, W., Yoshizawa, M., and Ohno, H. (2002) Chem. Lett., 880–881, (b) Ogihara, W., Kosukegawa, H., and Ohno, H. (2006) Chem. Commun., 3637–3639. 137 (a) Yoshizawa, M., Hirao, M., I-Akita, K., and Ohno, H. (2001) J. Mater. Chem., 11, 1057–1062; (b) Narita, A., Shibayama, W., Sakamoto, K., Mizumo, T., Matsumi, N., and Ohno, H. (2006) Chem. Comm., 1926–1928; (c) Narita, A., Shibayama, W., and Ohno, H. (2006) J. Mater. Chem., 16, 1475–1482.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

83

4 Electrodeposition of Metals Thomas Schubert, Sherif Zein El Abedin, Andrew P. Abbott, Katy J. McKenzie, Karl S. Ryder, and Frank Endres

Between 1980 and about 2000 most of the studies on the electrodeposition in ionic liquids were performed in the first generation of ionic liquids, formerly called “room-temperature molten salts” or “ambient temperature molten salts”. These liquids are comparatively easy to synthesize from AlCl3 and organic halides such as 1-ethyl-3-methylimidazolium chloride. Aluminum can be quite easily be electrodeposited in these liquids as well as many relatively noble elements such as silver, copper, palladium and others. Furthermore, technically important alloys such as Al–Mg, Al–Cr and others can be made by electrochemical means. The major disadvantage of these liquids is their extreme sensitivity to moisture which requires handling under a controlled inert gas atmosphere. Furthermore, Al is relatively noble so that silicon, tantalum, lithium and other reactive elements cannot be deposited without Al codeposition. Section 4.1 gives an introduction to electrodeposition in these first generation ionic liquids. In the 1990s John Wilkes and coworkers introduced air- and water-stable ionic liquids (see Chapter 2.2) which have attractive electrochemical windows (up to ± 3 V vs. NHE) and extremely low vapor pressures. Furthermore, they are free from any aluminum species per se. Nevertheless, it took a while until the first electrodeposition experiments were published. The main reason might have been that purity was a concern in the beginning, making reproducible results a challenge. Water and halide were prominent impurities interfering with the dissolved metal salts and/or the deposits. Today about 300 different ionic liquids with different qualities are commercially available from several companies. Section 4.2 summarizes the state-of-the-art of electrodeposition in air- and water-stable ionic liquids. These liquids are for example well suited to the electrodeposition of reactive elements such as Ge, Si, Ta, Nb, Li and others. Section 4.3 is devoted to electrodeposition in a special class of deep eutectic solvents/ionic liquids which are based on well-priced educts such as e.g. choline chloride. The impressive aspect of these liquids is their easy operation, even under air, as well as their large-scale commercial availability. One disadvantage has to be mentioned: the choline chloride-based liquids especially are currently not yet

Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

84 4 Electrodeposition of Metals

suited to the electrodeposition of reactive elements such as aluminum or elemental semiconductors like silicon. In Section 4.4, finally, troublesome aspects are shortly summarized. An important aspect is that the electrochemical window alone is not sufficient and one can be pretty surprised if the electroreduction of e.g. TaCl5 rather delivers nonstoichiometric halides instead of the desired tantalum metal. For an electroplating bath the solution chemistry also plays an important role and a new concept of additives seems to be necessary. 4.1 Electrodeposition in AlCl3 -based Ionic Liquids 4.1.1 Introduction

Historically, AlCl3 -based ionic liquids were the first to be used for the electrodeposition of metals. As described before, they are easy to synthesize by simple addition of the Lewis acidic AlCl3 to a 1,3-dialkyl-imidazolium, alkyl-pyridinium or quaternary ammonium compound under an inert atmosphere. The main disadvantages of these materials are their corrosiveness and their instability against air and moisture. Nevertheless, they have a kind of universal character to dissolve other metal salts. The Lewis acidity of these materials can be varied by varying the relative amount of organic salt and AlCl3 : with a molar excess of AlCl3 they are Lewis acidic; with a molar excess of the organic salt they are Lewis basic. To yield neutral liquids it is necessary to buffer the 50:50 mol% mixture with NaCl, because minimum variations from the equimolar composition would shift them towards basic or acidic compositions [1]. It is well known that the chemistry and electrochemistry of many elements are influenced significantly by the Lewis acidity of AlCl3 -based ionic liquids. By far the most studies on metal and alloy deposition have been performed in AlCl3 -based ionic liquids. In the following subsections all metals are categorized in groups of the periodic system of the elements. 4.1.2 Group I Metals

Alkali metals have high oxidation–reduction potentials and low atomic masses. Thus they are attractive candidates for anodes in secondary batteries. In this context, it was shown in a couple of investigations that lithium and sodium can be electrodeposited from tetrachloroaluminate-based ionic liquids. 4.1.2.1 Electrodeposition of Lithium Lithium is of particular interest for use as an anode material for secondary batteries because it has the highest electricity storage density of the active metals.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.1 Electrodeposition in AlCl3 -based Ionic Liquids

Fig. 4.1 Simplified electrochemical windows of 1-butyl-pyridinium chloride and 1-ethyl-3-methyl-imidazolium chloride.

The first electrodeposition of lithium from an ionic liquid was reported in 1985 by Lipsztajn and Osteryoung [2]. They were able to deposit lithium from a 1-ethyl3-methyl-imidazolium chloride/aluminum trichloride ionic liquid. They noticed that a “neutral” ionic liquid, a “neutral basic” ionic liquid (neutral + small excess of RCl) and a “neutral acidic” ionic liquid (neutral + small excess of AlCl3 ) each have unique features. Both the basic and the neutral acidic ionic liquids show an extension of 1.5 V of the electrochemical window, making them interesting for electrochemical applications. They found that lithium chloride was not soluble in the neutral but dissolved in the neutral acidic ionic liquid. From the latter no reduction of lithium was observed prior to the cathodic limit of those ionic liquids. Thus they prepared first the neutral acidic ionic liquid with a certain excess of AlCl3 and then added an equivalent amount of lithium chloride to obtain a LiAlCl4 solution in a neutral ionic liquid. From that they were able to reduce lithium ions on tungsten, glassy carbon, and aluminum electrodes. Piersma et al. demonstrated that lithium can be electrodeposited from 1-ethyl3-methyl-imidazolium tetrachloroaluminate ionic liquid, when lithium chloride was dissolved in the melt [3]. Platinum, glassy carbon and tungsten were used as working electrodes with molybdenum and platinum foils as counter electrodes. At −2.3 V a reduction peak of Li+ is observed and at about −1.6 V the stripping of lithium occurs. They noticed that the efficiency was much less than 100%. In addition, they were able to demonstrate that the addition of proton sources like triethanolamine·HCl widens the electrochemical window and allows the plating and stripping of lithium (and also sodium).

85

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

86 4 Electrodeposition of Metals

Fig. 4.2 Cathodic scan cyclic voltammograms of near-Lewis neutral and LiCl-buffered [EMIM]Cl/AlCl3 ionic liquids at a W working electrode. Scan rate: 100 mV s−1 . (a) Negative scan limit: −2.2 V; (b) negative scan limit: −2.7 V.

4.1.2.2 Electrodeposition of Sodium The first work in this field was reported by Winnick et al. in 1995 [4]. In order to design a sodium/iron(II) chloride battery, they examined a 1-ethyl-3-methylimidazolium chloride/aluminum chloride-based system. As described by Lipsztajn and Osteryoung for lithium it was first necessary to synthesize the acidic ionic liquid by adding an excess of AlCl3 and then adding an equivalent amount of sodium chloride as a buffer to obtain again the neutral species.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.1 Electrodeposition in AlCl3 -based Ionic Liquids

Fig. 4.3 Cyclic voltammogram of neutral buffered, unprotonated melt at tungsten (a) and 303 stainless steel (b).

For their experiments they used tungsten and 303 stainless steel as substrates. They used ionic liquids protonated with HCl in order to enhance the platingstripping efficiency. Electrodeposition of sodium from AlCl3 /[EMIM]Cl/NaCl ionic liquid was not observed because of limited cathodic stability. The addition of gaseous HCl increased the cathodic stability so that a Na+ /Na redox was observed: at a pressure of 6.1 Torr a reduction of Na+ at −2.3 V followed by a stripping wave at −2.1 V on the W electrode was observed. Piersma et al. were able to enhance the deposition–stripping behavior by adding triethanolamine hydrochloride instead of gaseous HCl. The addition widened the electrochemical window and the resulting mixtures were reported to be stable for a month.

87

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

88 4 Electrodeposition of Metals

Fig. 4.4 Cyclic voltammogram at tungsten of a neutral buffered, ionic liquid protonated to a partial pressure of 6.1 Torr HCl.

Finally, Kohl et al. used quaternary ammonium chlorides, e.g. benzyltrimethylammonium chloride, together with AlCl3 and sodium chloride (working electrode: Pt, counter electrode : Pt wire, Al wire in melt). The addition of SOCl2 was required in order to reduce Na+ . At −2.4 V the deposition of sodium started and at −1.8 V the reoxidation was observed [5]. 4.1.3 Group II Metals

From the elements of this group, magnesium is probably the most interesting to deposit on other metals because of its property of forming dense layers of the corresponding oxides. To date no method has been published that describes the electrodeposition of magnesium or any other Group II metal from tetrachloroaluminate-based ionic liquids. 4.1.4 Group III Metals 4.1.4.1 Electrodeposition of Aluminum and Aluminum Alloys The electrodeposition of aluminum has enormous potential in industrial applications. The main reason for this is that aluminum reacts with oxygen to form dense layers of aluminum oxides, protecting metals from corrosion. By far most of the publications concerning the electrodeposition of metals from tetrachloroaluminatebased ionic liquids focus on aluminum. In this context, Osteryoung and Robinson were the first, in 1980, when they described the electrodeposition of aluminum on platinum and glassy carbon from

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.1 Electrodeposition in AlCl3 -based Ionic Liquids

an acidic composition of butylpyridinium chloride and AlCl3 where 50% benzene was added [6]. The reduction was observed at −0.43 V. From basic reaction mixtures they were not able to observe a deposition. Osteryoung and Welch demonstrated by coulometry using a tungsten electrode that the deposition process of aluminum is reversible (in a butylpyridinium chloride/AlCl3 ionic liquid) [7]. The deposition occurred at −0.43 V and the oxidation was observed at −0.22 V. In addition, they determined the rate of the corrosion process to be 1 × 10−11 mol cm−2 s−1 . In a bulk deposition they were able to deposit aluminum on a brass substrate at a thickness up to 15 µm. Lay and Skyllas-Kazacos were the first to describe a deposition from imidazoliumbased tetrachloroaluminate ionic liquid [8]. On glassy carbon, aluminum was deposited at −0.2 V (instead of −0.43 V for the pyridinium-based system of Osteryoung and Welch). Furthermore, they were able to show that the deposition process has complicated kinetics and is not simply controlled by diffusion. Using a tungsten electrode they were able to demonstrate in chronopotentiometric measurements that initially a potential of −0.65 V is necessary due to the nucleation process, but after reaching the barrier the potential drops below −0.2 V. Hussey et al. carried out an aluminum bulk deposition on copper foil using a Lewis acidic aluminum chloride 1-ethyl-3-methyl-imidazolium chloride-based ionic liquid [9]. The thickness of the observed deposits were in the range 24–30 µm. Without additives the deposits were not shiny and only poorly adherent. The addition of benzene enhanced the quality of the deposit. XRD measurements confirmed that the composition of the deposits was 100% aluminum metal. Very significant investigations concerning important parameters for the commercialization were performed by Abbott et al. They used benzyltrimethylammonium chloride/AlCl3 instead of 1-ethyl-3-methyl-imidazolium chloride/AlCl3 to deposit aluminum on a number of substrates [10]. The reason for using benzyltrimethylammonium chloride was that this material is less water sensitive, easier to purify, has greater thermal stability and is potentially more cost effective than the materials used before. Surprisingly, when using an iron electrode an underpotential deposition of aluminum at +0.20 V was observed, which was not the case on aluminum (−0.20 V) and platinum (−0.25 V) substrates. In the corresponding cyclovoltammogram a nucleation loop was observed for Al and Pt, which suggests a kinetic nucleation control of the deposition. A test in a hull cell was also performed on a nickel foil. It showed that the brightest and most uniform deposit was obtained at 5.1 × 10−5 A cm−2 . Endres et al. were able to deposit nanocrystalline aluminum from an aluminum chloride/1-butyl-3-methyl-imidazolium chloride-based ionic liquid (molar ratio: 55/45 mol%) and to characterize it by using XRD and TEM [11]. Figure 4.5 shows the corresponding XRD pattern. The size distribution is shown in Figure 4.6. The grain size was determined to be 12 ± 1 nm. Furthermore, Endres et al. were also able to deposit aluminum alloys such as Al–Mn, which are widely used in the automotive and aviation industries for lightweight construction. The deposition was performed from a Lewis acidic ionic

89

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

90 4 Electrodeposition of Metals

Fig. 4.5 XRD pattern of nanocrystalline Al with a grain size of 12 ± 1 nm.

liquid (as described above), where MnCl2 was added. The average grain size of the deposits was 26 ± 1 nm.

4.1.4.2 Electrodeposition of Indium Sun et al. reported the electrodeposition of indium on glassy carbon, tungsten and Nickel. In basic chloroaluminates, elemental indium is formed via one threeelectron reduction step from the [InCl5 ]2− complex [12]. Furthermore, Carpenter reported the deposition of an indium(I) species [13].

Fig. 4.6 Size distribution of nanocrystalline Al TEM image.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.1 Electrodeposition in AlCl3 -based Ionic Liquids

4.1.4.3 Electrodeposition of Gallium The electrodeposition of gallium is of interest for its extraction and purification and for the production of III-V semiconductors. Sun et al. were the first to report the electrodeposition of gallium from Lewis acidic aluminum chloride–1-ethyl3-methyl-imidazolium chloride melts (ratio 60:40 mol%) on tungsten and glassy carbon in 1999 [14] The Ga(I) species was introduced by anodization of the gallium metal. Using a tungsten electrode the electroreduction was observed at +0.255 V. At a deposition temperature of 30 ◦ C the gallium deposits were liquid, covering the tungsten wire. If removed from the electrolyte they solidified in droplet-like form. The Ga(I)/Ga(III) electrode reaction exhibits slow charge transfer kinetics with an anodic transfer coefficient and a standard heterogeneous rate constant of 0.24 and 3.16 × 10−4 cm s−1 , respectively, at tungsten. Using glassy carbon they observed a three-dimensional nucleation of gallium, with diffusion-controlled growth of the nuclei. The diffusion coefficients for the Ga(III) and the Ga(I) species were 2.28 × 10−7 and 9.12 × 10−7 cm2 s−1 , respectively. 4.1.5 Group IV Metals 4.1.5.1 Electrodeposition of Tin Pitner and Hussey studied the electrochemistry of tin in acidic and basic AlCl3 /1ethyl-3-methyl-imidazolium chloride-based ionic liquids by using voltammetry and chronoamperometry at 40 ◦ C [15]. They reported that the Sn(II) reduction process is uncomplicated at a platinum substrate, where in the acidic ionic liquid the reduction wave was observed at +0.46 V on the Pt electrode and the oxidation at +0.56 V. When they used a gold electrode instead of platinum, they observed an underpotential deposition of a tin monolayer and an additional underpotential deposition process that was attributed to the formation of tin–gold alloy at the surface. The deposition of tin on glassy carbon was controlled by nucleation. The formal potentials of the Sn(II)/Sn couple in the 66.7/33.3 and 44.4/55.6 mol% ionic liquids were determined to be 0.55 ± 0.01 V and −0.85 ± 0.03 V, respectively, vs. Al(III)/Al in the 66.7/33.3 mol% ionic liquid, and the diffusion coefficients of Sn(II) were determined to be (5.3 ± 0.7) × 10−7 and (5.1 ± 0.6) × 10−7 cm2 s−1 , respectively. 4.1.6 Group V Metals 4.1.6.1 Electrodeposition of Antimony Habboush and Osteryoung were the first to describe the electrodeposition of a Group V metal from AlCl3 /1-butyl-pyridinium chloride-based ionic liquids. As antimony sources they used SbCl3 or Sb-rods, dissolved by anodic dissolution [16]. For the composition AlCl3 :BuPyCl (0.8:1) a deposition of Sb was observed at −0.885 V

91

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

92 4 Electrodeposition of Metals

and dissolution at −0.420 V and for the solution composed of AlCl3 :BuPyCl (1.5:1) a deposition of Sb occurred at +0.53 V and anodic dissolution at +1.11 V. They remarked that SbCl2 + was the dominant species in the acidic ionic liquids. The reduction of this species on glassy carbon exhibited irreversible behavior. In the basic melts SbCl4 − and SbCl6 − were believed to be the dominant species. In basic media the reduction of Sb(III) to the metal on glassy carbon was also irreversible while its oxidation to Sb(V) showed quasi-reversible behavior. In another publication Lipsztain and Osteryoung used imidazolium-based ionic liquids to study the behavior of Sb(III) under conditions where the unbuffered properties of a neutral ionic liquid played an important role [17]. 4.1.7 Group VI Metals 4.1.7.1 Electrodeposition of Tellurium Sun et al. used basic melts of 1-ethyl-3-methyl-imidazolium chloride and aluminum chloride of different molar ratios to dissolve TeCl4 [18]. At −0.68 V reduction of tellurium was observed, which was clearly controlled by the nucleation/growth rate. The bulk deposition led only to poorly adherent powder which was confirmed to be Te◦ by XRD.

4.2 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids 4.2.1 Introduction

Ionic liquids have attracted extensive attention since they have extraordinary physical properties, superior to those of water or organic solvents. They are usually nonvolatile, nonflammable, less toxic, good solvents for both organics and inorganics and can be used over a wide temperature range (see Chapters 2 and 3). Moreover, ionic liquids have quite large electrochemical windows, up to 6 V, and hence they give access to elements which cannot be electrodeposited from aqueous or organic solutions. Another advantage of ionic liquids is that problems associated with hydrogen ions in conventional protic solvents, for example hydrogen embrittlement, can be eliminated as most of them are aprotic. Chloroaluminate ionic liquids are regarded as the first generation of ionic liquids. However, their hygroscopic nature has delayed progress in many applications since they must be prepared and handled under an inert gas atmosphere. Thus, the synthesis of air- and water-stable ionic liquids, which are considered as the second and third generations of ionic liquids, has attracted further interest in the use of ionic liquids in various fields. Unlike the chloroaluminate ionic liquids, these ionic liquids can be prepared and safely stored outside an inert atmosphere. Generally, they are water insensitive. However, water-containing [BMIM]PF6 can

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.2 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids 93

be pretty aggressive due to the formation of HF by decomposition of the ionic liquid in the presence of water. Therefore, ionic liquids based on more hydrophobic anions such as trifluoromethanesulfonate (CF3 SO3 − ), bis (trifluoromethanesulfonyl) amide [(CF3 SO2 )2 N− ] and tris (trifluoromethanesulfonyl) methide [(CF3 SO2 )3 C− ] have been developed [19–21]. These ionic liquids have received extensive attention, not only because of their low reactivity towards water but also because of their large electrochemical windows. In general, the wide electrochemical windows of the ionic liquids have opened the door to the electrodeposition of reactive elements, such as Al, Ta, Si, Se and others, which cannot be obtained from aqueous solutions at moderate temperatures. For more information on the use of ionic liquids as solvents for electrodeposition of metals and semiconductors, we refer to recently published review articles [22–25]. We have reviewed the electrodeposition of metals and semiconductors in the most popular air- and water-stable ionic liquids in a short review [23]. In this section we present a review of the recent efforts for the electrodeposition of less reactive metals (such as Zn, Cu, Cd, Cr, Pd, Ag, Pt and Sb), highly reactive metals (such as Al, Mg and Li) and, finally, refractory metals (such as Ta and Ti) in air- and water-stable ionic liquids. 4.2.2 Electrodeposition of Less Reactive Metals

Most of the metals that can be electrodeposited from aqueous solutions can also be electrodeposited from ionic liquids. As many ionic liquids are environmentally friendly, they are considered as suitable alternatives for poisonous plating baths. Furthermore, as ionic liquids have very low vapor pressures (often between 10−11 and 10−10 mbar at or near room temperature, at 100 ◦ C the vapor pressure depends on the liquid, in the region of 10−6 –10−4 mbar) they could principally be used in open galvanic baths at variable temperatures without releasing harmful vapors which reduces the amount of volatile organic compounds released into the atmosphere. Another advantage of using ionic liquids instead of aqueous baths is that their thermal stability makes it easier to electrodeposit metals through direct electrodeposition without subsequent annealing. Moreover, electrodeposition carried out in aqueous solutions is often complicated by problems involving hydrogen embrittlement and low current efficiency. As a result, ionic liquids wherein the coevolution of hydrogen is excluded are good alternatives to aqueous plating baths. Examples of some important metals that can be electrodeposited from aqueous electrolytes and ionic liquids will be presented below. 4.2.2.1 Zinc Zinc and its alloys are good materials for corrosion-resistant coatings and they are widely used in the automobile industry. The electrodeposition of zinc or its alloys is normally performed in aqueous electrolyte solutions. However, zinc and its alloys can be obtained in improved quality from ionic liquids. It was shown that Lewis acidic ZnCl2 –[EMIM]Cl (1-ethyl-3-methylimidazolium chloride) liquids in which

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

94 4 Electrodeposition of Metals

the concentration of ZnCl2 is higher than 33 mol%, are potentially useful for the electrodeposition of zinc and zinc-containing alloys [26–28], see also Chapters 4.3 and 5. The cathodic electrochemical window of these liquids is determined by the reduction of Zn(II) to Zn metal. As a result, the electrodeposition of Zn and its alloys is possible from these melts. As these ionic liquids are aprotic solvents, hydrogen embrittlement is excluded. Such systems and others under development may allow deposition of zinc on safety-relevant steel supports, with high-strength steel qualities, which are not allowed to be zincated in aqueous solutions. Moreover, in contrast to the AlCl3 -based ionic liquids (see Chapters 2.1 and 4.1), the ZnCl2 –[EMIM]Cl liquid does not react vigorously with moisture and is thus easier to handle. Recently Abbott et al. [29] reported that Zn can be electrodeposited from a solution of ZnCl2 in urea and ethylene glycol/choline chloride-based ionic liquid (see Chapter 4.3).

4.2.2.2 Copper Copper is a widely used metal with extensive industrial applications, especially in the semiconductor industry. Almost all connections on semiconductor chips are made with copper due to its low electrical resistance, good mechanical properties and high corrosion resistance. Therefore, the electrochemical deposition of Cu becomes more attractive in the semiconductor manufacturing industry because of its low processing temperature, high selectivity and low cost. One problem in semiconductor technology is that the tantalum diffusion barrier, on which copper is deposited, reacts to tantalum oxide at the surface. The electrodeposition of copper has been intensively investigated in chloroaluminate ionic liquids (see for example Ref. [30]). Sun et al. [31] demonstrated that Cu can be electrodeposited in a basic chloride containing 1-ethyl-3-methyl imidazolium tetrafluoroborate ([EMIM]BF4 ). Murase et al. [32] stated that Cu can be electrodeposited in the air- and water-stable ionic liquid trimethyl-n-hexylammonium bis (trifluoromethylsulfonyl) amide and the Cu deposition and dissolution involved one-electron redox reactions. Recently, we reported that nanocrystalline copper with an average crystallite size of about 50 nm can be obtained without additives in the ionic liquid 1-butyl1-methylpyrrolidinium bis (trifluoromethylsulfonyl) amide ([BMP]Tf2 N) [33]. Because of the limited solubility of many tested copper compounds in the ionic liquid [BMP]Tf2 N, copper cations were introduced into the ionic liquid by anodic dissolution of a copper electrode [33]. The SEM micrograph in Figure 4.7 shows the surface morphology of such an electrodeposited copper layer on a gold substrate obtained at a constant potential of −0.250 V (vs. Pt) for 2 h in the ionic liquid [BMP]Tf2 N containing 60 mmol L−1 of Cu(I) at 25 ◦ C. As seen, the deposit is dense and contains fine crystallites with average sizes of about 50 nm. Interestingly, the deposited copper is nanocrystalline without any additive. The electrodeposition of nanocrystalline copper is quite interesting as a coating since nano-Cu has excellent mechanical and electronic properties, superior to those of microcrystalline copper, furthermore nano-Cu is an important catalyst [34].

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.2 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids 95

Fig. 4.7 SEM micrograph of nanocrystalline copper obtained potentiostatically on Au in the ionic liquid [BMP]Tf2 N containing 60 mmol L−1 Cu(I) at a potential of −0.25 V (vs. Pt) for 2 h at room temperature.

4.2.2.3 Cadmium Despite environmental and health issues, cadmium is still an important metal because of the wide variety of its applications, e.g. in solar cells [35] and rechargeable batteries [36]. It was reported [37] that Cd can be electrodeposited from a basic 1-ethyl-3-methyl imidazolium tetrafluoroborate [EMIM]BF4 ionic liquid containing CdCl2 . The cadmium electrodeposits were found to be very pure and adhered well to the tungsten substrate. Furthermore, Cd can also be deposited in the acidic zinc chloride-1-ethyl-3-methyimidazolium chloride (ZnCl2 –[EMIM]Cl) ionic liquid [38]. At a more negative deposition potential, zinc can be codeposited with cadmium. The Cd content in the Cd–Zn electrodeposits can be increased by increasing the Cd(II) concentration or by increasing the deposition temperature [38].

4.2.2.4 Chromium The electrodeposition of chromium in a mixture of choline chloride and chromium(III) chloride hexahydrate has been reported recently [39]. A dark green, viscous liquid is obtained by mixing choline chloride with chromium(III) chloride hexahydrate and the physical properties of this deep eutectic solvent are characteristic of an ionic liquid. The eutectic composition is found to be 1:2 choline chloride/chromium chloride. From this ionic liquid chromium can be electrodeposited efficiently to yield a crack-free deposit [39]. Addition of LiCl to the choline chloride–CrCl3 ·6H2 O liquid was found to allow the deposition of nanocrystalline black chromium films [40]. The use of this ionic liquid might offer an environmentally friendly process for electrodeposition of chromium instead of the current chromic acid-based baths. However, some efforts are still necessary to get shining

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

96 4 Electrodeposition of Metals

chromium deposits which can compete with the conventional Cr(VI)- or Cr(III)based aqueous galvanic processes.

4.2.2.5 Palladium Palladium is employed in a number of industrial applications and fundamental studies because of its high catalytic activity for many chemical reactions, e.g. its ability to absorb hydrogen [41]. On the other hand, due to hydrogen absorption, only brittle Pd deposits can be obtained in aqueous solutions. The advantage of performing electrodeposition of Pd in ionic liquids is that hydrogen evolution does not occur. Sun et al. demonstrated that Pd and some of its alloys, namely Pd–Ag [42], Pd–Au [43] and Pd–In [44], can be obtained from the basic 1-ethyl-3methylimidazolium chloride/tetrafluoroborate ionic liquid. Compact alloy deposits were obtained and the Pd content in the deposits increased with the increase in Pd mole fraction in the plating bath.

4.2.2.6 Silver The electrodeposition of silver from chloroaluminate ionic liquids has been studied by several authors [45–47]. Katayama et al. [48] reported that the room-temperature ionic liquid 1-ethyl-3-methylimidazolium tetrafluoroborate ([EMIM]BF4 ) is applicable as an alternative electroplating bath for silver. The ionic liquid [EMIM]BF4 is superior to the chloroaluminate systems since the electrodeposition of silver can be performed without contamination of aluminum. Electrodeposition of silver in the ionic liquids 1-butyl-3-methylimidazolium tetrafluoroborate ([BMIM]BF4 ) and 1-butyl-3-methylimidazolium hexafluorophosphate was also reported [49]. Recently we showed that isolated silver nanoparticles can be deposited on the surface of the ionic liquid 1-butyl-3-methylimidazolium trifluoromethylsulfonate ([BMIM]TfO) by electrochemical reduction with free electrons from low-temperature plasma [50] (see Chapter 10). This unusual reaction represents a novel electrochemical process, leading to the reproducible growth of nanoscale materials. In our experience silver is quite easy to deposit in many air- and water-stable ionic liquids.

4.2.2.7 Platinum Films or nanoparticles of platinum are of particular interest since they are important catalysts for many chemical reactions. The electrodeposition of platinum in the ionic liquids [BMIM]BF4 and [BMIM]PF6 has been reported [51]. The Pt deposit was shiny, dense and contained nanocrystals with sizes less than 100 nm. Furthermore, the deposited Pt films obtained in the ionic liquids exhibited higher catalytic performance for the electroxidation of methanol compared with the Pt films obtained in HClO4 aqueous solution [51]. The electrodeposition of PtZn from a Lewis acidic 40–60% ZnCl2 –[EMIM]Cl containing PtCl2 has also been reported [52]. Similar to the electrodeposition of palladium there is no hydrogen evolution during platinum deposition in ionic liquids which can alter the quality of platinum deposited in aqueous solutions.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.2 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids 97

4.2.2.8 Antimony Antimony is a brittle silvery-white metal. Although the unalloyed form of antimony is not often used in industry, alloys of antimony have found wide commercial applications. The integration of antimony gives certain desirable properties, such as increased corrosion resistance and hardness. Moreover, antimony is also the component of some semiconductors such as InSb and InAs1–x Sbx . Sb electrodeposits with good adherence were obtained in a water-stable 1-ethyl-3-methylimidazolium chloride-tetrafluoroborate ([EMIM]Cl-BF4 ) room-temperature ionic liquid [53]. Furthermore, it was stated that a crystalline InSb compound can be obtained through direct electrodeposition in the ionic liquid [EMIM]Cl-BF4 containing In(III) and Sb(III) at 120 ◦ C [54]. It is just a question of time until antimony electrodeposition is reported in the third generation of ionic liquids. 4.2.3 Electrodeposition of Reactive Metals

In this section we will show that air- and water-stable ionic liquids can be used for the electrodeposition of highly reactive elements which cannot be obtained from aqueous solutions, such as aluminum, magnesium and lithium, and also refractory metals such as tantalum and titanium. Although these liquids are no longer airand water-stable when AlCl3 , TaF5 , TiCl4 and others are dissolved, quite interesting results can be obtained in these liquids. 4.2.3.1 Electrodeposition of Aluminum As is known, the commercial production of aluminum is carried out by electrolysis of molten cryolite (Na3 AlF6 ) in which aluminum oxide is dissolved at an elevated temperature of about 1000 ◦ C [55]. This method is still the main industrial method for primary aluminum production. However, it is not suitable for coating other metals with a layer of aluminum since the electrolysis is performed at a temperature where Al is liquid. Nowadays, there are various methods for aluminum coating such as, hot dipping, thermal spraying, sputter deposition, vapor deposition and electroplating in e.g. organic solvents. The electroplating process offers some advantages: the deposits are usually adherent and do not affect the structural and mechanical properties of the substrate. Furthermore, the thickness and the quality of the deposits can be adjusted by controlling the experimental parameters. Moreover, the electroplating process is rather cost-efficient, since it is performed at moderate temperature. Because of its high reactivity (−1.67 V vs. NHE), the electrodeposition of aluminum from aqueous solutions is not possible. Therefore, electrolytes for Al deposition must be aprotic, such as ionic liquids or organic solvents. The electrodeposition of aluminum in organic solutions is commercially available (SIGAL-process [56, 57]) but due to volatility and flammability there are some safety issues. Therefore, the development of room-temperature ionic liquids in recent years has resulted in another potential approach for aluminum electrodeposition. Many papers have been published on the electrodeposition of aluminum from chloroaluminate (first

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

98 4 Electrodeposition of Metals

generation) ionic liquids [58–68]. Although high quality Al deposits can be obtained using such liquids, a main disadvantage of them is that they are extremely hygroscopic and thus must be strictly handled under inert gas conditions. Furthermore, the organic halides are very difficult to dry. Therefore, the electrodeposition of aluminum in less reactive air- and water-stable ionic liquids is of great interest. Quite recently we reported for the first time that nano- and microcrystalline aluminum can be electrodeposited in three different air- and water-stable ionic liquids, namely 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl) amide [BMP]Tf2 N, 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl) amide [EMIM]Tf2 N and trihexyl-tetradecyl phosphonium bis(trifluoromethylsulfonyl) amide (P14,6,6,6 Tf2 N) [69, 70]. It was found that the ionic liquids [BMP] Tf2 N and [EMIM] Tf2 N form biphasic mixtures in an AlCl3 concentration range 1.6–2.5 mol L−1 and 2.5–5 mol L−1 , respectively [70]. Moreover, the electrodeposition of aluminum at room temperature occurs only from the upper phase at AlCl3 concentrations ≥ 1.6 mol L−1 and ≥ 5 mol L−1 in the ionic liquids [BMP]Tf2 N and [EMIM] Tf2 N, respectively. The biphasic behavior of such liquids was first reported by Wasserscheid [71], but a comprehensive understanding of the aluminum species in the phases is still missing. Interestingly, we have found that Al can only be electrodeposited from the upper phase of the biphasic mixture. This means that the reducible aluminum-containing species exists only in the upper phase of the biphasic mixtures and hence the electrodeposition of Al occurs only from the upper phase. In the case of the ionic liquid [BMP]Tf2 N, shiny, dense and adherent deposits with very fine crystallites in the nanometer regime can be obtained without any addition of organic brighteners or use of pulse plating techniques (Figure 4.8(a)). In contrast, coarse cubic-shaped aluminum particles in the micrometer regime are obtained in the ionic liquid [EMIM]Tf2 N (Figure 4.8(b)). As the temperature and electrochemical parameters were varied it is unlikely that this observation is due to viscosity effects alone. Probably, the [BMP]+ cation acts as a grain refiner and plays its role by adsorption on the substrates and on growing nuclei, thus hindering the further growth of crystallites. Many more experiments are required to elucidate the effect of the ionic liquid on the deposit quality. 4.2.3.2 Electrodeposition of Magnesium Magnesium and its alloys offer a high potential for use as lightweight structural materials in automotive and aircraft applications. As magnesium is a very reactive metal (E ◦ = −2.37 V vs. NHE), it can be only obtained from aprotic electrolytes. It is worth noting that the electrodeposition of magnesium in organic electrolytes or in ionic liquids is feasible but not straightforward. Recently, it was claimed in several papers (with similar data) that magnesium can be electrodeposited from the ionic liquid 1-butyl-3-methylimidazolium tetrafluoroborate [BMIM]BF4 using magnesium trifluoromethylsulfonate [Mg(CF3 SO3 )2 ] as a source of magnesium [72–75]. Apart from the comparatively low reduction stability of imidazolium ions with magnesium deposition, there is no hard evidence presented for the deposition of metallic magnesium. We ourselves could not electrodeposit

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.2 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids 99

Fig. 4.8 (a) SEM micrograph of an electrodeposited Al film on gold formed in the upper phase of the mixture AlCl3 /[BMP] Tf2 N after potentiostatic polarization at −0.45 V (vs. Al) for 2 h at 100 ◦ C. (b) SEM

micrograph of an electrodeposited Al film on gold made in the upper phase of the mixture AlCl3 /[EMIm] Tf2 N after potentiostatic polarization at −0.05 V (vs. Al) for 2 h at 100 ◦ C.

magnesium using the recipes described in the mentioned papers. In part we had solubility problems with water-free ionic liquids, and a deposit forms which seems to contain mainly decomposition products. On the other hand the reduction stability of ionic liquids with tetra-alkylammonium or pyrrolidinium cations (∼ −3 V vs. NHE) should, thermodynamically, be sufficient to allow magnesium deposition. In a recent paper, we have tried to electrodeposit magnesium in the ionic liquids 1-butyl-1-methyl-pyrrolidinium bis(trifluoromethylsulfonyl) amide and 1-butyl-1-methylpyrrolidinium trifluoromethylsulfonate ([BMP]TfO) using a Grignard reagent and magnesium perchlorate as sources of magnesium, respectively [76]. Pyrrolidinium ions are cathodically about 700 mV more stable than

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

100 4 Electrodeposition of Metals

imidazolium ions; furthermore, it is well known that magnesium can be electrodeposited from Grignard compounds in ether solvents [77, 78]. It was found that the electroreduction of Grignard reagent in the ionic liquid [BMP]Tf2 N might lead to the formation of thin Mg films which, under air, are subject to oxidation to magnesium oxide and hydroxide. Furthermore, the reduction of Mg(ClO4 )2 in [BMP]TfO is followed by an anodic process showing typical stripping peak behavior; however, the current efficiency for magnesium deposition is not very high. The electrodeposition of magnesium in ionic liquids should, thermodynamically, be possible. Nevertheless more effort is required to find a suitable ionic liquid and suitable magnesium precursors for a technically relevant process. 4.2.3.3 Electrodeposition of Lithium Lithium metal is of particular importance as an anode material for high energy batteries. Lithium batteries are used widely in portable electronic devices and electric vehicles. They show the highest energy density among the applicable chemical and electrochemical energy storage systems (up to 180 Wh kg−1 ). Due to its high reactivity (E ◦ = −3.05 V vs. NHE), lithium cannot be electrodeposited from any aqueous electrolytes. The electrodeposition of lithium in the ionic liquid [BMP]Tf2 N containing Li(Tf2 N) was reported by Katayama [79] and MacFarlane [80]. The latter group investigated the cycling properties (repetitive deposition and stripping) of lithium in the ionic liquid [BMP]Tf2 N containing Li(Tf2 N). It was shown that uniform lithium deposit morphology over many cycles can be achieved at moderate current densities. Cycling efficiencies exceeding 99% were obtained [80]. However, the Tf2 N ion breaks down in the presence of lithium. On the one hand, the decomposition products stabilize the Li, on the other hand anion decomposition leads to a limited lifetime of such secondary batteries, therefore much more effort is required to make Li secondary batteries based on ionic liquids. 4.2.3.4 Electrodeposition of Tantalum High-temperature molten salts were found to be efficient baths for the electrodeposition of tantalum [81–86]. Senderoff and Mellors reported the first results on the electrodeposition of Ta using the ternary eutectic mixture LiF–NaF–KF as a solvent and K2 TaF7 as a source of Ta at temperatures between 650 and 850 ◦ C [81, 82]. Despite enormous importance, these baths have many technical and economic problems, such as loss in the current efficiency of the electrolysis process due to the dissolution of metal after its deposition [87] and the expected corrosion problems at high temperatures. Furthermore, from a practical point of view, molten salts are hardly suited for the coating of sensitive materials like NiTi shape memory alloy with tantalum since the electrolysis process is performed at too high temperatures. With ionic liquids a technical electroplating processes might be performed at moderate temperature. Recently, we reported for the first time that tantalum can be electrodeposited as thin layers in the water and air stable ionic liquid 1-butyl-1-methyl pyrrolidinium bis (trifluoromethylsulfonyl) amide at 200 ◦ C using TaF5 as a source of tantalum [88]. The quality of the deposit was found to be improved on addition of LiF to the

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.2 Electrodeposition of Metals in Air- and Water-stable Ionic Liquids 101

Fig. 4.9 (a) SEM micrograph of the electrodeposit formed potentiostatically on Pt in ([BMP]Tf2 N) containing 0.25 M TaF5 and 0.25 M LiF at a potential of −1.8 V for 1 h at 200 ◦ C. (b) XRD patterns of the deposited layer.

deposition bath. The SEM micrograph of the Ta electrodeposit, Figure 4.9(a), made potentiostatically at −1.8 V in ([BMP]Tf2 N) containing 0.25 M TaF5 and 0.25 M LiF on a Pt electrode at 200 ◦ C for 1 h shows a smooth, coherent and dense layer. XRD patterns of the electrodeposit clearly show the characteristic patterns of crystalline tantalum, Figure 4.9(b). We would like to point out clearly that tantalum deposition is not straightforward in ionic liquids, especially at low temperatures. Under the wrong conditions (high current density) mainly subvalent XRD-amorphous tantalum species like [Ta6 Cl12 ]2+ are obtained. There seems to be a limiting current density above which one obtains only subhalides and below which thin crystalline tantalum layers (Figure 4.9(a)) are obtained. Currently we are able to deposit about 1 µm thick tantalum layers. In our opinion thicker deposits are feasible, but the

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

102 4 Electrodeposition of Metals

development of a technical process will require some effort. Some problems in the electrodeposition of refractory and rare earth elements will be presented in Section 4.4. Furthermore, we showed that adherent, dense and uniform layers of Ta can be electrodeposited on NiTi alloy in the ionic liquid [BMP]Tf2 N containing 0.25 M TaF5 and 0.25 M LiF at 200 ◦ C [89]. NiTi alloys are widely used as orthodontic wires, selfexpanding cardiovascular and urological stents, and bone fracture fixation plates and nails [90–92]. The biocompatibility of NiTi implants depends on their corrosion resistance. The major risk associated with NiTi implants is the breakdown of the passive film which occurs owing to the aggressiveness of human body fluids, leading to a release of Ni ions that may cause allergic, toxic and carcinogenic effects [93–95]. We found that the electrodeposition of only a 500 nm thick film of Ta on NiTi alloy improves its corrosion resistance considerably, leading to a decreased release of Ni ions into solution which enhances its biocompatibility [89]. Furthermore, we think that micrometer thick Ta layers, e.g. on stents to improve the X-ray contrast, are possible. 4.2.3.5 Electrodeposition of Titanium Titanium owes its great importance to its excellent mechanical and corrosion performance. As for most refractory metals, high-temperature molten salts are considered as the most efficient baths for titanium electrodeposition. Recently, there was an attempt to electrodeposit titanium at room temperature in the airand water-stable ionic liquid 1-butyl-3-methyl-imidazolium bis (trifluoromethylsulfonyl) amide [BMIM] Tf2 N containing TiCl4 as a source of titanium. Using in situ STM there were hints that titanium may be electrodeposited in ultrathin layers [96]. Our own experience has shown that attempts to deposit micrometer thick titanium deposits with the recipe in Ref. [96] fail. Instead of elemental titanium soluble, polymeric subvalent titanium halide species are obtained. In situ electrochemical quartz crystal microbalance (EQCM) measurements show that there is a tremendous increase in viscosity during TiCl4 electroreduction, furthermore Tf2 N breakdown (see Refs. [97] and [98]) might alter titanium deposition. Thermodynamically, Ti deposition should be possible in thick layers in ionic liquids, but the right ionic liquid and especially the right titanium precursors still have to be found. An idea might be to make Ti(Tf2 N)4 or similar compounds for titanium electrodeposition. 4.2.4 Summary

In this chapter we have briefly discussed the high potential of air- and waterstable ionic liquids as electrolytes for metal deposition. Their extraordinary physical properties, superior to those of water or organic solvents, and their stability, open the door to the electrodeposition of many metals. Some advantages of air- and waterstable ionic liquids in electrodeposition are that they are quite easy to purify and handle and in most cases they do not decompose under environmental conditions. They can have pretty wide electrochemical windows of up to 6 V, and hence they

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 103

give access to reactive metals which cannot be electrodeposited from aqueous or organic solutions. This branch of electrodeposition is quite novel and will require much more effort to develop technical processes. However, there is also a price to pay. In our experience only rarely can the know-how from aqueous electrochemistry be transferred to ionic liquids. A quick success, especially with refractory and rare earth metals, is unlikely as cluster chemistry has to be considered.

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures

The preceding chapters have shown that the majority of metals can now be electrodeposited from ambient-temperature ionic liquids. However, this does not necessarily mean that the liquid with the widest potential window will negate the use of all other ionic liquids. Rather, it is most likely that ionic liquids will be taskspecific with discrete anions being used for metals that cannot be electrodeposited from aqueous solutions such as Al, Li, Ti, V and W. Type I eutectics will probably be the most suitable for Al, Ga and Ge. Type II eutectics are most suitable for Cr and Type III are most suited to Zn, Cu, Ag and associated alloys. Type III will also find application in metal winning, oxide recycling and electropolishing. To date most practically important metals have been electrodeposited from ionic liquids and a comprehensive review is given in articles by Abbott [99] and Endres [100–102]. In this chapter we will concentrate on the deposition of metals from eutecticbased ionic liquids. These have been developed since the end of the 1990s, primarily by our group and that of Sun. Figure 4.10 shows just some of the metals that

Fig. 4.10 A range of metal and metal alloy coatings deposited electrolytically from type II (Cr) and type III (Ni, Cu Zn Sn, Ag) choline chloride-based ionic liquids.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

104 4 Electrodeposition of Metals

have been deposited from these types of eutectics. The principles underlying the eutectic-based liquids are all the same although clearly speciation is dominated by the Lewis or Brønsted acidity of the components. The electrochemistry of the liquids is largely unaffected by the cation although this does have a significant effect upon the physical properties, most noticeably the phase behavior, viscosity and conductivity. A significant number of studies have characterized the physical properties of eutectic-based ionic liquids but these have tended to focus on bulk properties such as viscosity, conductivity, density and phase behavior. These are all covered in Chapter 2.3. Some data are now emerging on speciation but little information is available on local properties such as double layer structure or adsorption. Deposition mechanisms are also relatively rare as are studies on diffusion. Hence the differences between metal deposition in aqueous and ionic liquids are difficult to analyse because of our lack of understanding about processes occurring close to the electrode/liquid interface. One issue is that most metal complexes formed in ionic liquids are anionic and these will have a significant effect on viscosity and mass transport. The effect of metal ion concentration on reduction current will therefore not be linear. Relative Lewis acidity will affect mass transport, ionic strength and speciation and accordingly the nucleation and growth mechanism of metals would be expected to be concentration dependent. The only models that exist for double layer structure in ionic liquids suggest a Helmholz layer at the electrode/solution interface [103, 104]. If the reduction potential is below the point of zero charge (pzc) then this would result in a layer of cations approximately 5 Å thick across which most of the potential would be dropped, making it difficult to reduce an anionic metal complex. Hence, the double layer models must be incorrect. Electrodeposition using eutectic-based ionic liquids has almost exclusively used quaternary ammonium halides with metal halides primarily in the chloride form. Aqueous plating solutions rarely use chlorides as they tend to yield black powdery deposits and the inclusion of halides into metallic coatings is seen as undesirable due to the possibility that it can lead to the breakdown of passivating layers and exacerbate corrosion. The morphology issue is thought to be due to the ease of nucleation from halide salts which leads to large numbers of small nuclei forming at the electrode surface. Lewis basic anions cannot be circumvented for eutecticbased ionic liquids as they need to be good ligands to interact strongly with the Lewis acid. The question that needs to be posed is whether chloride ions actually cause a problem when their activity is negligible due to the presence of a strong Lewis acid. The issue that needs to be addressed is that Type I and II eutecticbased ionic liquids necessarily have high concentrations of metal chlorides and will tend to promote nucleus formation. In many cases the working concentration is 5 to 10 mol dm−3 which, although seemingly high, is not overly different to many aqueous plating solutions. Further ionic liquid formulation needs to address how nucleation can be suppressed while growth is supported.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 105

4.3.1 Type I Eutectics 4.3.1.1 Chlorozincate Ionic Liquids In general the potential windows are not as wide as those for the haloaluminates or the discrete anions and they tend to be limited by the deposition of metal at the cathodic limit and the evolution of chlorine at the anodic limit. Since ionic liquids are aprotic solvents, hydrogen evolution and hydrogen embrittlement that often occur in aqueous baths are circumvented in these liquids. Moreover, because of their thermal stability, these ionic liquids make it easier to electrodeposit crystalline metals and semiconductors through direct electrodeposition without subsequent annealing. From a practical perspective the chlorozincate liquids are easier to make and handle than the corresponding chloroaluminates as they are less susceptible to hydrolysis. As with the aluminum-based liquids the electrochemistry is dominated by the complex anions present in the liquid, which depend upon the composition and the relative Lewis acidity. There is some evidence, however, that hydrolysis of zinc Lewis basic melts does occur as Hsiu et al. used fast atom bombardment mass spectroscopy (FAB MS) to show that some zinc species do contain oxygen [105]. The same group studied the potential limits of [EMIM]Cl/ZnCl2 in the molar ratio range 3:1 to 1:3 [105]. It was found that in the Lewis acidic region (excess ZnCl2 ) the potential window was ca. 2 V; the negative potential limit is due to the deposition of metallic zinc and the positive potential limit is due to the oxidation of the chlorozincate complexes to form chlorine. In the Lewis basic region the potential window could be as large as 3 V, corresponding to the cathodic reduction of [EMIM]+ and the anodic oxidation of Cl− , which is similar to the basic chloroaluminate melts. The potential windows for the Lewis acidic ionic liquids are surprisingly close to the difference in the standard cell potentials for the corresponding half◦ ◦ − E Cl = 2.1 V). This suggests that cell reactions in aqueous solutions (E Zn 2+ /Zn /Cl− 2

while the reduction potential for Zn2 Cl5 − will be shifted with respect to the Zn2+ /Zn couple the oxidation of Cl− to Cl2 will be affected by the same amount. In the Lewis acidic melts underpotential deposition (UPD) of zinc was observed on Pt and Ni electrodes. The potential window and UPD of zinc in Lewis acidic choline chloride (ChCl):ZnCl2 was found to be exactly the same as the corresponding [EMIM]Cl system, suggesting that the cation has little or nothing to do with the electrochemistry of the liquid. An in depth study of the deposition mechanism was carried out by Sun et al. who studied the 1:1 [EMIM]Cl/ZnCl2 system at various temperatures on glassy carbon (GC), nickel and platinum electrodes [106]. The GC electrode required the largest overpotential for deposition. The stripping process showed a single peak on GC, whereas on Ni two oxidation processes were observed, separated by ca. 0.6 V. It was proposed that the more positive oxidation process corresponded to the dissolution of an intermetallic compound formed during electrodeposition.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

106 4 Electrodeposition of Metals

Chronoamperometry on the GC and Ni electrodes at 50 ◦ C showed that the electrodeposition of zinc proceeded by a three-dimensional instantaneous nucleation and growth process. The results also suggested that the growth process is under mixed diffusion and kinetic control. The zinc deposits formed by bulk electrolysis consisted of hexagonal grains with a size of 4–5 µm. These crystals were covered by numerous small needles which could be subsequently removed, however the underlying grains showed good adherence to the substrate. Deposits formed at larger overpotentials were poorly adherent flakes and increasing the temperature to 80 ◦ C also had little effect upon the morphology. Sun also studied the effect of adding a diluent to a liquid to improve mass transport [106]. Propylene carbonate was added from 20 to 60% (v/v) and found to have little effect on the voltammetry. Chronoamperometry showed the same instantaneous three-dimensional growth under mixed diffusion and kinetic control. The propylene carbonate did, however, lead to an improvement in deposit morphology compared to the neat melt and no small needles were seen in any of the deposits. The grain size was affected by the deposition potential, indicating that nucleation density increases with increased overpotential. The grain size was also affected by diluent concentration with larger grains forming with higher propylene carbonate content although this also increased the grain size distribution. Grain size could also be increased by increasing the temperature of the melt. Iwagishi studied the deposition of zinc from Lewis basic [EMIM]Br/ZnBr2 eutectics at 120 ◦ C and investigated the effect of adding ethylene glycol as a diluent [107, 108]. Analysis of the choronoamperometric current–time transients indicated that the overpotential was related to the progressive nucleation with diffusioncontrolled growth of the nuclei. The nucleation loop observed using cyclic voltammetry disappeared on adding more than 45 mol% ethylene glycol to the ionic liquid. The cathodic current increased with increasing ethylene glycol content and it was proposed that this promoted the dissociation of [EMIM]Br to [EMIM]+ cation and Br− , and accordingly the concentration of ZnBr4 2− in the liquid was increased. The study was further extended to investigate the effect of a range of dihydric alcohols (ethylene glycol, 1,3-propanediol, 1,2-butanediol and 1,3-butanediol). The addition of the dihydric alcohol improved the smoothness and color of the deposits and also increased the cathodic current efficiency at high current density. Of the four dihydric alcohols, ethylene glycol gave the best results. The same group studied the effect of water on zinc deposit morphology in the same ionic liquid in the [EMIM]Br/ZnBr2 (70:30 mol%). Smooth layers of silver-colored Zn were obtained at cathodic current densities < 100 A m−2 , whereas smooth grey Zn layers were electrodeposited at cathodic current densities >150 A m−2 . The liquid with a water content of less than 10 ppm was superior to the liquid with a water content of 400 ppm in cathodic current efficiency, smoothness, and metallic luster. The [EMIM]Br/ZnBr2 –ethylene glycol ternary systems were also studied with a water content of 30 ppm. The cathodic current efficiencies were all 100%, even at a current density as high as 300 A m−2 , in Lewis basic liquids with an EG content of between 30 and 75 mol%.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 107

Abbott et al. studied the deposition of zinc from a 1:2 choline chloride (ChCl):ZnCl2 ionic liquid [109] at 60 ◦ C and found deposits with a similar morphology to that shown by Sun. The optimum current density was found to be between 2 and 5 A m−2 and higher current densities led to powdery, non-adherent deposits. This is due primarily to the high viscosity and low conductivity of the choline-based liquids. The current plating efficiency in this liquid was found to be effectively 100% and the deposition process was shown to be almost totally reversible, with only the UPD material remaining on the surface. Chlorozincate liquids have also been studied for the deposition of numerous zinc-containing alloys including Pt–Zn, Zn–Fe, Sn–Zn and Cd–Zn alloys. These alloys will all be discussed in greater detail in Chapter 5. One explanation for the change in deposit morphology with time observed by Sun and others [105–108] could be the structure of the double layer during deposition. As the chlorozincate anions are reduced at the electrode surface the liquid close to the electrode surface will become more Lewis basic − − i.e. Zn2 Cl− 5 + 4 e → 2Zn + 5Cl

and this will in turn affect the composition of the zinc-containing species − − Zn2 Cl− 5 + Cl → 2ZnCl3

The effect of this will be exacerbated if the liquid is viscous. It can be seen from the above studies that the nucleation and growth mechanisms are dependent upon the Lewis acidity of the liquid and this may help to explain the growth of needle-shaped crystals on top of the hexagonal crystallites. The addition of a diluent decreases the viscosity and could allow the chloride ions to diffuse away. 4.3.1.2 Other Type I Eutectics Chlorostannate and chloroferrate [110] systems have been characterized but these metals are of little use for electrodeposition and hence no concerted studies have been made of their electrochemical properties. The electrochemical windows of the Lewis acidic mixtures of FeCl3 and SnCl2 have been characterized with ChCl (both in a 2:1 molar ratio) and it was found that the potential windows were similar to those predicted from the standard aqueous reduction potentials [110]. The ferric chloride system was studied by Katayama et al. for battery application [111]. The redox reaction between divalent and trivalent iron species in binary and ternary molten salt systems consisting of 1-ethyl-3-methylimidazolium chloride ([EMIM]Cl) with iron chlorides, FeCl2 and FeCl3 , was investigated as possible half-cell reactions for novel rechargeable redox batteries. A reversible one-electron redox reaction was observed on a platinum electrode at 130 ◦ C. Elemental gallium has been electrodeposited from chlorogallate ionic liquids formed between [EMIM]Cl and GaCl3 [112]. The direct electrodeposition of GaAs from ionic liquids was studied mainly by two groups. Wicelinski et al. [113] used an acidic chloroaluminate liquid to co-deposit Ga and As. However, it was reported

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

108 4 Electrodeposition of Metals

that Al underpotential deposition on Ga occurred. Carpenter and Verbrugge studied the deposition of GaAs using an ionic liquid based on GaCl3 to which AsCl3 was added [112, 114]. Unfortunately poor quality deposits were obtained and both pure As and Ga were present in the deposits but they did show that semiconductor deposition was possible and thermal annealing could improve the quality of the deposits. The same principle has been used for the deposition of InSb alloys [115]. An ionic liquid based on InCl3 is made and SbCl3 is added. InSb alloy was electrodeposited but there was also some elemental In and Sb in the deposits. The In:Sb ratio could be varied by altering the deposition potential.

4.3.2 Type II Eutectics

Type I eutectics only form liquids at ambient temperatures with metal salts that melt below about 400 ◦ C (e.g. ZnCl2 /ChCl). This is related to the lattice energy of the salt and its ability to interact with the quaternary ammonium salt. There is a limited number of such salts and these preclude several of the technologically important metals being incorporated in eutectic-based ionic liquids. In general, it is metal salts with tetrahedral geometries that have lower lattice energies. One way of decreasing the lattice energy, especially of octahedrally coordinated metal salts, is to use hydrate salts, as these have relatively low melting points. A wide variety of hydrated salt mixtures with choline chloride have been found to form these ionic liquids, including CrCl3 ·6H2 O, CaCl2 ·6H2 O, LaCl3 ·6H2 O CoCl2 ·6H2 O, LiNO3 ·4H2 O and Zn(NO3 )2 ·4H2 O [116] and hence this technology could be generic to the deposition of a range of metals and alloys. However, only chromium and cobalt have been deposited from these liquids. These liquids could be viewed as concentrated aqueous solutions, but because the ionic strength is extremely high the water molecules are strongly coordinated to the ions and hence they are difficult to reduce at the electrode surface. Accordingly the deposition of metals such as chromium can be carried out with high current efficiencies [117]. Figure 4.11 shows the voltammetry of a 1ChCl: 2 CrCl3 ·6H2 O and the corresponding 1ChCl: 2 CoCl2 ·6H2 O. It is immediately apparent that the two metals have different electrochemical behavior. Chromium is reduced via a Cr(II) state which forms an insoluble intermediate at the electrode–solution interface and the deposition process is irreversible. The cobalt analogue is quasi-reversible as there is a ca. 500 mV difference between the deposition potential and the stripping potential. Metal can be deposited from both eutectic mixtures. In the case of chromium a dull, metallic-looking coating is obtained from the bulk electrolysis of 1ChCl: 2 CrCl3 ·6H2 O at − 4 V and 60 ◦ C [116]. The corresponding experiment using CoCl2 ·6H2 O yields a bright adherent metallic film. The addition of up to 10 wt.% LiCl, however, leads to a matt black chromium layer. The film has an amorphous morphology that is free of surface cracks, unlike

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 109

Fig. 4.11 Cyclic voltammetry of a 1ChCl: 2 CrCl3 .6H2 O and the corresponding 1ChCl: 2 CoCl2 ·6H2 O ionic liquids on a Pt microelectrode at a sweep rate of 20 mV s−1 . Cr data at 60 ◦ C and Co data at 25 ◦ C.

samples deposited using aqueous Cr(III) and Cr(VI) which are highly crystalline and have a highly cracked surface [116]. Cross-sectional analysis of the film produced using an ionic liquid showed that it was homogeneous with no structural characteristics, even under the highest magnification. Relatively fast deposition rates could be obtained (up to 60 µm h−1 ) although deposits thicker than about 30 µm tended to become quite powdery and less adherent. Chromium films were deposited onto 304 stainless steel but the substrate was found to corrode relatively quickly ( 1600 h in a salt spray test without any visible signs of corrosion. The black coloration of the film was due to the deposits being nanoparticulate and XRD analysis shows that the 110 and 211 were the predominant crystal faces present. The deposit thickness, adherence and morphology could be further improved using pulse-plating. A range of brighteners and additives used in aqueous plating baths was tested, but none was found to cause an improvement in the deposit morphology. The relatively high viscosity of these ionic liquids allows improved stability of particulate suspensions. Consequently, deposition of a range of Cr composites using Si3 N4 , BN, Al2 O3 or particulate (0.3–1.0 µm) diamond is facile. The scanning electron micrographs presented in Figure 4.12 show a Cr composite deposited from a type II Cr/ChCl liquid containing 5 wt.% diamond powder. The crystals of C (diamond) are clearly visible, held within a Cr metal matrix. EDAX analysis confirms large amounts of both Cr and C. Similar results were obtained with Si3 N4 and Al2 O3 . The particles do not aggregate upon deposition in the film but rather remain as discrete entities, suggesting that they are just dragged onto the surface as the metal deposits.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

110 4 Electrodeposition of Metals

Fig. 4.12 (a) and (b) Scanning electron micrographs of a Cr composite deposited from a type II Cr/ChCl liquid containing 5 wt.% diamond powder.

4.3.3 Type III Eutectics

Eutectics formed between quaternary ammonium salts and hydrogen bond donors (HBD) have potential windows that tend to be controlled by the stability of the carboxylic acid, amide or alcohol. In general the potential windows depend upon the pK a of the HBD. Figure 4.13 shows the potential windows of eutectics formed between ChCl with ethylene glycol, urea and malonic acid. The potential windows are significantly smaller than some imidazolium-based liquids with discrete anions, however, the windows are sufficiently wide for metals such as zinc and nickel to be electrodeposited with high current efficiencies. The potential windows are naturally wider on metals that are less catalytic than Pt. Type III eutectics have the advantages that they are relatively benign and inexpensive and can thus be applied to large scale processes.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 111

Fig. 4.13 Cyclic voltammograms of (a) urea/ChCl, (b) ethylene glycol/ChCl and (c) malonic acid/ChCl ionic liquids. Voltammograms were acquired at room temperature (20 ◦ C), using a Pt disk (2 mm diameter) working electrode, a Ag wire reference electrode and a potential scan rate of 50 mV s−1 .

The deposition of zinc, tin and zinc–tin alloys has been studied in eutectics based on choline chloride with ethylene glycol and urea [118]. SEM images of Zn–Sn alloys deposited from choline chloride with ethylene glycol and urea liquids are shown in Figure 4.14. The morphologies are clearly very different. The electrochemistry of the different metals was affected by the different HBDs and this was assigned to the different speciation of the metals in the liquids. In urea the only zinc-containing species is ZnCl3 − whereas in ethylene glycol ZnCl3 − , Zn2 Cl5 − and Zn3 Cl7 − were detected. This is because urea acts as a far stronger ligand for ZnCl3 − than ethylene glycol [118]. The zinc deposits obtained from electrolysis of both the ethylene glycol and urea-based liquids were similar; dull grey metallic films with good adherence. The SEM images were similar to those reported by Sun et al. [106], which is unsurprising given that a type III eutectic with high ZnCl2 concentrations is very similar in composition to a chlorozincate melt to which ethylene glycol has been added as a diluent. What is clear however is that the diluent is affecting not only the viscosity but also the speciation, which will accordingly change the nucleation and growth processes. The HBD also affects the way in which alloys form. Voltammetry and XRD showed that when both Zn and Sn were present in the same melt a homogeneous ZnSn alloy phase was formed when urea was used as the HBD whereas ethylene glycol caused separate zinc and tin phases to form. As with the type II metal-based liquids (e.g. Cr) particulate suspensions are stabilized by the relatively high viscocity. A dispersion of 3 wt.% Al2 O3 was made in type III eutectics and mild agitation was sufficient to retain the alumina as

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

112 4 Electrodeposition of Metals

Fig. 4.14 Scanning electron micrographs obtained by the electrolysis of 0.5 M ZnCl2 /0.05 M SnCl2 in (a) 1ChCl: 2urea and (b)1ChCl: 2EG, both at a current density of 10 mA cm−2 for 1 h.

a homogeneous dispersion. EDAX and SEM analysis showed the inclusion of approximately 1 wt.% Al2 O3 in the film. Recent work has focussed on converting these liquids into practical plating solutions by investigating ways of improving the morphology of the deposits. In aqueous solutions a range of compounds is routinely added to act as brighteners. These are thought to work by either shifting the redox potential of the metal through complexation or by hindering metal nucleation and growth at the electrode surface. It was shown that the anion of the metal salt has a significant effect on the reduction potential for the Cu2+ /Cu+ and Cu+ /Cu couples in a ChCl: 2urea eutectic, even though it was only present in a small concentration (20 mM) compared to the

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.3 Deposition of Metals from Non-chloroaluminate Eutectic Mixtures 113

chloride ion (>1 M). The anion can also change the trend in Cu2+ /Cu+ redox with respect to that of the Cu+ /Cu couple [99]. The addition of well known complexing agents such as ethylene diamine and EDTA can have a more significant effect on redox potentials where the position of stripping potentials can be shifted by over 250 mV. The complexing agents make it more difficult to reduce the metal and hinder nucleation, which leads to less nuclei formation and allows the crystals to grow larger before they encounter a neighboring grain. Bulk deposition from an ionic liquid containing just CuCl2 produces black, powdery deposits whereas the addition of a complexing agent can lead to lustrous copper deposits [99]. The addition of strong complexing agents to ionic liquids may not be trivial, however, as it will also affect the charge on the metal center and the interaction between the metal center and the halide anion of the ammonium salt, influencing the phase behavior and viscosity. In Type III eutectics the possibility exists to choose a Brønsted acid that could act as a built-in brightener. Most deposition experiments in ionic liquids have been carried out primarily using halide salts, which is different to most aqueous processes. Oxides have been found to dissolve in high concentrations in Type III eutectics [119, 120] and these have been studied for the electrowinning of metals from ores or waste materials [121]. In principle, however. they could be used for metal plating as they produce coatings with morphologies similar to those obtained using chlorides. One issue that arises is the speciation of the oxide following dissolution. FAB-MS data have shown that some metals, particularly with acid-based hydrogen bond donors revert to the halometalate complexes, e.g. CuO gives CuCl3 − [120]. Urea-based liquids give complexes where the oxygen is still attached to the metal center e.g. ZnO gives [ZnOCl·urea]− [119].

4.3.4 Future Developments

From an academic standpoint there are numerous fundamental issues that still need to be addressed, including the effect of speciation on the mechanism of nucleation and growth. An understanding of the double layer structure and processes occurring during deposition is essential for an informed choice of suitable metal salts and the design of brighteners. Diluents to reduce the liquid viscosity and make them easier to handle will also have to be identified. It is highly probable that practical plating solutions will be a complex mixture of salts, viscosity modifiers, brighteners and wetting agents, analogous to current aqueous plating solutions. To develop practical plating systems information about the long term stability of the ionic liquids under high applied current densities needs to be determined. The effect of adsorbed moisture on deposit morphology also needs to be ascertained as practical liquids will have to be as robust as possible.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

114 4 Electrodeposition of Metals

4.4 Troublesome Aspects 4.4.1 Deposition of Reactive Elements

The deposition of anti-corrosion coatings is one of the main goals of electrochemical research. The majority of useful metals for this application have extremely negative reduction potentials. Metals such as vanadium, niobium, tantalum, titanium, magnesium and others are precluded from deposition in ambient temperature systems due to the narrow potential window of potential solvents. They should, however, be easily electrodeposited from ionic liquids. The cathodic limits of e.g. the 1,1-dialkylpyrrolidinium-based ionic liquids are below the electrodeposition potential for lithium (around – 3 V vs. NHE), thus it should be trivial to deposit vanadium (−1.17 V), titanium (−1.21 V) and magnesium (−2.34 V), to mention a few. However, thus far there has been no convincing report on the electrodeposition of magnesium or titanium available in the literature, and we ourselves needed almost two years to find a suitable way to electrodeposit crystalline tantalum in micrometer thick layers. In the case of tantalum deposition we started initially with 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide, a liquid which has a cathodic limit of about −3 V vs. NHE on Au(111), in which we dissolved TaCl5 (0.1–0.3 mol l−1 ). Indeed we got a relatively simple cyclic voltammogram with two reduction processes and poorly defined oxidation reactions. We were quite optimistic in the beginning as the second reduction process corresponds to the formation of a black deposit which was potentially the first electrochemical route to make thick tantalum layers. After having washed off all ionic liquid from the sample we were already a bit sceptical as the deposit was quite brittle and did not look metallic. The SEM pictures and the EDX analysis supported our scepticism and the elemental analysis showed an elemental Ta/Cl ratio of about 1/2. Thus, overall we have deposited a low oxidation state tantalum choride. Despite the initial disappointment we were still eager to obtain the metal and found some old literature from Cotton [122], in which he described subvalent clusters of molybdenum, tungsten and tantalum halides. In the case of tantalum the well-defined Ta6 Cl12 2+ complex was described with an average oxidation number of 2.33 and thus with a Ta/Cl molar ratio very close to 1/2. Such clusters are depicted in Figure 4.15. In these clusters tantalum atoms are bound to other tantalum atoms and are also edge bridged via halide. As our deposit was completely amorphous without any XRD peak we concluded that it did not consist of crystalline tantalum but rather of such clusters. We varied the electrode potential for deposition and tried deposition with very low constant current densities, but in no case was crystalline tantalum obtained. Thus, the electrochemical window of our liquid was surely wide enough, but for some reason the electrodeposition stopped before Ta(0) was obtained. When we studied the literature dealing with metal clusters we found that the cluster chemistry with fluoride seems to be less comprehensive. Consequently

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.4 Troublesome Aspects 115

Fig. 4.15 Some metal/halide clusters described in 1969 by Cotton, Ref. [122].

we tried the deposition in the same liquid with TaF5 (0.25 mol l−1 ) as a source of tantalum. In Figure 4.16 the cyclic voltammogram of TaF5 in the above-mentioned liquid is shown at three different temperatures [123]. The cyclic voltammogram is quite similar to the voltammogram of TaCl5 and consists mainly of two reduction processes. There is no visible surface process at

Fig. 4.16 Cyclic voltammogram of TaF5 in 1-butyl-1methylpyrrolidinium bis(trifluoromethylsulfonyl)amide at variable temperature. The second cathodic process is correlated to the deposition of a black amorphous material.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

116 4 Electrodeposition of Metals

Fig. 4.17 SEM picture of a “tantalum” deposit made from TaF5 in 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide. The deposit is obviously amorphous, with EDX a Ta/F ratio of 4/1 is obtained.

the first cathodic peak, at the second cathodic peak a black deposit again forms. Unfortunately the deposit is still XRD amorphous, and the SEM picture does not show a crystalline material, Figure 4.17. However, the EDX analysis made us more optimistic as we got an atomic Ta/F ratio of 4/1. Our assumption that with fluoride we would get a lower amount of subvalent tantalum fluorides seemed to be right. Furthermore, underneath the black brittle deposit there was always a thin shining layer which looked metallic. An in situ STM study showed that triangularly shaped crystals grew on the nanoscale with a typically metallic behavior in the tunnelling spectrum [123] which, taken both together, is quite unusual for an amorphous deposit. Upon addition of LiF we finally found (see Chapter 4.2) parameters with which we could deposit crystalline tantalum layers with thicknesses of about 1 µm. The key parameter was a low current density for the electrodeposition. With high current densities in the 10 mA cm−2 region we mainly got an amorphous deposit, whereas with current densities of about 10 µA cm−2 we got metallic tantalum, although the deposition rate was naturally slow. At least three aspects have to be considered: 1. It is known in inorganic chemistry that all refractory and rare earth elements tend to form subvalent halides from their iodides, bromides and chlorides. 2. In the case of tantalum deposition the addition of LiF was required. On the one hand the Li+ ion might destabilize the Ta–F bond and thus facilitate the deposition of tantalum, on the other hand Li+ ions might influence the electrochemical double layer and facilitate charge transfer. We are about to perform in situ STM studies, the results will be reported in the peer reviewed literature.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.4 Troublesome Aspects 117

3. Mass transport may influence material growth in ionic liquids. 1-Butyl-1methylpyrrolidinium bis(trifluoromethylsulfonyl)amide, for example, is, at room temperature, about 60 times more viscous than water. At temperatures above 150 ◦ C its viscosity is similar to most molecular solvents at ambient conditions. Indeed, temperatures between 150 and 200 ◦ C were best to deposit tantalum from TaF5 in the presence of LiF. One has to keep in mind that the deposition of tantalum from TaF5 or an anionic complex delivers one Ta atom and 5–7 fluorides. If the deposition is too fast F− may not diffuse rapidly enough from the surface to the bulk of the solution and may be trapped in the deposit. This might explain why we only got crystalline tantalum layers at low current densities.

In our opinion non-stoichiometric metal halide compounds have to be expected for the electrodeposition of refractory and rare earth metals if the deposition is performed from halides as precursors. The electrodeposition of these metals requires tailor-made metal salts. 4.4.2 Viscosity/Conductivity

The view of many electroplaters is that the viscosity of ionic liquids is too high and the specific conductivity too low to be viable for the deposition of metals. At room temperature it is surely right that many liquids are viscous compared with aqueous solutions. However, it is totally neglected that the situation changes completely when the liquids are heated. Even at a moderate temperature, i.e. 100 ◦ C, many liquids have viscosities of only a few mPa s, quite similar to water, thus reducing IR drops in electrochemical experiments considerably. From our point of view we recommend: “If it doesnt work at room temperature, just heat it up!”. As some ionic liquids have practical thermal windows as high as 200–300 ◦ C they can be regarded as the missing link to high-temperature molten salts [124]. Methods for estimating maximum process operating times and temperatures have been developed [125]. Thus, variation of temperature is rather a benefit and we ourselves were able to show that at T > 100 ◦ C the grey phase of selenium can be electrodeposited exclusively [126]. Furthermore one should take into account that there is constant progress in the synthesis of ionic liquids, thus it can be expected that liquids with viscosities around 5 mPa s at room temperature (just 5 times more than water) will be available in future. 4.4.3 Impurities

It is commonly accepted in the ionic liquids community that the purification of ionic liquids can be relatively complex. Currently, they cannot be distilled at reasonable rates, crystallized or sublimed. Thus, the only reasonable solution is to synthesize them from high quality starting materials. Apart from organic impurities

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

118 4 Electrodeposition of Metals

(decomposition products of anions and/or cations, side products) halide, Li+ and K+ are common inorganic impurities. Li+ and K+ can be found in the 1000 ppm regime if the liquids are made by metathesis reaction from metal salts and organic halides. Sometimes even low amounts of impurities washed off from silica or alumina (often used to remove the yellowish color of ionic liquids) can be found in the liquids. These impurities can only be removed by extensive washing with highest quality water or by an electrochemical treatment with separated cathodic and anodic compartments. Water is introduced during the washing process, but usually it can easily be removed by putting the liquids under vacuum at elevated temperature. Water concentrations of 3 ppm and below are easily obtained for 1butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide with this method. For a technical electrochemical process 10 ppm of Li+ or K+ might be negligible, for fundamental electrochemical studies on the nanoscale with the in situ STM (Chapter 9) it is rather a nightmare if there is underpotential deposition of lithium in a potential regime where the deposition of e.g. silicon is expected. It took us four months to confirm results for silicon electrodeposition because of the contamination of one of our liquids with Li+ (see also Refs. [127, 128]). As will be shown in a later paper the decomposition of the organic cation of an ionic liquid (e.g. by high galvanostatic pulses) can also strongly alter the morphology of materials. Whereas the electrodeposition in 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide usually delivers a microcrystalline aluminum, a nanocrystalline deposit is obtained if the deposition is performed during cathodic breakdown of the imidazolium cation. This is also a kind of in situ made impurity which can strongly alter electrodeposition. Although almost all of the modern ionic liquids are per se air- and water-stable one has to bear in mind that upon addition of SiCl4 , TaF5 , SeCl4 , AlCl3 and other moisture sensitive compounds the resulting solutions are no longer water-stable. Thus, inert gas conditions are required to get reproducible results. From our experience only ultrapure ionic liquids should be employed for fundamental electrochemical studies unless the influence of impurities has been understood. Our own experience and that of many other groups have shown that even a few ppm of impurities can strongly alter fundamental studies.

4.4.4 Additives

There is a phalanx of different additives available in aqueous electroplating. We have often had the experience that aqueous electroplaters add personal additive recipes to ionic liquids and are surprised, or even disappointed, that they do not work. One has to bear in mind that all known additives were developed over several decades and their mode of action in aqueous solutions is still not fully understood. How can it be expected that one can just replace water by the ionic liquid and get the same or even better results? In our opinion a deep understanding of cation/anion

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

4.4 Troublesome Aspects 119

interactions with dissolved substances is required to develop additives that are suited for ionic liquids. 4.4.5 Cation/Anion Effects

We were the first to find that there are cation/anion effects in the electrodeposition of metals. In the case of aluminum we found that it is deposited as a nanomaterial in 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide from AlCl3 , whereas it is deposited as a microcrystalline material under quite similar conditions in 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide [129]. The likely explanation is that the pyrrolidinium ion interferes with the electrode surface and the growing nuclei, thus hindering crystal growth [130]. Maybe these cation/anion effects explain why in first generation ionic liquids there are about 100 papers on the electrodeposition of Al and its alloys from AlCl3 and 1-ethyl-3-methylimidazolium chloride and only a few with tetraalkylammonium chlorides/AlCl3 . Our own experiments have shown that the deposition of Al from 1-butyl-1-methylpyrrolidinium chloride/AlCl3 delivers a crystalline but rather flakelike black product. Thus it might be a bad choice to employ cheap liquids for electrodeposition. In our opinion a deep understanding of cation/anion interferences is required and one should be aware of unexpected effects by just employing a different ionic liquid. 4.4.6 Price

A main reproach is that the cost of ionic liquids is too high at the moment. Indeed, 1 kg of ultrapure 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide costs up to 2000 € . One has to bear in mind that currently one pays more or less the salary of the technician in the laboratory who synthesizes the liquid from the educts. If a large scale production line was available, operated automatically, the costs would be reduced drastically. There is a dispute in the community about what future prices will be. Currently it is believed that in a few years from now the prices will start at about 10 € per liter for standard ionic liquids with prices up to 10 000 € per kg for tailor-made “research liquids”. The first generation ionic liquids based on AlCl3 and dialkylimidazolium chlorides are candidates for such comparatively cheap liquids and a price between 10 and 15 € per kg is conceivable. One should not forget that ionic liquids have practically no vapor pressure and that they can easily be recycled, as shown in Chapter 11.4. Thus the overall costs for a process will decide whether an ionic liquids process will be established or not. In the case of Al electrodeposition there would be an immediate advantage of ionic liquids: in contrast to the SIGAL process where Al is deposited from explosive alkyl-aluminum compounds, thus strictly requiring inert gas, dry air would be sufficient in the case

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

120 4 Electrodeposition of Metals

of an ionic liquids process. It is likely that the overall costs would be at the same level or even lower. 4.4.7 One Liquid for All Purposes?

The dream would be that there will be – like water – one ionic liquid that is suited as a general liquid for all electrochemical reactions. It cannot be excluded that such a liquid will be produced in the future, but at present the field is in rather a developmental state. We ourselves were pretty surprised when we realized that the cation of an ionic liquid can have a dramatic effect on the electrodeposition of metals. A deeper understanding of ionic liquids will be required before ionic liquids become standard electrolytes for electroplating. The motivation of this chapter was to show that despite the enormous prospects of ionic liquids in electrodeposition some troublesome aspects have to be expected. Apart from the purity and price of ionic liquids the optimum temperature for any process has to be found. Furthermore, suitable additives for electrodeposition will have to be developed and cation/anion effects that can strongly alter the morphology of deposits have to be expected. Finally, the electrochemical window alone is not the only factor that needs to be considered for the deposition of reactive metals. Suitable precursors will have to be tailor-made and it is our personal opinion that the electrodeposition of metals like Mg, Ti, Ta and Mo may not be possible from metal halides but rather metal bis(trifluoromethylsulfonyl)amide salts and other ones may be more suitable. References 1 Melton, T.J., Joyce, J., Maloy, J.T., Boon, J.A., and Wilkes, J.S. (1990) J. Electrochem. Soc., 137, 3865. 2 Lipsztajn, M. and Osteryoung, R.A. (1985) Inorg. Chem., 24, 716–719. 3 Piersma, B.J., Ryan, D.M., Schumacher, E.R., and Reichel, T.L. (1996) J. Electrochem. Soc., 143, 908–913. 4 Gray, G.E., Kohl, P.A., and Winnick, J. (1995) J. Electrochem. Soc., 142, 3636–3642. 5 Kim, K., Lang, Ch., Moulton, R., and Kohl, P.A. (2004) J. Electrochem. Soc., 151, A1168–A1172. 6 Robinson, J. and Osteryoung, R.A. (1980) J. Electrochem. Soc., 127, 122–128. 7 Welch, B.J. and Osteryoung, R.A. (1981) J. Electroanal. Chem., 118, 456–466. 8 Lai, P.K. and Skyllas-Kazacos, M. (1988) J. Electroanal. Chem., 248, 431–440.

9 Liao, Q., Pitner, W.R., Stewart, G., Hussey, Ch.L., and Stafford, G.R. (1997) J. Electrochem. Soc., 144, 936–943 10 Abbott, A.P., Eardley, Ch.A., Farley, N.R.S., Griffith, G.A., and Pratt, A. (2001) J. Appl. Electrochem., 31, 1345–1350. 11 Endres, F., Bukowski, M., Hempelmann, R., and Natter, H. (2003) Angew. Chem. Int. Ed., 42, 3428–3430. 12 Liu, J.S.-Y. and Sun, I.W. (1997) J. Electrochem. Soc., 144, 140. 13 Carpenter, M.K. (1994) J. Mater. Res., 9, 2584. 14 Chen, P.Y., Lin, Y.F., and Sun, J.W. (1999) J. Electrochem. Soc., 146 (9), 3290–3294. 15 Pitner, W.R. and Hussey, C.L. (1997) J. Electrochem. Soc., 144, 3095. 16 Habboush, D.A. and Osteryoung, R.A. (1984) Inorg. Chem., 23, 1726–1734.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

References 17 Lipsztajn, M. and Osteryoung, R.A. (1985) Inorg. Chem., 24, 3492. 18 Jeng, E.G.-S. and Sun, I.-W. (1997) J. Electrochem. Soc., 144, 2369–2374. 19 Bonhˆote, P., Dias, A., Papageorgiou, N., Kalyanasundaram, K., and Gr¨atzel, M. (1996) Inorg. Chem., 35, 1168. 20 Fuller, J. and Carlin, R.T. (1998) in Molten Salts (eds P.C. Trulove, H.C. De Long, G.R. Stafford, and S. Deki), PV 98–11, p. 227, The Electrochemical Society Proceedings Series, Pennington, NJ. 21 MacFarlane, D.R., Meakin, P., Sun, J., Amini, N., and Forsyth, M. (1999) J. Phys Chem. B., 103, 4164. 22 Abbott, A.P. and McKenzie, K.J. (2006) Phys. Chem. Chem. Phys., 8, 4265. 23 Zein El Abedin, S. and Endres, F. (2006) Chem. Phys. Chem., 7, 58. 24 Buzzeo, M.C., Evans, R.G., and Compton, R. (2004) Chem. Phys. Chem., 5, 1106. 25 Galinski, M., Lewandowski, A., and Stepniak, I. (2006) Electrochim. Acta, 51, 5567. 26 Lin, Y.-F. and Sun, I-W. (1999) Electrochim. Acta, 44, 2771. 27 Chen, P.-Y., Lin, M.-C., and Sun, I-W. (2000) J. Electrochem. Soc., 147, 3350. 28 Chen, P.-Y. and Sun, I-W. (2001) Electrochim. Acta, 46, 1169. 29 Abbott, A.P., Capper, G., McKenzie, K.J., and Ryder, K.S. (2007) J. Electroanal. Chem., 599, 288. 30 Endres, F. and Schweizer, A. (2000) Phys. Chem. Chem. Phys., 2, 5455. 31 Chen, P.-Y. and Sun, I.-W. (1999) Electrochim. Acta, 45, 441. 32 Murase, K., Nitta, K., Hirato, T., and Awakura, Y. (2001) J. Appl. Electrochem., 31, 1089. 33 Zein El Abedin, S., Saad, A.Y., Farag, H.K., Boressinko, N., Liu, Q.X., and Endres, F. (2007) Electrochim. Acta, 52, 2746. 34 Rostovshchikova, T.N., Smirnov, V.V., Kozhevin, V.M., Yavsin, D.A., Zabelin, M.A., Yassievich, I.N., and Gurevic, S.A. (2005) Appl. Catal. A: General, 296, 70. 35 Bube, R.H. and Mitchell, K.W. (1993) J. Electron. Mater., 22, 17.

36 Iwai, K. and Kojima, T. (1982) Prog. Batteries Sol. Cells, 4, 491. 37 Chen, P.-Y. and Sun, I.-W. (2000) Electrochim. Acta, 45, 3163. 38 Huang, J.-F. and Sun, I.-W. (2002) J. Electrochem. Soc., 149, E348. 39 Abbott, A.P., Capper, G., Davies, D.L., and Rasheed, R.K. (2004) Chem. Eur. J., 10, 3769. 40 Abbott, A.P., Capper, G., Davies, D.L., Rasheed, R.K., Archer, J., and John, C. (2004) Trans. Inst. Met. Finish., 82, 14. 41 Somorjai, G.A. (1994) Introduction to Surface Chemistry and Catalysis, John Wiley & Sons, Inc., New York. 42 Tai, C.-C., Su, F.-Y., and Sun, I.-W. (2005) Electrochim. Acta, 50, 5504. 43 Su, F.-Y., Huang, J.-F., and Sun, I.-W. (2004) J. Electrochem. Soc., 151, C811. 44 Hsiu, S.-I., Tai, C.-C., and Sun, I.-W. (2006) Electrochim. Acta, 51, 2607. 45 Xu, X.-H. and Hussey, C.L. (1992) J. Electrochem. Soc., 139, 1295. 46 Endres, F. and Freyland, W. (1998) J. Phys. Chem. B, 102, 10229. 47 Zell, C.A., Endres, F., and Freyland, W. (1999) Phys. Chem. Chem. Phys., 1, 697. 48 Katayama, Y., Dan, S., Miura, T., and Kishi, T. (2001) J. Electrochem. Soc., 148, C102. 49 He, P., Liu, H.T., Li, Z.Y., Liu, Y., Xu, X.D., and Li, J.H. (2004) Langmuir, 20, 10260. 50 Meiss, S.A., Rohnke, M., Kienle, L., Zein El Abedin, S., Endres, F., and Janek, J. (2007) Chem. Phys. Chem., 8, 50. 51 He, P., Liu, H., Li, Z., and Li, J. (2005) J. Electrochem. Soc., 152, E146. 52 Huang, J.-F. and Sun, I.-W. (2004) Electrochim. Acta, 49, 3251. 53 Yang, M.-H. and Sun, I.-W. (2003) J. Appl. Electrochem., 33, 1077. 54 Yang, M.-H., Yang, M.-C., and Sun, I.-W. (2003) J. Electrochem. Soc., 150, C544. 55 Brukin, A.R. (1987) Production of Aluminum and Alumina, Critical Reports in Applied Chemistry, Vol. 20, John Wiley & Sons, Ltd., Chichester. 56 Ziegler, K. and Lehmkuhl, H. (1956) Chemie, 283, 414.

121

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

122 4 Electrodeposition of Metals

57 Kautek, W. and Birkle, S. (1989) Electrochim. Acta, 34, 1213. 58 Zhao, Y. and VanderNoot, T.J. (1997) Electrochim. Acta, 42, 3. 59 Melton, T.J., Joyce, J., Maloy, J.T., Boon, J.A., and Wilkes, J.S. (1990) J. Electrochem. Soc., 137, 3865. 60 Carlin, R.T. and Osteryoung, R.A. (1989) J. Electrochem. Soc., 136, 1409. 61 Moffat, T.P. (1994) J. Electrochem. Soc., 141, L115. 62 Ali, M.R., Nishikata, A., and Tsuru, T. (1997) Electrochimica Acta, 42, 2347. 63 Stafford, G.R. (1994) J. Electrochem. Soc., 141, 945. 64 Tsuda, T., Hussey, C.L., and Stafford, G.R. (2004) J. Electrochem. Soc., 151, C379. 65 Stafford, G.R. (1989) J. Electrochem. Soc., 136, 635. 66 Endres, F., Bukowski, M., Hempelmann, R., and Natter, H. (2003) Angew. Chem., 115, 3550; (2003) Angew. Chem. Int. Ed. Engl., 42, 3428. 67 Liu, Q.X., Zein El Abedin, S., and Endres, F. (2006) Surf. Coat. Technol., 201, 1352. 68 Jiang, T., Chollier, M.J., Dube, G., Lasia, A., and Brisard, G.M. (2006) Surf. Coat. Technol., 201, 1. 69 Zein El Abedin, S., Moustafa, E., Hempelmann, R., Natter, H., and Endres, F. (2005) Electrochem. Commun., 7, 1111. 70 Zein El Abedin, S., Moustafa, E., Hempelmann, R., Natter, H., and Endres, F. (2006) Chem. Phys. Chem., 7, 1535. 71 Brausch, N., Metlen, A., and Wasserscheid, P. (2004) Chem. Commun., 13, 1552. 72 NuLi, Y., Yang, J., and Wang, P. (2006) Appl. Surf. Sci., 252, 8086. 73 Wang, P., NuLi, Y., Yang, J., and Feng, Z. (2006) Surf. Coat. Technol., 201, 3783. 74 Feng, Z., NuLi, Y., Wang, J., and Yang, J. (2006) J. Electrochem. Soc., 135, C689. 75 NuLi, Y., Yang, J., and Wu, R. (2005) Electrochem. Commun., 7, 1105. 76 Cheek, G.T., O’Grady, W.E., Zein El Abedin, S., Moustafa, E.M., and Endres, F., J. Electrochem. Soc., accepted.

77 Aurbach, D., Turgeman, R., Chusid, O., and Gofer, Y. (2001) Electrochem. Commun., 3, 252. 78 Liebnow, C., Yang, Z., and Lobitz, P. (2000) Electrochem. Commun., 2, 641. 79 Katayama, Y., Morita, T., Yamagata, M., and Miura, T. (2003) Electrochemistry, 71, 1033. 80 Howlett, P.C., MacFarlane, D.R., and Hollenkamp, A.F. (2004) Electrochem. Solid-State Lett., 7, A97. 81 Mellors, G.W. and Senderoff, S. (1965) J. Electrochem. Soc., 112, 840. 82 Senderoff, S. and Mellors, G.W. (1966) Science, 153, 1475. 83 Dutra, A.J.B., Vazquez, J.C., and Espinola, A. (1993) Miner. Eng., 6, 663. 84 Chamelot, P., Taxil, P., and Lafage, B. (1994) Electrochim. Acta, 39, 2571. 85 Chamelot, P., Palau, P., Massot, L., Savall, A., and Taxil, P. (2002) Electrochim. Acta, 47, 3423. 86 Lantelme, F., Barhoun, A., Li, G., and Besse, J.P. (1992) J. Electrochem. Soc., 139, 1255. 87 Haarberg, G.M. and Thonstad, J. (1989) J. Appl. Electrochem., 119, 789. 88 Zein El Abedin, S., Farg, H.K., Moustafa, E.M., Welz-Biermann, U., and Endres, F. (2005) Phys. Chem. Chem. Phys., 7, 2333. 89 Zein El Abedin, S., Welz-Biermann, U., and Endres, F. (2005) Electrochem. Commun., 7, 941. 90 El Medawar, L., Rocher, P., Hornez, J.-C., Traisnel, M., Breme, J., and Hildebrand, H.F. (2002) Biomol. Eng., 19, 153. 91 Gil, F.X., Manero, J.M., and Planell, J.A. (1996) J. Mater. Sci.: Mater. Med., 7, 403. 92 Yachia, D. and Beyar, M. (1993) Eur. Urol., 24, 500. 93 Jia, W.Y., Beatty, M.W., Reinhardt, R.A., Petro, T.M., Cohen, D.M., Maze, C.R., Strom, E.A., and Hoffman, M. (1999) J. Biomed. Mater. Res., 48, 488. 94 Uo, M., Watari, F., Yokoyama, A., Matsuno, H., and Kawasaki, T. (1999) Biomaterials, 20, 747. 95 Shih, C., Lin, S., Chen, Y., Su, Y., Lai, S., Wu, G.J., Kwok, C., and Chung, K. (2000) J. Biomed. Mater. Res., 52, 395.

c04 (JWBG008-Endres)

December 25, 2007

13:28

Char Count=

References 96 Mukhopadhyay, I., Aravinda, C.L., Borissov, D., and Freyland, W. (2005) Electrochim. Acta, 50, 1275. 97 Howlett, P.C., Izgorodina, E., Forsyth, M., and MacFarlane, D.R. (2006) Z. Phys. Chem., 220, 1483. 98 Endres, F., Zein El Abedin, S., and Borissenko, N. (2006) Z. Phys. Chem., 220, 1377. 99 Abbott, A.P. and McKenzie, K.J. (2006) Phys. Chem. Chem. Phys., 8, 4265–4279. 100 Endres, F. (2002) Chem. Phys. Chem., 3, 144. 101 Endres, F. and Zein El Abedin, S. (2006) Phys. Chem. Chem. Phys., 8, 2101. 102 Endres, F. (2004) Z. Phys. Chem., 218, 255. 103 Gale, R.J. and Osteryoung, R.A. (1980) Electrochim. Acta, 25, 1527. 104 Nanjundiah, C., McDevitt, S.F., and Koch, V.R. (1997) J. Electrochem. Soc., 144, 3392; (b) Nanjundiah, C., Goldman, J.L., McDevitt, S.F., and Koch, V.R. (1997) Proc. Electrochem. Soc., 96–25, 301. 105 Hsiu, S.-I., Huang, J.-F., Sun, I.-W., Yuan, C.-H., and Shiea, J. (2002) Electrochim. Acta, 47, 4367–4372. 106 Lin, Y.-F. and Sun, I.-W. (1999) Electrochim. Acta, 44, 2771. 107 Koyama, K., Iwagishi, T., Yamamoto, H., Shirai, H., and Kobayashi, H. (2002) Electrochem., 70, 178–182. 108 Iwagishi, T., Yamamoto, H., Koyama, K., Shirai, H., and Kobayashi, H. (2002) Electrochem., 70, 671–674. 109 Abbott, A.P., Capper, G., Davies, D.L., Munro, H., Rasheed, R., and Tambyrajah, V. (2003) in (eds R.D. Rogers and K.R. Seddon), , Ionic Liquids as Green Solvents: Progress and Prospects 439. 110 Abbott, A.P., Capper, G., Davies, D.L., Munro, H., and Rasheed, R. (2004) Inorg. Chem., 43, 3447. 111 Katayama, Y., Konishiike, I., Miura, T., and Kishi, T. (2002) J. Power Sources, 109, 327–332. 112 Verbrugge, M.W. and Carpenter, M.K. (1990) AIChE J., 36, 1097. 113 Wicelinski, S.P. and Gale, R.J. (1987) Proc. Electrochem. Soc., 134, 262.

114 Carpenter, M.K. and Verbrugge, M.W. (1987) J. Electrochem. Soc., 87, 591 115 Carpenter, M.K. and Verbrugge, M.W. (1994) J. Mater. Res., 9 (2), 584. 116 Abbott, A.P., Capper, G., Davies, D.L., and Rasheed, R. (2004) Chem. Eur. J., 10, 3769. 117 Abbott, A.P., Capper, G., Davies, D.L., Rasheed, R.K., Archer, J., and John, C. (2004) Trans. Inst. Metal Finish, 82, 14. 118 Abbott, A.P., Capper, G., McKenzie, K.J., and Ryder, K.S. (2007) J. Electroanal. Chem., 599, 288. 119 Abbott, A.P., Capper, G., Davies, D.L., Rasheed, R., and Shikotra, P. (2005) Inorg Chem., 44, 6497. 120 Abbott, A.P., Capper, G., McKenzie, K.J., and Shikotra, P. (2006) J. Chem. Eng. Data, 51, 1280–1282. 121 Abbott, A.P., Capper, G., and Shikotra, P. (2006) Trans. Inst. Miner. Metall. C, 115, 15. 122 Cotton, F.A. (1969) Acc. Chem. Res., 2, 240. 123 Zein El Abedin, S., Farag, H.K., Moustafa, E.M., Welz-Biermann, U., and Endres, F. (2005) Phys. Chem. Chem. Phys., 7, 2333. 124 Endres, F. and Zein El Abedin, S., Acc. Chem. Res., ASAP Article, 10.1021/ar700049w. 125 Baranyai, K.J., Deacon, G.B., MacFarlane, D.R., Pringle, J.M., and Scott, J.L. (2004) Aust. J. Chem., 57, 145. 126 Zein El Abedin, S., Saad, A.Y., Farag, H.K., Borissenko, N., Liu, Q., and Endres, F. (2007) Electrochim. Acta, 52, 2746. 127 Endres, F., Zein El Abedin, S., and Borissenko, N. (2006) Z. Phys. Chem., 220, 1377. 128 Borisenko, N., Zein El Abedin, S., and Endres, F. (2006) J. Phys. Chem. B, 110 (12), 6250–6256. 129 Zein El Abedin, S., Moustafa, E.M., Hempelmann, R., Natter, H., and Endres, F. (2006) Chem. Phys. Chem., 7, 1535–1543. 130 Moustafa, E.M., Zein El Abedin, S., Shkurankov, A., Zschippang, E., Saad, A.Y., Bund, A., and Endres, F. (2007) J. Phys. Chem. B, 111 (18), 4693–4704.

123

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

125

5 Electrodeposition of Alloys I.-Wen Sun, and Po-Yu Chen

5.1 Introduction

Electrodeposition of alloys is an important subject as alloys often provide properties superior to those of single-metal electrodeposits. The electrodeposited alloys can be more corrosion resistant, more wear resistant, better in catalytic properties and better in magnetic properties. Similar to the deposition of pure metals, the properties of the electrodeposited alloys can be varied by experimental factors such as plating bath composition, current density, overpotential and temperature. Furthermore, applying pulsed electrodeposition allows one to influence the grain size of the deposits (see Chapter 9). The interest in the investigation of electrodeposition of alloys is increasing, quite simply because the number of alloy combinations is vast. While aqueous plating baths are widely employed for the electrodeposition of alloys, the limited electrochemical window and hydrogen evolution problems have considerably restricted the number of alloys that can be electrodeposited from aqueous plating baths. Over the past two decades, ionic liquids (ILs) have attracted considerable interest as media for a wide range of applications. For electrochemical applications they exhibit several advantages over the conventional molecular solvents and high temperature molten salts: they show good electrical conductivity, wide electrochemical windows of up to 6 V, low vapor pressure, non-flammability in most cases, and thermal windows of 300–400 ◦ C (see Chapter 4). Moreover, ionic liquids are, in most cases, aprotic so that the complications associated with hydrogen evolution that occur in aqueous baths are eliminated. Thus ILs are suitable for the electrodeposition of metals and alloys, especially those that are difficult to prepare in an aqueous bath. Several reviews on the electrodeposition of metals and alloys in ILs have already been published [1–4]. A selection of published examples of the electrodeposition of alloys from ionic liquids is listed in Table 5.1 [5–40]. Ionic liquids can be classified into water/air sensitive and water/air stable ones (see Chapter 3). Historically, the water-sensitive chloroaluminate first generation ILs are the most intensively studied. However, in future the focus will rather be on air- and waterstable ionic liquids due to their variety and the less strict conditions under which Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

126 5 Electrodeposition of Alloys Table 5.1 Metal alloys that have been electrodeposited from ionic liq-

uids. Alloy

Ref.

Ionic liquid

Al–Nb Al–Ni Al–Co Al–Cr Al–Cu Al–Mn Al–La Al–Ag Al–Ti Al–Mo Al–Zr Al–Pt Al–Mg Al–Mo–Mn Al–Cr–Ni Zn–Cu Zn–Cd Zn–Sn Zn–Co Zn–Fe Zn–Ni Zn–Mg Zn–Pt Pt–Zn Au–Zn Ag–Zn Nb–Sn Pd–Au Pd–Ag Pd–In In–Sn Cu–Sn Zn–Mn

[5] [6] [7–10] [11] [12] [13] [14] [15] [16, 17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27, 28] [29] [30] [31] [32] [34] [35] [36] [37] [38] [39] [40] [41] [42] [46]

[EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 , [BP]+ Cl− /AlCl3 [BP]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 , [BMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [BTMA]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /AlCl3 [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 [BP]+ Cl− / ZnCl2 , [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 /NiCl2 [EMIM]+ Br− /ZnBr2 /MgBr2 /EG [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /ZnCl2 [EMIM]+ Cl− /SnCl2 /NbCl5 [EMIM]+ BF4 − [EMIM]+ BF4 − [EMIM]+ BF4 − [EMIM]+ BF4 − [TMHA]+ Tf2 N− [TBMA]+ Tf2 N−

they can be handled. The principles of alloy electrodeposition are beyond the scope of this book therefore for an overview on alloy electrodeposition the book by Brenner [44] is recommended.

5.2 Electrodeposition of Al-containing Alloys from Chloroaluminate Ionic Liquids

The Lewis acidic chloroaluminate ILs are suitable for the electrodeposition of aluminum-containing alloys. Many examples have been published but those that have been reviewed in detail by Stafford and Hussey [1] will not be included in this section.

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.2 Electrodeposition of Al-containing Alloys from Chloroaluminate Ionic Liquids 127

5.2.1 Al–Ti

The electrodeposition of Al–Ti alloy has been investigated in the acidic aluminum chloride-1-ethyl-3-methylimidazolium chloride ([EMIM]+ Cl− /AlCl3 ) ionic liquid containing Ti(II) up to 0.17 mol l−1 at 353 K [16]. Such alloys are technically interesting due to their high temperature resistance. Ti(II) can be introduced into the liquid by direct dissolution of TiCl2 or by the reduction of TiCl3 with Al metal in the liquid. It was proposed that TiCl2 dissolves in the liquid by forming [Ti(AlCl4 )3 ]− and its solubility increases with increasing IL acidity. TiCl2 (s) + 2Al2 Cl7 − → [Ti(AlCl4 )3 ]− + AlCl4 −

(5.1)

Lowering the liquid acidity from 66.7–33.3% to 60.0–40.0% mole fraction results in the disproportionation of [Ti(AlCl4 )3 ]− producing TiCl3 and Ti precipitates. 3[Ti(AlCl4 )3 ]− + 3AlCl4 − ↔ 2TiCl3 (s) + Ti + 6Al2 Cl7 −

(5.2)

Ti(II) tends to form polymers or aggregates upon increasing the Ti(II) concentration or the liquid acidity. Electrochemical, either galvanostatic or potentiostatic, oxidation of Ti metal produces either passive TiCl3 film or volatile TiCl4 which escapes from the liquid. The oxidation of metallic titanium to Ti(II) by direct anodization of Ti metal in this liquid has not yet been described. Cyclic voltammograms recorded on polycrystalline stationary and rotating Pt disk electrodes in the acidic [EMIM]+ Cl− /AlCl3 ionic liquids demonstrated that the reduction of Ti(II) to Ti(0) occurs at a potential where the deposition of Al also occurs. As an Al–Ti alloy forms, both the deposition and stripping waves shift to values more positive than the pure Al oxidation. The magnitude of the shifts increases with increasing Ti(II) concentration. Bulk deposits of Al–Ti alloys were prepared by using DC galvanostatic electrolysis on a copper rotating disk electrode (Cu-RDE) at a current density of −10 mA cm−2 in a Ti(II)-saturated liquid. The Ti metal content of bulk Al–Ti alloys prepared in this way decreased with increasing applied current density, suggesting that the reduction potential of the Ti(II)/Ti couple would be positive of that for the Al(III)/Al couple. Increasing the total applied reduction current densities or making the applied potential more negative probably leads to a mass-transport-limited value for Ti deposition, whereas the partial current density for the deposition of Al still increases, resulting in alloy deposits with decreased Ti content. The Ti content in the Al–Ti electrodeposited from this ionic liquid is limited by the solubility of Ti(II) in the liquid and by the minimum practical current density that can be applied. Scanning electron micrographs of the electrodeposited Al–Ti alloys revealed that the deposits were compact dense nodules of single crystals. The nodule size decreases with decreasing current density and increasing Ti content. X-ray powder diffraction (XRD) patterns of the electrodeposits containing 7.0 to 18.4 a/o Ti metal showed a disordered face-centered cubic (fcc) structure, very similar to that of pure Al. The aluminum lattice parameter decreases as the smaller Ti atoms substitute for Al. Furthermore, X-ray reflections

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

128 5 Electrodeposition of Alloys

of the Al–Ti alloys broaden with increasing Ti content, suggesting a decrease in the grain size of the deposit. Potentiodynamic anodic polarization curves recorded for Al–Ti alloys electrodeposited on copper electrodes in deaerated aqueous NaCl solution revealed that, similar to what is found for Al–Ti alloys prepared by sputter deposition, the electrodeposited Al–Ti alloys exhibit a significant increase in pitting potential relative to pure Al. The electrodeposition of Al–Ti alloys has also been examined at 298 K on Au(111) in an acidic aluminum chloride-1-butyl-3-methylimidazolium chloride ([BMIM]+ Cl− /AlCl3 ) containing 10 mM TiCl4 [17]. Cyclic voltammograms showed that Ti(IV) can be electrochemically reduced to Ti(III) in the form of hardly soluble TiCl3 , which can be further reduced to Ti(II) at a potential close to the underpotential deposition (UPD) of Al on Au(111), followed by the co-deposition of Al–Ti prior to the Al bulk deposition. The stripping of Al–Ti can be observed during the anodic scan. Comparing the electrochemical scanning tunneling microscopy (EC-STM) images of the deposits revealed that Al UPD clusters preferentially deposit along the Au step edges in the absence of Ti whereas the UPD of Al–Ti begins with the formation of monoatomically high clusters on the Au terraces without any site preference. The formation of an Al–Ti phase in the electrodeposits was further confirmed by X-ray photoelectron spectra (XPS) analysis. 5.2.2 Al–Mo

The electrodeposition of Al–Mo high-temperature and corrosion resistant alloy was investigated in a Lewis acidic [EMIM]+ Cl− /AlCl3 ionic liquid using the octahedral hexanuclear Mo(II) cluster compound, (Mo6 Cl8 )Cl4 [18]. A previous study [42] showed that in basic [EMIM]+ Cl− /AlCl3 liquid, (Mo6 Cl8 )Cl4 picks up two excess chloride ions from the liquid to form [(Mo6 Cl8 )Cl6 ]2− complex anion but the reduction of this species does not produce Mo metal. The (Mo6 Cl8 )Cl4 is soluble in the acidic [EMIM]+ Cl− /AlCl3 and preserves its {Mo6 Cl8 }4+ core structure. However, it cannot be oxidized within the anodic potential limit of this liquid and cannot be reduced prior to the electrodeposition of Al. Cyclic voltammograms recorded at Pt stationary and rotating disk electrodes in 66.7 mol% [EMIM]+ Cl− /AlCl3 liquid show that the addition of (Mo6 Cl8 )Cl4 results in slightly negative shifts in the potential of the electrodeposition process. Furthermore, the stripping wave of pure Al deposits is replaced by a new stripping wave at a more positive potential. These results indicate that Al–Mo alloys are formed through the following reduction process. x{Mo6 Cl8 }4+ + 8(3 − 2x) Al2 Cl7 − + 6(3 − x) e− = 6Al1−x Mox + 2(21 − 13x)AlCl4 −

(5.3)

Electrodeposits of Al–Mo alloys were prepared at Cu rotating disk electrode and rotating wire electrode substrates with galvanostatic electrolysis and examined with

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.2 Electrodeposition of Al-containing Alloys from Chloroaluminate Ionic Liquids 129

energy dispersive X-ray (EDX), SEM, and XRD for compositional and morphological analysis. As the concentration of {Mo6 Cl8 }4+ in the plating bath is much smaller than that of Al2 Cl7 − , the partial current density for {Mo6 Cl8 }4+ reduction is fixed and small, and increasing the total applied deposition current density simply leads to more Al deposition. As a result, the Mo content of the alloy decreases with increasing applied current density. Increasing the {Mo6 Cl8 }4+ concentration increases the partial current density for Mo deposition. Thus, at a fixed total applied current density, the Mo content increases with increasing {Mo6 Cl8 }4+ concentration. The partial current for Mo deposition relative to that for Al deposition can also be increased by increasing the deposition temperature, leading to higher Mo content in the alloy. SEM images of the Al–Mo electrodeposits reveal a surface morphology consisting mainly of spherical nodules and a nearly specular surface could be obtained. EDX maps for Al and Mo in this electrodeposit indicate that both elements are distributed more or less evenly over the surface of the deposit. XRD analysis of the Al–Mo alloy deposits shows that those containing less than 5 atom% Mo are single phase, supersaturated solid solutions having an fcc structure very similar to that of pure Al. Broad reflection indicative of an amorphous phase appears in deposits containing more than 6.5 atom% Mo. As the Mo content of the deposits is increased, the amount of fcc phase in the alloy decreases whereas that of the amorphous phase increases. When the Mo content is more than 10 atom%, the deposits are completely amorphous. As the Mo atom has a smaller lattice volume than Al, the lattice parameter for the deposits decreases with increasing Mo content. Potentiodynamic anodic polarization experiments in deaerated aqueous NaCl revealed that increasing the Mo content for the Al–Mo alloy increases the pitting potential. It appears that the Al–Mo deposits show better corrosion resistance than most other aluminum–transition metal alloys prepared from chloroaluminate ionic liquids. 5.2.3 Al–Zr

The electrodeposition of Al–Zr alloys was examined in the 66.7–33.3 mol% [EMIM]+ Cl− /AlCl3 liquid [19]. The reduction of Zr(IV), which was introduced as ZrCl4 in the liquid, produces a small ill-defined cathodic wave and a small negative shift to the Al deposition wave. Voltammetric data show that the small ill-defined cathodic wave corresponds to the Zr(IV)/Zr(III) reaction. It is noted that a surface passivating film is formed on the electrode surface after this reaction, indicating that the Zr(III) is insoluble in the liquid. Solutions of Zr(II) can be prepared by chemical reduction of Zr(IV) with Al or Zr metals. The in situ formtion of Zr(II) is, however, more efficient with Al metal in more acidic liquid. The maximum concentration of Zr(II) that could be produced by the Al reduction of Zr(IV) was about 0.02 mol l−1 in 66.7–33.3 mol% liquid. The measured diffusion coefficient for the Zr(II) species is much smaller than that for the Zr(IV) species and decreases as the Zr(II) concentration increases. This phenomenon suggests that polymerization of Zr(II) may possibly have occurred.

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

130 5 Electrodeposition of Alloys

The electrodeposition of Al–Zr alloys was investigated by using galvanostatic electrolysis using Cu rotating wire electrodes at 353 K in the 66.7–33.3 mol% liquid containing either Zr(IV) or Zr(II). As a limiting current is more rapidly obtained for the reduction of Zr(II) than for the reduction of Al2 Cl7 − , the partial current for the deposition of Zr reaches a constant value whereas that for the deposition of Al continuously increases with increasing current density. As a result, the Zr content of the electrodeposited alloys decreases with increasing current density. It was also noted that because the diffusion coefficient of Zr(II) is much smaller than that of Zr(IV), solutions of Zr(II) lead, in Al–Zr alloys, to smaller amounts of Zr- relative to Zr(IV)-containing solutions of equal concentration. Results from SEM and XRD examinations revealed that the structure of the Al–Zr deposits varies with the alloy composition. Al–Zr deposits containing less than 5 atom% Zr consist of nodules of fcc crystals similar to pure Al. The nodules decrease in size with increasing Zr composition and decreasing current density, suggesting that a certain grain refining is driven by the incorporation of Zr into the alloy rather than by the deposition overpotential. In addition to fcc Al, an amorphous phase becomes apparent when the Zr content for the Al–Zr alloy is further increased, and the alloy deposit containing 16.6 atom% Zr is completely amorphous. The corrosion resistance of the electrodeposited Al–Zr alloy was examined by pitting potential measurements. It was found that the addition of 8 atom% or more Zr increases the pitting potential of the alloy by about +0.3 V vs. pure Al. 5.2.4 Al–Pt

The electrodeposition of Al–Pt, which is an interesting material for catalysis, has been studied in Lewis acidic ionic liquids formed from AlCl3 with benzyltrimethyl ammonium chloride ([BTMA]+ Cl− /AlCl3 ) [20]. This ionic liquid is slightly less water sensitive than [EMIM]+ Cl− /AlCl3 . Cyclic voltammograms recorded at an Fe electrode in a 1:2 [BTMA]+ Cl− /AlCl3 liquid containing BTMA2 PtCl6 showed that the reduction of Pt(II) is slightly less negative than the reduction of Al(III). This is surprising at first glance but shows to what amount complexation can alter electrode potentials. Constant potential electrolysis was employed to prepare Al–Pt deposits in liquids containing different platinum complex ions such as tetraaminoplatinum(II) and bis(acetylacetonato)platinum(IV). At a more negative potential (−1 V vs. an Al quasi-reference electrode immersed in the same liquid) where bulk deposition of Al would occur, poorly adherent powders were obtained which contained primarily Al and some trapped chloride. At a less negative applied potential (−0.6 V) where the deposition of Al may still be in the kinetic controlled region, bright, adherent Al–Pt alloy deposits with dense nodules could be obtained with negligible amounts of chloride. The Pt content for the Al–Pt deposits increased (from 5% to 13% by weight) with increasing Pt(II) (or Pt(IV)) concentration, which depended on the solubility of the Pt complex used in the solution. EDX analysis suggested that the Al–Pt deposits were homogeneous. At even less positive deposition potential (−0.2 V) bright, adherent pure Pt could be obtained although the deposition rate was very slow. Information from XRD analysis was, however, not provided in this study.

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.2 Electrodeposition of Al-containing Alloys from Chloroaluminate Ionic Liquids 131

5.2.5 Al–Mg

Al-Mg alloys are widely used for chemical-processing and food-handling equipment. While electrodeposition may be an effective and cost efficient option for the preparation of thin alloy coatings, the fact that the standard potential of the Mg(II)/Mg couple is much more negative than that of the Al(III)/Al makes the electrodeposition of Al–Mg alloys, at first glance, implausible. Nevertheless, induced codeposition of Al–Mg alloys from Lewis acidic [EMIM]+ Cl− /AlCl3 (mole ratio 1:2) ionic liquid at 30 ◦ C has been examined by Morimitsu et al. [21]. Mg(II) was introduced to the liquid by dissolution of 0.2 mol kg−1 MgCl2 . Cyclic voltammograms were recorded at a tungsten working electrode vs. an Al(III)/Al reference electrode. They show that the deposition of pure Al at a W substrate requires a large nucleation overpotential. Addition of MgCl2 reduces the overpotential and shifts the deposition process to less negative electrode potentials. Multiple stripping waves appear in the presence of MgCl2 and the relative magnitude of these stripping waves depends on the reversing potential, indicating the codeposition of Mg with Al. The Mg atomic content of the deposits was found to increase with increasing current density (or cathodic overpotential), supporting the fact that deposition of Mg starts at a potential more negative than that of pure Al. The alloy deposits obtained in this study were single phase Al–Mg solid solutions. As the Mg atomic content was very low, 2.2 atom%, the XRD patterns of the alloy deposits were almost identical to that of pure Al. The codeposition of Mg with Al in this study was classified as “induced codeposition” of which the alloy deposition occurs at more positive potentials than the deposition of the more noble metal (Al in this case) of the alloy components [43]. 5.2.6 Al–Mo–Mn

Since both Al–Mo [18] and Al–Mn [13] alloys can be electrodeposited from Lewis acidic [EMIM]+ Cl− /AlCl3 liquid, Tsuda et al. investigated the electrodeposition of the Al–Mo–Mn ternary alloys in the 33.3–66.7% mole ratio [EMIM]+ Cl− /AlCl3 liquid containing Mo(II) as (Mo6 Cl8 )Cl4 and Mn(II) as MnCl2 at 55 ◦ C [22]. Cyclic voltammograms recorded at a Pt disk electrode for the solutions revealed that the Al–Mo–Mn electrodeposition process varies with the concentration ratio CMn(II) /CMo(II) and that the presence of Mn(II) in the solution inhibits the nucleation of Al. Controlled current techniques were employed to prepare Al–Mo–Mn alloy samples at a copper rotating wire electrode. The relationships between the total applied current density, jt , and the partial current densities of Mo, Mn, and Al were expressed as jMo , jMn , and jAl . Plots of −jMo , −jMn , and −jAl vs. −jt show that −jAl varies linearly with jt . It appears that the deposition of Mo is inhibited by Mn(II) so that −jMo reaches a limiting value and becomes a smaller fraction of the total current density as –jt is increased. Because of this, the Mo content of the Al–Mo–Mn alloy decreases with increasing −jt . On the other hand, −jMn increases with −jt so that the Mn content of the alloys increases with increasing −jt . Overall,

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

132 5 Electrodeposition of Alloys

Al–Mo–Mn alloys rich in Mo are obtained at small −jt , whereas alloys rich in Mn are obtained at large −jt , provided that CMn(II) /CMo(II)  2. The inhibition of the deposition of the more noble component, Mo, by the less noble component, Mn, makes the codeposition of Al, Mo, and Mn an anomalous process. SEM images of the Al–Mo–Mn alloy deposits reveal that the morphology varies from spherical nodules to a shining surface, depending on the current density and the Mo and Mn concentrations. XRD analysis revealed that the deposits containing less than approximately 10 atom% Mo + Mn exhibited a face-centered cubic Al and an amorphous phase. When the concentration of Mo + Mn exceeded 10 atom%, only an amorphous phase was observed. Pitting potential measurements of the electrodeposited Al–Mo–Mn alloys revealed that the addition of relatively modest amounts of Mo and Mn to the alloy resulted in a significant increase in corrosion resistance compared to pure Al and the comparable binary alloy containing only one of the transition metal components. 5.2.7 Al–Cr–Ni

Alloys with Al, Cr and Ni are of technical importance due to their excellent temperature and corrosion resistance. The electrodeposition of an Al–Cr–Ni layer was attempted in 1:2 [EMIM]+ Cl− /AlCl3 liquid containing 5 × 10−2 mol l−1 NiCl2 and 6 × 10−2 mol l−1 CrCl2 at 338 K [23]. Cyclic voltammetric experiments showed that the reduction of Ni(II) occurs at a potential more positive than for Cr(II) and Al(III) whereas the reduction potential of Cr(II) almost overlapped with that of Al(III). Al–Cr–Ni alloy samples were prepared by constant potential electrolysis at glassy carbon substrates and their compositions were analyzed by fluorescence X-ray spectroscopy. The results showed that the atomic ratio of Al:Cr:Ni was 97:2:1 at a potential where bulk deposition of Al occurred and 90:1:9 at a less negative potential. The low atomic ratio of Cr and Ni in the deposits is partly due to the low concentration of Ni(II) and Cr(II) in comparison to that of Al(III). To increase the Ni and Cr content in the deposit, pulse potential electrolysis was adopted. In a typical pulse electrolysis cycle, the potential was first stepped to a sufficiently negative value for the codeposition of Al, Cr, and Ni, and then the potential was applied to a more positive value for the dissolution of Al. By changing pulse conditions, including negative and positive potentials and frequency, the concentration of Ni and Cr in the deposits was enhanced to 20–27 and 15 atom%, respectively.

5.3 Electrodeposition of Zn-containing Alloys from Chlorozincate Ionic Liquids

Ionic liquids can be obtained by the combination of zinc halide with certain organic halides (see Chapter 3.3). As the cathodic potential limit of Lewis acidic liquids (with more than 33 mol% zinc halide) is due to the deposition of metallic zinc, the

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.3 Electrodeposition of Zn-containing Alloys from Chlorozincate Ionic Liquids 133

electrodeposition of zinc and its alloys from these ionic liquids is feasible. Some of the studies are described below. 5.3.1 Alloys of Zn with Cu, Cd and Sn

The electrodeposition of Zn–Cu, Zn–Sn, and Zn–Cd has been investigated in Lewis acidic [EMIM]+ Cl− /ZnCl2 liquid containing Cu(II) [24], Cd(II) [25] and Sn(II) [26], respectively. Figure 5.1 illustrates the cyclic voltammograms of the 50.0–50.0 mol% [EMIM]+ Cl− /ZnCl2 with and without Cu(I). A typical alloy formation was observed. The deposition of Cu, Sn, and Cd occurs at a potential of 0.5, 0.3 and 0.1 V, respectively, more positive than the deposition of Zn. In these studies samples of the alloys were prepared on Ni substrates by constant potential electrolysis and examined with EDX, SEM, and XRD. It was found that the Zn content in the electrodeposits increased as the deposition potential became more negative but decreased with increasing concentrations of Cu(II), Sn(II), and Cd(II) in the solution. Increasing the deposition temperature increases the mass-transport rates

Fig. 5.1 Staircase cyclic voltammograms for the 50.0–50.0 mol% [EMIM]+ Cl− /ZnCl2 liquid on tungsten and nickel electrodes at 80 ◦ C, (a) and (b) with 200 mM Cu(I), (c) without Cu(I). Scan rate 50 mV s−1 . [Ref. 24].

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

134 5 Electrodeposition of Alloys

of the metal ions in the plating bath and decreases the overpotential required for the deposition. As a result, increasing the temperature increases the content of the more noble metal (Cu, Cd, and Sn) in the deposit.

5.3.2 Zn–Co

The electrodeposition of Zn–Co and Zn–Fe alloys in an aqueous bath is classified as an anomalous codeposition [44] because the less noble Zn is preferentially deposited with respect to the more noble metal. This anomaly was attributed to the formation of Zn(OH)+ which adsorbs preferentially on the electrode surface and inhibits the effective deposition of the more noble metal. This anomaly was circumvented by using zinc chloride-n-butylpyridinium chloride ([BP]+ Cl− / ZnCl2 ) [27] or [EMIM]+ Cl− /ZnCl2 [28] ionic liquids containing Co(II). The Zn–Co deposits can be varied from Co-rich to Zn-rich by decreasing the deposition potential or increasing the deposition current. XRD measurement reveals the presence of Co5 Zn21 in the deposited Zn–Co alloys and that the Co-rich alloys are amorphous and the crystalline nature of the electrodeposits increases as the Zn content of the alloys increases. Addition of propylene carbonate cosolvent to the ionic liquid decreases the melting temperature of the solution and allows the electrodeposition to be performed at a lower temperature. The presence of CoZn alloy is evidenced by the XRD patterns shown in Figure 5.2.

Fig. 5.2 (A) XRD patterns of electrodeposits produced from a 60.0–40.0 mol% [EMIM]+ Cl− /ZnCl2 liquid containing 1.16 wt% of CoCl2 at 80 ◦ C at deposition potential of (b) 0.13, (c) −0.17, (d) −0.22, and (e) −0.26 V. For comparison, the XRD pattern of a pure zinc deposit is given in

(a). (B) XRD patterns of electrodeposits produced from 6.0 g of 60.0–40.0 mol% [EMIM]+ Cl− /ZnCl2 liquid containing 1.16 wt% of CoCl2 and 6.0 g of propylene carbonate at 40 ◦ C at a deposition potential of (a) 0.20 and (b) −0.27 V. [Ref. 28].

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.3 Electrodeposition of Zn-containing Alloys from Chlorozincate Ionic Liquids 135

5.3.3 Zn–Fe

Similar to the electrodeposition of Zn–Co, the electrodeposition of corrosion resistant Zn–Fe alloy in an aqueous bath is an anomalous codeposition and the Zn/Fe ratio in the deposit is higher than that in the electrolyte. However, nonanomalous deposition of Zn–Fe was achieved by conducting the deposition in a 60.0–40.0 mol% [EMIM]+ Cl− /ZnCl2 ionic liquid containing Fe(II) [29]. Cyclic voltammograms showed that the deposition of Fe occurs at a potential less negative than that of Zn. Underpotential deposition of Zn on Fe occurred through an instantaneous two-dimensional nucleation process observed prior to the deposition of bulk Zn. It is interesting to note that, as shown in Figure 5.3, the Zn adatoms from UPD were able to diffuse into the bulk iron to form Zn–Fe alloy. Zn–Fe alloy could also be prepared at potentials where bulk deposition of Zn occurred. The Fe content in the deposit can be varied from 100 to 50 atom% by decreasing the deposition potential or the Fe(II) concentration in the solution. SEM images of the deposits revealed that they were dense and compact, and the morphology varied from nodules to pyramidal and hexagonal as the iron content in the deposits decreased.

Fig. 5.3 SEM analysis of Zn electrodeposits obtained at −0.12 V where only UPD of Zn on Fe substrate occurs from pure 60.0–40.0 mol% [EMIM]+ Cl− /ZnCl2 ionic liquid, on to a 0.25 cm2 iron foil: (a) sec-

ondary electron image of a polished crosssection; (b) EDS line scan of the polished cross-section(scanned along the white line in the micrograph). [Ref. 29].

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

136 5 Electrodeposition of Alloys

5.3.4 Zn–Ni

The electrodeposition of Zn–Ni alloy from the [EMIM]+ Cl− /ZnCl2 /NiCl2 ionic liquid was examined by Koura et al. [30]. The Ni content in the deposit decreases from 98.6 to 12.3 mol% with increasing deposition current density. Due to the high viscosity of the [EMIM]+ Cl− /ZnCl2 /NiCl2 liquid, the current efficiency was low, even when the temperature was increased to 100 ◦ C. XRD analysis of the deposits revealed both amorphous and crystalline ZnNi, but the formation of the Zn21 Ni5 compound was not observed. In order to reduce the viscosity and to enhance the current efficiency, ethanol (EtOH) was added to the [EMIM]+ Cl− /ZnCl2 /NiCl2 liquid at 40 ◦ C. The current efficiency was improved to almost 100% in all the electrodepositions performed in the [EMIM]+ Cl− /ZnCl2 /NiCl2 /EtOH solution. The XRD patterns of the deposits obtained from the [EMIM]+ Cl− /ZnCl2 /NiCl2 /EtOH showed both crystalline Zn21 Ni5 and amorphous ZnNi. Differential scanning calorimetry (DSC) of the deposit showed exothermic peaks that were attributed to the amorphous-to-crystalline transformation together with crystal growth. 5.3.5 Zn–Mg

The electrodeposition of Zn–Mg alloy was examined in mixtures of 1-ethyl-3methylimidazolium bromide ([EMIM]+ Br− )/ZnBr2 /MgBr2 /ethylene glycol (EG) at 120 ◦ C.[31] The total concentration of ZnBr and MgBr in the bath was kept at 10 mol% while the ZnBr2 /MgBr2 mole ratio was varied. Linear scan voltammetry revealed a single reduction wave resulting from the deposition of Zn in the [EMIM]+ Br− /ZnBr2 /EG solution. The addition of MgBr2 shifted the reduction wave to more negative electrode potentials due to the codeposition of Mg. Dense Zn–Mg alloys could be electrodeposited by potentiostatic electrolysis at a Cu substrate. The XRD diffraction analysis showed that the alloys contained Zn11 Mg2 in addition to Zn, and CuZn5 . EDX analysis showed that the Zn and Mg were distributed uniformly in the alloy. The composition of the alloy could be controlled by the ionic liquid bath composition. A steel sample that was coated with Zn–Mg alloy containing 2.5 mol% Mg showed significantly improved corrosion resistance. 5.3.6 Pt–Zn

The electrodeposition of Pt–Zn from a 60.0–40.0 mol% [EMIM]+ Cl− /ZnCl2 liquid containing Pt(II) was investigated at 90 ◦ C [32]. Cyclic voltammograms showed that the reduction of Pt(II) to Pt occurs at a potential slightly less negative than the reduction of Zn(II). Multiple anodic stripping waves were observed for the Pt–Zn electrodeposits, indicative of a multiphasic structure of the deposits. The Zn component of the deposits was stripped at a less positive potential than the Pt component of the deposits. Samples of Pt–Zn deposits containing 8 to 42 atom%

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.4 Fabrication of a Porous Metal Surface by Electrochemical Alloying and De-alloying 137

Fig. 5.4 The XRD pattern of the Pt–Zn coating (Pt a/o = 40.91%) that was electrodeposited on a tungsten foil at a deposition potential of −0.2 V in the 60.0–40.0 mol% [EMIM]+ Cl− /ZnCl2 ionic liquid containing 120 mM PtCl2 at 90 ◦ C. [Ref. 32].

Pt were prepared on tungsten by constant potential electrolysis. EDX analysis of the deposits indicated that Pt and Zn were distributed uniformly in the deposits. The Pt content in the deposit decreases as the deposition potential approaches the value where bulk deposition of Zn occurs. Increasing the Pt(II) concentration in the liquid increases the Pt content in the deposits. As shown in Figure 5.4, XRD results indicated the presence of crystalline Zn and amorphous PtZn. If Zn is electrodeposited on a Pt substrate, the deposited Zn atoms interact with the Pt to form Pt–Zn surface alloys.

5.4 Fabrication of a Porous Metal Surface by Electrochemical Alloying and De-alloying

Porous metals are of interest due to their potential applications in catalysis, fuel cells, chemical sensors and so on. The fact that the electrodeposition of Zn on certain metals, M, can lead to surface alloys Mx Zn1–x and that the Zn in the alloy can be subsequently removed by anodic stripping makes it possible to prepare porous metal surfaces by an electrochemical alloying/de-alloying process. Some examples including Pt, Au, and Ag have been demonstrated [33–35]. The formation of a porous metal surface during electrochemical de-alloying can be accounted for with the model described by Erlebacher [45]. For example, Figure 5.5 illustrates the dealloying of a Ag–Zn surface alloy. The process starts with selective dissolution of the zinc atoms from the outermost Ag–Zn alloy surface, leaving behind the more noble Ag atoms, which agglomerate to islands, leading to the formation of tiny pits. The more pits formed, the more the original alloy is exposed to the electrolyte. The selective dissolution of zinc atoms from the newly exposed Ag–Zn releases more Ag atoms to the surface. These atoms diffuse to the Ag clusters left over from dissolution of previous layers, continuing to leave more Ag–Zn exposed to

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

138 5 Electrodeposition of Alloys

Fig. 5.5 Plan-view SEM images of silver wire samples that have been electrodeposited with 12.73 C cm−2 of zinc followed by de-alloying at 0.6 V. The amounts of the zinc that was de-alloyed are: (a) 1.59, (b) 4.77, (c) 9.55, and (d) 12.73 C cm−2 . The temperature was 150 ◦ C. [Ref. 35].

electrolyte and resulting in increased pore size. Such selective dissolution of zinc (roughening) and surface diffusion of Ag (agglomeration or smoothing) continues as the de-alloying proceeds and an interconnected porous structure is formed. The structure and morphology of the porous metal surface are affected by electrochemical variation of the composition and the thickness of the M–Zn surface alloys (M = Pt, Au, or Ag). Higher deposition temperature favors the effective formation of the M–Zn alloy and increases the thickness of the alloy. Higher de-alloying temperature enhances the surface diffusion of the metals and results in larger pores. As both the deposition and de-alloying steps are performed in a single bath of [EMIM]+ Cl− /ZnCl2 ionic liquid, the Zn(II) species consumed in the deposition step returns to the ionic liquid during the de-alloying step, the composition of the ionic liquid is essentially unchanged and thus can be re-used. The fabricated nanostructured platinum electrode was tested for the electro-oxidation of methanol. A much higher current density is observed on the nanostructured platinum electrode than on the polished platinum electrode, indicating that the former has a much higher surface area. The prospective application of the fabricated porous Ag electrode was tested for the electrochemical reduction of chloroform. The advantageous catalytic effect of the Ag electrode over a glassy carbon electrode is illustrated in Figure 5.6 [35] which shows that while no significant current due to chloroform could be observed

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.5 Nb–Sn 139

Fig. 5.6 Typical cyclic voltammograms showing the electroreduction of 50 mM chloroform in acetonitrile containing 1 M H2 O and 0.1 M tetraethyl ammonium perchlorate (TEAP) as the supporting electrolyte recorded at (a) a bare glassy carbon, (b) a polished Ag, and (c) a porous Ag electrode. [Ref. 35].

at the glassy carbon electrode, appreciable current was observed at a polished Ag electrode at about −1.3 V. The current density was further enhanced at the porous Ag electrode compared with that on the polished Ag electrode. It has been demonstrated that a nanoporous Au surface can be prepared by electrochemical alloying/ dealloying from the [EMIM]+ Cl− /ZnCl2 liquid. It should be mentioned here that according to IUPAC there are no nanoporous structures: 50 nm: macroporous. However, even in the peer-reviewed literature the expression “nanoporous” is increasingly employed for materials with pores in the nanometer regime. Porous Au can for example be successfully functionalized with self-assembled monolayers of L-cysteine. Such functionalization greatly improves the utility of the nanoporous gold, as was demonstrated in the sensitive and selective determination of Cu(II) [34].

5.5 Nb–Sn

Koura et al. investigated the electrodeposition of Nb–Sn alloy in the ionic liquid formed from a mixture of [EMIM]+ Cl− /SnCl2 /NbCl5 [36]. For the [EMIM]+ Cl− / SnCl2 liquid, only one redox couple due to the cathodic deposition and anodic stripping of Sn was observed in the cyclic voltammogram. When NbCl5 was introduced into the [EMIM]+ Cl− /SnCl2 liquid, new waves appeared in the voltammogram, suggesting the codeposition of Nb–Sn alloy. The potential of these waves shifted negatively on increasing the mole fraction of EMIC in the liquid, indicating that the

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

140 5 Electrodeposition of Alloys

Sn(II) and Nb(V) species changed their coordinations with the liquid composition. Nb–Sn alloy samples were prepared by the potentiostatic method and analyzed. The results showed that the Nb content in the alloy could be increased by increasing the bath temperature to 160 ◦ C and increasing the NbCl5 content in the bath. However, increasing the NbCl5 mole fraction in the bath also increased the viscosity of the bath. Pulse electrolysis was found to be effective in increasing the Nb content in the alloy. The maximum Nb content in the alloy was 60.8 wt% from constant potential electrolysis and 69.1 wt% from pulse electrolysis. XRD diffraction patterns showed that the electrodeposits contained crystalline Sn and Nb3 Sn which is a superconductor material.

5.6 Air- and Water-stable Ionic Liquids

For the electrodeposition of metals or alloys from air- and water-stable ionic liquids, it is necessary first to dissolve the corresponding metal ions in the ionic liquid. Such a dissolution process is made possible by introducing excess amounts of halide ions (such as Cl− ) to form soluble metal-halide complex anions. Alternatively, the metal is electrochemically oxidized in the ionic liquid to form the soluble salt such as Sn(Tf2 N) in the trimethyl-n-hexylammonium [bis(trifluoromethyl)sulfonyl]amide ([TMHA]+ Tf2 N− ) ionic liquid. 5.6.1 Pd–Au, Pd–Ag, Pd–In

The electrodeposition of Pd–Au [37] and Pd–Ag [38] was investigated in the temperature range from 30 to 120 ◦ C in the air- and water-stable ionic liquid 1-ethyl3-methylimidazolium chloride-tetrafluoroborate ([EMIM]+ BF4 − ) containing Pd(II) and Au(I), or Pd(II) and Ag(I), as well as excess chloride ions. Cyclic voltammograms indicated that the reduction of Au(I) and Ag(I) occurs at potentials less negative than that of the Pd(II). Pd–Au and Pd–Ag alloys could be prepared by galvanostatic or potentiostatic deposition on nickel substrates. EDX analysis of the deposited alloys indicated that the Pd content of the alloy increased with increasing Pd concentration in the bath and with lowering the deposition potential (or increasing the current density). XRD measurements indicated that the alloys were solid solutions of Pd–Au and Pd–Ag. SEM images showed that increasing the deposition temperature made the deposited alloys more compact. The electrodeposition of Pd–In was investigated in the [EMIM]+ BF4 − ionic liquid containing Pd(II) and In(III), and excess chloride ions [39]. The cyclic voltammograms shown in Figure 5.7 indicate that the reduction of Pd(II) occurs at a potential less negative than that for the bulk deposition of In. However, UPD of In on Pd occurs at the same potential as the deposition of Pd. Pd–In alloys can be prepared within the indium UPD regime or at more negative potentials where overpotential deposition of In occurs. As Figure 5.8 shows, in the UPD regime the In content

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.6 Air- and Water-stable Ionic Liquids 141

Fig. 5.7 Cyclic voltammograms of (a) 10 mM Pd(II), (b) 20 mM In(III) and (c) 10 mM Pd(II) + 20 mM In(III) in a [EMIM]+ Cl− /BF4 − ionic liquid at a GC electrode at 120 ◦ C. Scan rate = 100 mV s−1 . [Ref. 39].

Fig. 5.8 Variation of the Pd–In electrodeposit composition with deposition potential. The deposits were prepared in a [EMIM]+ Cl− /BF4 − ionic liquid at 120 ◦ C containing: () 10 mM Pd(II) and 10 mM In(III); () 10 mM Pd(II) and 40 mM In(III); () 40 mM Pd(II) and 10 mM In(III). [Ref. 39].

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

142 5 Electrodeposition of Alloys

of the alloys was less dependent on the In(III) concentration of the bath because the UPD of In on Pd is a slow process. In the mass-transport-limited potential regime the alloy composition corresponds to the Pd(II)/In(III) composition in the plating bath. SEM images showed that the alloys prepared within the UPD regime appeared to have more compact and smooth morphologies. 5.6.2 In–Sn

The electrodeposition of In–Sn alloys from the Lewis basic [EMIM]+ BF4 − ionic liquid containing 0.1 mol kg−1 InCl3 and 0.1 mol kg−1 SnCl2 was investigated by Morimitsu et al. using cyclic voltammetry and potentiostatic electrolysis [40]. The cyclic voltammograms indicated that the reduction of Sn(II) occurred at a potential less negative than the reduction of In(III). Furthermore, the deposition of Sn greatly reduced the overpotential required for the deposition of In, making the codeposition of InSn more feasible. The formation of at least two different phases during the deposition was indicated by the presence of multiple anodic stripping waves. Indium–tin alloy samples were prepared at a Pt flag electrode by constant potential electrolysis. The obtained samples were analyzed by inductively coupled plasma (ICP) atomic emission spectrometry and XRD measurements. The results showed that the indium content increased to 29 atom% as the applied potential became more negative. The InSn alloys could not only be obtained by the bulk deposition of both In and Sn but could also be prepared by UPD of indium on the predeposited tin. XRD measurements showed that the electrodeposited alloys were mixtures of Sn and InSn4 . The crystallinity of the InSn4 phase in the electrodeposits is significantly affected by electrolysis temperature. Characteristic diffraction patterns of InSn4 were not observed at room temperature but became evident when the alloy samples were electrodeposited at 80 ◦ C. The fact that the maximum In content in the deposits did not exceed 29 atom% indicates that the In–Sn alloy deposition is not simply controlled by the mass-transport of individual Sn(II) and In(III) species. Otherwise, the In content would be close to 50 atom% considering that the In(III) and Sn(II) concentration ratio is 1:1 in the plating solution. 5.6.3 Cu–Sn

The formation of Cu–Sn alloy by galvanic contact deposition in the trimethyl-nhexylammonium [bis(trifluoromethyl)sulfonyl]amide ([TMHA]+ Tf2 N− ) ionic liquid at a temperature above 100 ◦ C has been demonstrated by Katase et al. [41] Sn(II) was introduced into the liquid by dissolution of the Sn(Tf2 N) salt which has a solubility of 0.2 mol dm−3 . In the plating cell, a copper sheet was used as the cathodic substrate, a Sn sheet was used as the anode, and a Sn rod immersed in the same solution was used as a quasi-reference electrode. On short-circuiting, the Sn anode was oxidized to Sn(II) giving two electrons through external circuit to

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

5.6 Air- and Water-stable Ionic Liquids 143

the Cu cathode where Sn(II) was cathodically deposited. During deposition, the current density decreased gradually to a steady-state current due to the depletion of Sn(II) concentration in the vicinity of the cathode. The cathode potential remained positive against Sn(II)/Sn, ensuring that the activity of the deposited Sn atoms is lower than unity. After the conclusion of the deposition, silver-gray pinholes and crack-free coatings were obtained. The XRD pattern of the Cu–Sn deposits obtained at 120, 130, and 140 ◦ C showed that the deposits are composed of crystalline Cu6 Sn5 , Cu3 Sn, Cu10 Sn3 intermetallic phases. The amounts of Cu-rich phases and the alloy thickness increased with increasing temperature due to the enhanced diffusion of deposited Sn atoms into the Cu substrate. The temperature dependence of the thickness (x) of the deposits obtained by deposition for 72 h was studied and Arrhenius behavior was observed in the plot of log x vs. T −1 . From the slope of this plot, the apparent activation energy of the growth was estimated to be 58 kJ mol−1 .

5.6.4 Zn–Mn

The electrodeposition of Zn–Mn was investigated at 80 ◦ C in the hydrophobic tri-1butylmethylammonium bis((trifluoromethyl)sulfonyl)amide ([TBMA]+ Tf2 N− ) [46] ionic liquid containing Zn(II) and Mn(II) species that were introduced into the ionic liquid by anodic dissolution of the respective metal electrodes. Cyclic voltammograms indicated that the reduction of Zn(II) occurs at a potential less negative than that of the Mn(II). Due to some kinetic limitations, which is a common phenomenon in air- and water-stable ionic liquids, incomplete oxidation of Mn electrodeposits was observed in this system. The current efficiency of Mn electrodeposition in this ionic liquid approaches 100%, which is a great improvement compared to the results obtained in aqueous solution (20–70%). Electrodeposition of Zn–Mn alloy coatings has never been carried out in chloroaluminate ionic liquid because of the unavoidable codeposition of Mn and Al. Coatings containing Zn, Mn or Zn–Mn were obtained by controlled-potential electrolysis and analyzed by SEM, EDX and XRD. It is very interesting that the reduction wave of Zn(II) disappeared when the ionic liquid contained both Zn(II) and Mn(II) species; this is illustrated by the CVs shown in Figure 5.9. The reason is still not clear but compact and adherent Zn–Mn alloy deposits of various compositions can be obtained and the Mn/Zn ratio of these alloys depended almost completely on the Mn(II)/Zn(II) concentration ratio in the ionic liquid. The SEM images shown in Figure 5.10 demonstrate that the Zn–Mn alloy deposits were very smooth and the grain size increased with increasing concentration of Mn. The XRD results indicate that the Zn–Mn alloy deposits obtained from the [TBMA]+ Tf2 N− were metallic glasses or amorphous. Potentiodynamic anodic polarization experiments in deaerated aqueous NaCl revealed that the addition of Mn up to 50 atom% improves the corrosion resistance of Zn. However, the addition of Mn beyond this amount decreases the corrosion resistance of the Zn–Mn alloy.

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

144 5 Electrodeposition of Alloys

Fig. 5.9 Staircase cyclic voltammograms of (—) 0.3 M Zn(II), (. . .) 0.2 M Mn(II) and (—) a mixture of 0.16 M Zn(II) + 0.1 M Mn(II) recorded at a W electrode in [TBMA]+ Tf2 N− ionic liquid. Temperature, 80 ◦ C. Scan rate, 50 mV s−1 . [Ref. 46].

Fig. 5.10 SEM micrographs of pure Mn, pure Zn and various compositions of Zn–Mn alloys. The composition of the alloy coatings are shown as atomic ratios on each plot. The magnification of each micrograph is 5000×. [Ref. 46].

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

References

5.7 Summary

In this chapter some results on the electrodeposition of alloys from ionic liquids are summarized. Many fundamental studies have been performed in chloroaluminate first generation ionic liquids but the number of studies employing air- and waterstable ionic liquids rather than the chloroaluminates is increasing. Currently, new ionic liquids with better electrochemical properties are being developed. For example, Abbott et al. [47] have prepared a series of ionic liquids by mixing commercially available low-cost choline chloride and MCl2 (M = Zn, Sn) or urea and demonstrated that these ILs are good media for electrodeposition for pure metals (see Chapter 4.3). It can be expected that in the near future, the electrodeposition of alloys from ILs may become available for industrial applications. Furthermore, due to their variety, their wide electrochemical and thermal windows air- and water-stable ionic liquids have unprecedented prospects for electrodeposition.

References 1 Stafford, G.R. and Hussey, C.L. (2002) Electrodeposition of transition metal– aluminum alloys from chloroaluminate molten salts, in Advances in Electrochemical Science and Engineering, Vol. 7 (eds R.C. Alkire and D.M. Kolb), Wiley-VCH, Verlag GmbH. 2 Endres, F. (2002) Chem. Phys. Chem., 3, 144. 3 Katayama, Y. (2005) Electrodeposition of metals in ionic liquids, in Electrochemical Aspects of Ionic Liquids (ed. H. Ohno), Ch. 9, Wiley & Sons, Inc., New York. 4 El Abedin, S.Z. and Endres, F. (2006) Chem. Phys. Chem., 7, 58. 5 Koura, N., Kato, T., and Yumoto, E. (1994) J. Surf. Finish. Soc. Jpn., 45, 805. 6 Pitner, W.R., Hussey, C.L., and Stafford, G.R. (1996) J. Electrochem. Soc., 143, 130. 7 Carlin, R.T., Trulove, P.C., and De Long, H.C. (1996) J. Electrochem. Soc., 143, 2747. 8 Mitchell, J.A., Pitner, W.R., Hussey, C.L., and Stafford, G.R. (1996) J. Electrochem. Soc., 143, 3448. 9 Ali, M.R., Nishikata, A., and Tsurs, T. (1997) Electrochim. Acta, 42, 1819. 10 Carlin, R.T., De Long, H.C., Fuller, J., and Trulove, P.C. (1998) J. Electrochem. Soc., 145, 1598. 11 Ali, M.R., Nishikata, A., and Tsuru, T. (1997) Electrochim. Acta, 42, 2347.

12 Tierney, B.J., Pitner, W.R., Mitchell, J.A., Hussey, C.L., and Stafford, G.R. (1998) J. Electrochem. Soc., 145, 3110. 13 De long, H.C., Mitchell, J.A., and Trulove, P.C. (1998) High Temp. Mater. Proc., 2, 507. 14 Tsuda, T., Nohira, T., and Ito, Y. (2001) Electrochim. Acta, 46, 1891. 15 Zhu, Q., Hussey, C.L., and Stafford, G.R. (2001) J. Electrochem. Soc., 148, C88. 16 Tsuda, T., Hussey, C.L., Stafford, G.R., and Bonevich, J.E. (2003) J. Electrochem. Soc., 150, C234. 17 Aravinda, C.L., Mukhopadhyay, I., and Freyland, W. (2004) Phys. Chem. Chem. Phys., 6, 5225. 18 Tsuda, T., Hussey, C.L., and Stafford, G.R. (2004) J. Electrochem. Soc., 151, C379. 19 Tsuda, T., Hussey, C.L., Stafford, G.R., and Kongstein, O. (2004) J. Electrochem. Soc., 151, C447. 20 Abbott, A.P., Eardley, C.A., Farley, N.R.S., Griffith, G.A., and Pratt, A. (2001) J. Appl. Electrochem., 31, 1345. 21 Morimitsu, M., Tanaka, N., and Matsunaga, M. (2000) Chem. Lett., 1028. 22 Tsuda, T., Hussey, C.L., and Stafford, G.R. (2005) J. Electrochem. Soc., 152, C620. 23 Ueda, M., Ebe, H., and Ohtsuka, T. (2005) Electrochemistry, 73, 739.

145

c05 (JWBG008-Endres)

December 25, 2007

14:10

Char Count=

146 5 Electrodeposition of Alloys

24 Chen, P.-Y., Lin, M.-C., and Sun, I-W. (2000) J. Electrochem. Soc., 147, 3350. 25 Huang, J.-F. and Sun, I.-W. (2002) J. Electrochem. Soc., 149, E348. 26 Huang, J.-F. and Sun, I-W. (2003) J. Electrochem. Soc., 150, E299. 27 Koura, N., Endo, T., and Idemoto, Y. (1999) J. Non-Cryst. Solids, 205, 650. 28 Chen, P.-Y. and Sun, I-W. (2001) Electrochim. Acta, 46, 1169. 29 Huang, J.-F. and Sun, I-W. (2004) J. Electrochem. Soc., 151, C8. 30 Koura, N., Suzuki, Y., Idemoto, Y., Kato, T., and Matsumoto, F. (2003) Surf. Coat. Technol., 120, 169. 31 Iwagishi, T., Sawada, K., Yamamoto, H., Koyama, K., and Shirai, H. (2003) Electrochemistry, 71, 318. 32 Huang, J.-F. and Sun, I-W. (2004) Electrochim. Acta, 49, 3251. 33 Huang, J.-F. and Sun, I-W. (2004) Chem. Mater., 16, 1829. 34 Huang, J.-F. and Sun, I.-W. (2005) Adv. Funct. Mater., 15, 989. 35 Yeh, F.-H., Tai, C.-C., Huang, J.-F., and Sun, I.-W. (2006) J. Phys. Chem. B, 110, 5215. 36 Koura, N., Umebayashi, T., Idemoto, Y.,

37 38 39 40

41

42 43 44 45 46 47

and Ling, G. (1999) Electrochemistry, 67, 684. Su, F.-Y., Huang, J.-F., and Sun, I.-W. (2004) J. Electrochem. Soc., 151, C811. Tai, C.-C., Su, F.-Y., and Sun, I.-W. (2005) Electrochim. Acta, 50, 5504. Hsiu, S.-I., Tai, C.-C., and Sun, I.-W. (2006) Electrochim. Acta, 51, 2607. Morimitsu, M., Nakahara, Y., and Matsunaga, M. (2005) Electrochemistry, 73, 754. Katase, T., Kurosaki, R., Murase, K., Hirato, T., and Awakura, Y. (2006) Electrochem. Solid-State Lett., 9, C69. Barnard, P.A., Sun, I.-W., and Hussey, C.L. (1990) Inorg. Chem., 29, 3670. Brenner, A. (ed.) (1963) Electrodeposition of Alloys, Vol. 1, Academic Press, New York. Brenner, A. (1963) Electrodeposition of Alloys, Vol. 1, Academic Press, New York. Erlebacher, J. (2005) J. Electrochem. Soc., 151, C614. Chen, P.-Y. and Hussey, C.L. (2007) Electrochim. Acta, 52, 1857. Abbott, A.P., Capper, G., Davies, D.L., Munro, H.L., Rasheed, R.K., and Tambyrajah, V. (2001) Chem. Commun., 7, 1010.

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

147

6 Electrodeposition of Semiconductors in Ionic Liquids Natalia Borisenko, Sherif Zein El Abedin, and Frank Endres

In this chapter we report on the electrodeposition of semiconductors in ionic liquids. It is shown that ionic liquids are, due to their extraordinary physicochemical properties, well suited as a solvent medium for the electrodeposition of elemental semiconductors (like Si and Ge), their mixtures (Six Ge1−x ) and compound semiconductors (GaAs, AlSb, InSb, ZnTe, CdTe, CuInSe2 , etc.).

6.1 Introduction

There is a wide variety of applications for elemental and compound semiconductors. Compound semiconductor thin films, for example, are used in many optoelectronic devices like photon detectors, light emitting diodes (LED), photovoltaics and lasers. Cadmium based II–VI semiconducting thin films, such as CdTe, CdSe, CdS, and CdHgTe, display a variety of band gaps and lattice constants, which make them interesting for optoelectronic applications [1, 2]. The family of III–V compound semiconductors, in particular antimony-based semiconductors (AlSb, GaSb, and InSb), are of great interest as barrier materials in high-speed electronics and long-wavelength optoelectronic devices [3, 4]. Au–Cd alloys are employed as ohmic contacts with semiconducting films and may provide additional doping in these materials [5]. Ternary compound semiconductors, for example CuInSe2 (CIS), are promising materials for thin film photovoltaic applications due to their stability, direct energy band gap and high absorption coefficient [6]. Elemental semiconductors, such as Si and Ge, are widely used as wafer material for different electronic applications, and junctions of n- and p-doped Si are still interesting for photovoltaic applications. Ge quantum dots made by molecular beam techniques under ultrahigh vacuum (UHV) conditions have interesting optical properties. For example, Ge quantum dots on Si(111) show a photoluminescence at about 1 eV [7]. Silicon nanocrystals embedded in a SiO2 matrix have been discussed for the development of nanoscale silicon-based lasers [8]. Electrochemical deposition is one of the main fields in electrochemistry, both in industrial processes and in fundamental research. It has been applied to make Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

148 6 Electrodeposition of Semiconductors in Ionic Liquids

semiconductors by electrochemical means for over 30 years and a good general overview can be found in Ref. [9]. However, standard industrial procedures for semiconductor electrodeposition are rare. Many studies on the deposition of semiconductors and their characterization have been performed in different solutions such as aqueous media, organic solutions and molten salts. In fundamental research, most of the investigations have been performed by molecular beam epitaxy (MBE) under UHV conditions. In industrial processes, chemical or physical vapor deposition methods are preferred. The obtained layers are of a high quality, but such processes are cost-intensive thus making the deposits quite expensive. Therefore electrodeposition of semiconductors would be technically interesting as electrodeposition is, in contrast to UHV techniques, a comparatively cheap process; only the imagination of the user limits the size of the objects onto which the semiconductor can be deposited. Recently Stickney et al. demonstrated that compound semiconductors like CdTe, CdSe, CdS or HgSe can be electrodeposited in aqueous media by the electrochemical atomic layer epitaxy (ECALE) method [10–15]. The desired semiconductor is made by the subsequent layer-by-layer growth of the respective elements. The direct electrodeposition of compound semiconductors in one step is often difficult for kinetic reasons. At room temperature the elements are codeposited in varying amounts together with the desired semiconductor and variation in temperature can strongly affect the quality of the films [16]. The electrodeposition of III–V compound semiconductors, like InSb and InAs, has also been investigated in aqueous solutions [17–21]. Unfortunately, elemental semiconductors like silicon, germanium and their mixtures (Six Ge1−x ) cannot be obtained in aqueous solutions as the deposition is strongly disturbed by hydrogen evolution. Instead of Si deposition there would only be hydrogen evolution in aqueous media. Macroscopically thick and amorphous germanium films can be obtained from GeI4 dissolved in propylene glycol at elevated temperature under galvanostatic conditions [22]. Szekely et al. showed that a roughly 130 µm thick germanium layer was electrodeposited from GeCl4 dissolved in propylene glycol at 59 ◦ C and a constant current density of 0.4 mA cm−2 . Unfortunately, the current efficiency in these systems is only about 1% [23]. Ge can also be electrodeposited in high-temperature molten salts [24]. There have been several attempts to electrodeposit silicon in organic solvents [25–27], and smooth and uniform silicon deposits up to 0.25 µm thick were described. However, Auger electron spectroscopy analysis of such deposits evidenced oxygen content and it was not clear whether the deposit was oxidized during the deposition process or if it was a consequence of an open porosity in the film. Of course, silicon can also be electrodeposited in high-temperature molten salts [28]. Recently the electrodeposition of silicon from its halides in non-aqueous solutions was investigated [29]. These authors also reported strong oxidation of the electrochemically made silicon. The electrodeposition of semiconductors in ionic liquids is a comparably young research area. Ionic liquids are, due to their extraordinary physical properties, (see Chapter 3) very interesting as electrolytic media. They are good solvents for a variety of both organic and inorganic materials (depending on the liquid), they are immiscible with a variety of organic solvents and, in part, immiscible with water, they are nonvolatile (in most cases) and, hence, can be

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.3 Indium Antimonide 149

used even in ultra-high-vacuum systems. Ionic liquids exhibit wide electrochemical windows of up to 6 V, and wide thermal windows of 300–400 ◦ C. Depending on the cation/anion combination, the temperature can be varied over several hundred degrees, so that kinetic barriers in semiconductor compound formation could be overcome. The unique properties of ionic liquids open the opportunity to apply them as solvents for electrodeposition of semiconductors, which are hardly (or even not at all) accessible from aqueous solutions, such as Si, Ge, GaAs, etc. The electrodeposition of GaSb, InP, InSb, InSe and ternary semiconductors in ionic liquids is also interesting, especially at elevated temperatures as kinetic barriers in compound formation might be more easily overcome at temperatures around 200–300 ◦ C. In this chapter we summarize recently published results on the electrodeposition of semiconductors in different ionic liquids.

6.2 Gallium Arsenide

GaAs is a well-known III–V compound semiconductor with a direct band gap of 1.43 eV at room temperature. Due to the high mobility of the charge carriers, GaAs-based electronic devices can operate at higher frequencies than equivalent Si devices, resulting in faster electronics, that makes these semiconductor interesting for many optoelectronic applications including semiconductor lasers, LEDs and solar cells. The direct electrodeposition of GaAs in ionic liquids has been studied by two groups. In 1986 Wicelinski and Gale showed that GaAs can, in principle, be electrodeposited from GaCl3 and AsCl3 at 40 ◦ C in the Lewis acid chloroaluminate ionic liquid composed of AlCl3 and 1-butylpyridinium chloride [30]. The authors report that Al codeposition occurs in the underpotential deposition regime on the Ga surface. In order to minimize Al contamination Carpenter and Verbrugge employed a chlorogallate ionic liquid [31, 32]. It was shown that GaAs film can be obtained at room temperature in the Lewis basic GaCl3 /1-methyl-3-ethylimidazolium chloride ionic liquid, to which AsCl3 was added. The quality of the deposit can be improved by thermal annealing, which makes this method promising, in principle, for the electrodeposition of GaAs-based compound semiconductors. However, GaCl3 -based ionic liquids are extremely aggressive and AsCl3 is extremely poisonous so that such liquids would involve enormous security issues.

6.3 Indium Antimonide

InSb is an important compound semiconductor of the III–V family for optoelectronic purposes. At room temperature the semiconductor has a direct band gap of 0.17 eV and a high mobility of charge carriers. Similar to GaAs, it was reported that InSb can be directly electrodeposited at 45 ◦ C in the Lewis basic chloroindate ionic liquid InCl3 /1-methyl-3-ethylimidazolium chloride, to which SbCl3 was

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

150 6 Electrodeposition of Semiconductors in Ionic Liquids

added before the deposition [33]. The In/Sb ratio in the deposit is strongly dependent on the applied electrode potential. Consequently elemental Sb and In can also be found in the films. Recently Sun et al. employed the Lewis basic 1-ethyl3-methylimidazolium chloride/tetrafluoroborate ionic liquid containing InCl3 and SbCl3 for InSb deposition [34]. The composition of the InSb films also depends strongly on the deposition potential. However, the crystallinity of the deposits is strongly improved by increasing the deposition temperature, and polycrystalline InSb can be directly electrodeposited at 120 ◦ C without additional annealing. The band gap was determined by absorption spectroscopy to be 0.2 eV. Although the quality of the deposits depends on the absolute concentrations of In(III) and Sb(III) species and although individual indium and antimony crystals can be found in the films this result proves that the wide thermal windows of ionic liquids and the wide electrochemical windows allow one to find parameters under which the compound semiconductors can be made directly. 6.4 Aluminum Antimonide

The binary semiconductor AlSb is, like GaAs and InSb, one of the III–V semiconductors. In particular, AlSb is a highly efficient solar cell material. It exhibits a direct band gap of 2.5 eV and an indirect band gap of 1.2 eV at room temperature. The electrodeposition of AlSb was investigated at room temperature in the Lewis neutral ionic liquid AlCl3 /1-butyl-3-methyl-imidazolium chloride [35, 36]. A liquid containing Sb(III) was prepared by addition of SbCl3 /1-butyl-3-methylimidazolium chloride to the chloroaluminate ionic liquid. The electrodeposition of AlSb was investigated by in situ scanning tunneling microscopy (STM) and in situ scanning tunneling spectroscopy (STS). A band gap of about 2.0 eV was obtained. As in the case of GaAs and InSb, codeposition of the elements occurs, furthermore strong doping effects by the elements occur if the deposition is performed at electrode potentials away from the compound deposition potential. In future studies it should be investigated whether deposition at elevated temperatures (∼ 200 ◦ C) allows better control of AlSb-stoichiometry. Furthermore the use of air- and water stable ionic liquids might lead to more reproducible results. 6.5 Zinc Telluride

ZnTe is usually applied in switching devices and in solar cells. It is one of the II–VI compound semiconductors with a direct band gap of 2.3 eV at room temperature. The electrodeposition of ZnTe was investigated by Sun et al. in the Lewis basic ZnCl2 /1-ethyl-3-methylimidazolium ionic liquid containing propylene carbonate as a cosolvent at 40 ◦ C [37]. 8-Quinolinol was added to the solution to shift the reduction of Te(IV) to more negative potential, thus facilitating the codeposition. The composition of the ZnTe deposits is dependent on the deposition potential and

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.7 Germanium

on the concentration of Te(IV) in the solution. After thermal annealing the band gap was determined by UV/vis absorption spectroscopy to be 2.3 eV, which is in good agreement with ZnTe films made by others methods.

6.6 Cadmium Telluride

CdTe, a II–VI compound semiconductor with a direct band gap of 1.44 eV at room temperature, is, from its physical properties, a promising photovoltaic material. The electrodeposition of CdTe in ionic liquid was published recently by Sun et al. [38]. They were able to show that the semiconductor can be electrodeposited at elevated temperature (above 120 ◦ C) in the Lewis basic 1-ethyl-3-methylimidazolium chloride/tetrafluoroborate ionic liquid containing CdCl2 and TeCl4 . CdTe films were obtained by the underpotential deposition (UPD) of Cd on the deposited Te. The deposit composition was independent of the deposition potential within the Cd UPD regime. The crystallinity of the deposits is improved by increasing the deposition temperature, which again demonstrates the high potential of the wide thermal windows of ionic liquids for compound electrodeposition.

6.7 Germanium

Germanium is an elemental semiconductor with an indirect band gap of 0.67 eV at room temperature in the microcrystalline phase. Its crystal structure is determined by the tetrahedral symmetry of Ge in the crystalline phase. An interesting aspect is that Ge nanoparticles with diameters of only a few nanometers exhibit a sizedependent photoluminescence. Nanocrystalline Ge is a direct semiconductor and it is regarded today as a promising candidate for optical sensors. However, almost all studies on the production and characterization of Ge nanocrystals or quantum dots hitherto have been performed under UHV conditions, which require a high instrumental effort for a possible nanotechnological process. An electrochemical process would be quite interesting as there is, in principle, no limit on surface area and geometry, furthermore electrochemical experiments are comparatively easy to perform. The electrodeposition of germanium in ionic liquids was primarily investigated in our group. The original aim was to find out if germanium can be made by electrochemical means at all. In situ STM and in situ STS techniques were employed for this purpose. These techniques allow one to investigate the initial stages of the semiconductor electrodeposition and to understand the deposition process on the nanometer scale. The electrodeposition of Ge was studied at room temperature from GeX4 (where X = I, Br, Cl) on Au(111) in the ionic liquid 1-butyl-3methylimidazolium hexafluorophospate ([BMIM]PF6 ) [39–42]. At the time of the experiments this ionic liquid was one of a few which could easily be prepared

151

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

152 6 Electrodeposition of Semiconductors in Ionic Liquids

Fig. 6.1 CV of pure [BMIM]PF6 on Au(111). The scan rate is 1 mV s−1 . Mainly capacitive currents flow between the cathodic and the anodic limits. The electrochemical window is a little more than 4 V.

with water levels below 20 ppm and was, therefore, the best choice for such investigations. All experiments have to be performed under inert gas conditions as the germanium halides react rapidly with water. For comparison purposes the cyclic voltammograms were calibrated versus the overpotential deposition (OPD) of germanium. The electrochemical window of dry [BMIM]PF6 on Au(111) is a little more than 4 V, as can be seen in the cyclic voltammogram (CV) of Figure 6.1. In dry and high purity liquids only capacitive currents flow between the cathodic and anodic limits. At the cathodic limit the organic cation is irreversibly reduced and STM pictures in this potential regime show that a less defined deposit is formed on the electrode surface. This decomposition product is dissolved in the liquid when the potentiostat is switched off, turning the originally colorless liquid to purple. At the anodic limit gold oxidation occurs which, in the initial state, can also be probed by in situ STM. In the following the focus is on the electrodeposition of Ge from GeCl4 . The electrodeposition is quite similar from all the halides but the oxidation of the gold substrate was least severe in the case of GeCl4 . If [BMIM]PF6 ionic liquid is saturated with GeCl4 (Figure 6.2), two main reduction processes (P1 and P2 ) are observed in the cathodic regime [42]. The first reduction peak, with a minimum at +500 mV vs. Ge (P1 ) is attributed to the reduction of Ge(IV) to Ge(II). At potentials below 0 mV (P2 ) the bulk deposition of Ge from Ge(II) sets in, as can be seen with the naked eye. The rising cathodic current at about −1000 mV vs. Ge is attributed to the irreversible reduction of the organic cation. If only P1 is passed, an oxidation process is not observed. If Ge deposition is performed an oxidation peak at ∼1000 mV is observed, which means that this peak must be correlated to Ge electrooxidation. A series of oxidation peaks above +1500 mV is also observed if the electrode potential is cycled between +1000 and

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.7 Germanium

Fig. 6.2 CV of GeCl4 saturated in dry [BMIM]PF6 on Au(111). The scan rate is 1 mV s−1 . Upon addition of GeCl4 two mainly irreversible reduction peaks (P1 and P2 ) are observed. The process P1 is correlated with the reduction of Ge(IV) to Ge(II). At P2 the bulk deposition of Ge is

observed. The irreversible reduction of the imidazolium ion starts at −1000 mV (P3 ). A strong oxidation peak at +1000 mV is attributed to partial Ge dissolution. At electrode potentials above +1500 mV gold oxidation sets in.

+3000 mV vs. Ge, thus avoiding GeCl4 reduction. Therefore these processes are due to the oxidation of the gold substrate which is difficult to probe with in situ STM. At the open circuit potential, OCP, (+1200 mV vs. Ge) a typical Au(111) surface with step heights of 250 pm is probed (Figure 6.3(a)). When the electrode potential is reduced to +1000 mV, the step edges become quite obviously decorated, and this decoration is not observed in the pure ionic liquid. First, small two-dimensional islands with heights between 100 and 150 pm appear at about +950 mV (Figure 6.3(b)). These islands can be reversibly stripped from the surface. When the electrode potential is further reduced to +750 mV, islands with an average height of 250 pm form. If the potential is set back to +1200 mV the islands dissolve and tiny holes with a depth of about 100 pm appear. The holes completely heal in a few minutes and a flat terrace-like gold structure is again obtained. Between +300 mV and 0 mV a rough but completely closed layer with a maximum thickness of 300 pm forms (Figure 6.3(c)). The in situ I/U tunnelling spectrum clearly reveals the metallic behavior of this layer but the tunnelling barrier is much higher than for a pure gold substrate at OCP (Figure 6.3(d)), where even a linear increase in tunneling current can occur. It is therefore likely that, in the UPD regime, a surface alloying between Au and Ge takes place so that the UPD germanium gets a more noble metallic character. The deposit does not grow further if the electrode potential of the STM tip is set to sufficiently high values. However, if the tip potential is kept close to the bulk deposition potential of Ge, cluster agglomerates composed of small clusters of only some nanometers in diameter grow on the surface, probably due to a jump from the tip to the sample (Figure. 6.4(a)). In situ I/U tunnelling

153

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

154 6 Electrodeposition of Semiconductors in Ionic Liquids

Fig. 6.3 The series of STM pictures shows the UPD of Ge on Au(111) in GeCl4 /[BMIM]PF6 . At +1200 mV vs. Ge (OCP) a typical Au(111) structure with step height 250 pm is observed (a). When the electrode potential is reduced to +950 mV small two-dimensional islands with heights between 100 and 150 pm appear (b). Be-

tween +300 and 0 mV a completely closed but rough layer with an average height of 300 pm forms (c). The layer shows metallic behavior but the tunneling barrier is higher than that for the pure Au substrate at OCP. Most likely an alloying between Ge and Au occurs (d).

spectroscopy on different sites of these clusters shows a bias range of about 500 mV with almost zero tunneling current (Figure 6.4(b)). At −50 mV vs. Ge (slightly in the OPD regime) islands/crystallites with diameters of 50 nm and heights of 5–10 nm are observed (Figure 6.5(a)). The band gap of these individual crystals is 0.7 ± 0.1 eV, which is typical for intrinsic bulk germanium at room temperature. If the electrode potential is further reduced to −200 mV vs. Ge about 100 nm thick deposits with a band gap of 0.7 ± 0.1 eV form (Figure 6.5(b, c)). The film is composed of nanocrystals and tongue-like germanium islands. An XPS study of a micrometer

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.8 Silicon 155

Fig. 6.4 If the bias is only +200 mV small cluster agglomerates grow on the surface at +100 mV vs. Ge (a). A higher resolution image (inset in (a)) shows that the cluster agglomerates are composed of small clus-

ters with diameters of only a few nm. The in situ I/U tunneling spectrum reveals that these islands exhibit a bias range of about 500 mV with almoust zero tunneling current (b).

thick Ge film made from GeX4 shows that indeed elemental Ge was obtained [43]. The electrochemically made Ge is, however, subject to some attack by environmental oxygen (Figure 6.6(a, b)). We would like to mention a further interesting effect which we observed when GeI4 microcrystals were directly reduced in the ionic liquid, instead of microcrystals Ge crystals with sizes around 100–200 nm were obtained (Figure 6.7). Most likely the confined diffusion space between crystal and electrode surface led to comparatively small crystals. These results show that not only Ge layers but also Ge nanocrystals can be made electrochemically in ionic liquids by adjusting the experimental parameters.

6.8 Silicon

Currently silicon is still one of the most important semiconductors as it is the basis of any computer chip. It exhibits an indirect band gap of 1.1 eV at room temperature in the microcrystalline phase. Similar to Ge, silicon nanoparticles show a sizedependent photoluminescence. It was reported by Katayama et al. that a thin Si layer can be electrodeposited in 1-ethyl-3-methylimidazolium hexafluorosilicate at 90 ◦ C [44]. However, upon exposure to air the deposit reacted completely to SiO2 , which makes it difficult to decide whether the deposit was semiconducting or not. Recently, we showed for the first time that silicon can be well electrodeposited from SiCl4 in the air and water stable ionic liquid 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide ([BMP]Tf2 N) [45, 46]. This ionic liquid can be

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

156 6 Electrodeposition of Semiconductors in Ionic Liquids

Fig. 6.5 The STM images represent the OPD of Ge on Au(111) in GeCl4 /[BMIM]PF6 . At −50 mV vs. Ge clusters with diameters of 50 nm and heights of 5–10 nm are observed (a). When the electrode potential is set to −200 mV an about

100 nm thick Ge film, composed of small nanoclusters and tongue-like islands, is probed (b). The in situ tunneling spectrum also shows semiconducting behavior with a symmetrical band gap of 0.7 ± 0.1 eV, typical for elemental Ge (c).

dried to water contents below 1 ppm. Similar to Ge, the experiments were performed under inert gas conditions. The reversible ferrocene/ferrocinium (Fc/Fc+ ) redox couple was employed as a reference electrode. The liquid itself exhibits on Au(111) an electrochemical window slightly more than 5 V (Figure 6.8). At the cathodic limit a series of peaks (C1 –C3 ) is observed prior to the irreversible reduction of organic cation at −3200 mV vs. Fc/Fc+ . At the anodic limit at +2000 mV vs. Fc/Fc+ gold oxidation sets in. The oxidation processes A3 and A4 are only observed if the reduction processes C3 and C4 have been passed. For the peaks C1 and C2 the respective oxidation processes are missing. It was shown that the peaks C1 to C3 are correlated to the irreversible breakdown of the Tf2 N ion [47, 48]. If SiCl4 (0.1 mol l−1 ) is dissolved in [BMP]Tf2 N two main reduction processes preceding the reduction of the organic cation at C3 are observed (Figure 6.9). As

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.8 Silicon 157

Fig. 6.6 XPS spectra of germanium deposits made potentiostatically at −500 mV vs. Ge for 6 h from GeBr4 (a) and from GeCl4 (b) on Au(111). For both deposits Auger peaks with binding energies of ∼105 and ∼109 eV are found, which can be attributed to elemental Ge. A transition at 114–115 eV is likely due to a chemical shift as a higher oxidation state of Ge leads to

transitions at higer energies. For Ge from GeBr4 , the 3p transitions with binding energies of ∼122 and ∼126 eV are found, as expected for elemental Ge. For Ge from GeCl4 besides the 3p transitions for elemental Ge, two more transitions shifted by 4–5 eV with respect to 3p are obtained, indicating the presence of Ge(IV) in the deposit.

C2 only appears when SiCl4 is in the solution, this peak must be correlated to the bulk deposition of silicon. It is interesting that the reduction of the organic cation on a silicon surface is strongly hindered. Furthermore, the decomposition of the ionic liquid might passivate the Si surface as the deposition of Si can also start in the anodic branch at C* . Since the peak C1 was not observed on HOPG [45] and as there is no deposition at this potential regime, the process might be correlated with ionic liquid breakdown and surface restructuring/reconstruction. A further explanation might be the formation of low valent silicon species in the solution.

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

158 6 Electrodeposition of Semiconductors in Ionic Liquids

Fig. 6.7 SEM image of a germanium deposit, obtained upon electrodeposition under a GeI4 crystal at −1000 mV vs. Ge. The surface is completely covered by a thin germanium layer. A collection of nanoclusters with a grain size of about 100 mn is obtained.

The broad oxidation processes at E > 0 mV are partly due to Si oxidation, gold oxidation and oxidation of cation reduction products. STM images in Figure 6.10 show the growth of silicon in SiCl4 (0.1 mol l−1 )/[BMP]Tf2 N on Au(111). Between −300 mV (OCP) and −600 mV vs. Fc/Fc+ , instead of flat terraces, which would be typical for Au(111), the terraces show a worm-like restructuring (Figure 6.10(a)), which is less clear if SiCl4 is in

Fig. 6.8 CV of dry ultrapure [BMP]Tf2 N ionic liquid on Au(111) at a scan rate of 10 mV s−1 . The electrochemical window is about 5 V, limited by the reduction of the organic cation at C4 and by gold oxidation

at +2000 mV vs. Fc/Fc+ . Within the electrochemical window a series of reduction peaks C1 –C3 is observed, due to an irreversible breakdown of the Tf2 N anion.

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.8 Silicon 159

Fig. 6.9 CV of SiCl4 (0.1 M) in [BMP]Tf2 N ionic liquid on Au(111) at a scan rate of 10 mV s−1 . The bulk deposition of silicon starts at C2 . The reduction of the organic cation (C3 ) is hindered on the silicon surface. The first reduction process C1 is not

correlated with a definite surface process. In the reverse scan further silicon deposition is observed (C* ). Strong oxidation processes at a potential of more positive than 0 mV are correlated to the oxidation of silicon, gold and cation reduction products.

the solution. The step height between the terraces is still 250 ± 30 pm. The defects/vacancy islands are one monolayer deep and the width is about 10–20 nm. When the electrode potential is reduced to more negative values the number of these vacancy islands is strongly decreased. Therefore the allocation of the reduction process C1 in Figure 6.9 to a definite surface process is difficult. Between −1100 mV and −1600 mV a typical terrace-like Au(111) surface with a step height of 250 pm is probed (Figure 6.10(b)). At −1700 mV small silicon islands with diameters of less than 50 nm and heights between 150 and 450 pm start to grow (Figure 6.10(c)). With time, the number of these islands very slowly increases and they grow slightly in height. After 1 h a completely closed thin layer of silicon whose height is in the nanometer regime forms (Figure 6.10(d)). Some small islands rise above this thin layer. The in situ I/U spectrum reveals typical semiconducting behavior with a band gap of 1.1 ± 0.2 eV, which is in excellent agreement with literature data for microcrystalline silicon in the bulk phase at room temperature (1.1 eV). When the electrode potential is further reduced the islands grow above the surface and merge laterally, resulting in silicon agglomerates (Figure 6.11(a)). These structures can be as high as 10 nm, their width can reach 30 nm and they also exhibit a band gap of 1.1 eV (Figure 6.11(b)). SEM pictures show that a very thin layer of small clusters/crystallites forms first on the gold surface, followed by crystalline agglomerates that finally leads to a 500–1000 nm thick silicon layer (Figure 6.12). In order to get more ex situ information on the electrochemically made silicon we performed a detailed XPS study. For this purpose the Si was made inside the inert gas glove box. The deposit was subsequently purified by rinsing in isopropanol inside the glove box and transferred via a transport chamber to the XPS device. Thus

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

160 6 Electrodeposition of Semiconductors in Ionic Liquids

Fig. 6.10 The sequence of STM pictures shows the nanoscale growth of Si in the SiCl4 (0.1 M)/[BMP]Tf2 N on Au(111) probed by in situ STM. Instead of a typical flat Au(111) surface, a worm-like structure is probed at −300 mV vs. Fc/Fc+ (OCP) (a). When the electrode potential is reduced a

typical Au(111) terrace-like surface with a step height of 250 pm is observed (b). If the electrode potential is set to −1700 mV, small silicon islands with a width of less than 60 nm and 150–450 pm in height start to grow (c). After 1 h an about 5 nm high close silicon layer forms (d) close to the equilibrium potential.

the sample was never in air. As an example Figure 6.13 shows the XPS spectrum of the Si 2p peak. Besides the elemental peak at 101.3 eV there is strong evidence for SiOX at 104.4 eV. As discussed elsewhere [49] silicon is deposited electrochemically, but even in an inert gas glove box with an oxygen concentration as low as 1 ppm it is attacked by oxygen at the surface. The XPS study shows undoubtedly that elemental silicon can be electrodeposited in ionic liquids. 6.9 Grey Selenium

Grey selenium exhibits both photovoltaic and photoconductive properties, which make it useful in the production of photocells and solar cells. Moreover,

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.9 Grey Selenium

Fig. 6.11 When the electrode potential is reduced to −1800 mV vs. Fc/Fc+ silicon islands grow above the surface and merge laterally leading to agglomerates (a). The in situ I/U tunneling spectrum shows that both the layer and the islands exhibit a band gap of 1.1 ± 0.2 eV, typical for mycrocrystaline semiconducting Si (b).

Se-containing compound semiconductors, such as InSe, CdSe or CuInSe2 (CIS) have many optoelectronic applications, including advanced solar cells, IR detectors and solid-state lasers. The CIS solar cells, especially, are very promising as a highly efficient power supply. The electrodeposition of selenium has been intensively investigated in aqueous solutions [50–53]. However, the exclusive electrodeposition of grey selenium in aqueous solutions is pretty difficult as at temperatures below 100 ◦ C red and black selenium, which are both insulators, grow in certain amounts together with the desired grey phase. Consequently, in the technical process selenium is applied by gas phase condensation. Thermodynamically a phase transition

Fig. 6.12 SEM micrographs of electrodeposited Si, made potentiostatically at −2700 mV vs. Fc/Fc+ . The surface is completely covered by a thin silicon layer composed of individual clusters. Globular

50–150 nm wide crystallites consisting of many tiny crystals grow above this layer (a). A 500 nm thick silicon layer consists of coherent spherical crystallites (b).

161

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

162 6 Electrodeposition of Semiconductors in Ionic Liquids

Fig. 6.13 High resolution XPS spectrum of a nanoscale silicon deposit made potentiostatically at −2200 mV vs. Pt quasi-reference for 2 h on a stainless steel substrate. (Dots: original data obtained from the measurement. Solid lines: fitted data. Dotted line:

sum of both fitted contributions.) A Si 2p peak consists of two different contributions with binding energies of 101.3 and 104.4 eV, which can be attributed to elemental Si and SiOX , respectivelly.

from amorphous red to crystalline grey selenium occurs at about 80 ◦ C. But even at 100 ◦ C in aqueous solutions the deposit does not contain solely the grey phase of Se. Obviously grey selenium can only be electrodeposited at elevated temperatures of more than 100 ◦ C, which cannot be achieved in aqueous solutions. The direct electrodeposition of grey selenium can be performed in ionic liquids as the deposition process can be realized at elevated temperatures due to the high thermal stability and low vapor pressure of most of the ionic liquids. In a recent paper we reported that grey selenium can be well electrodeposited from SeCl4 in [BMP]Tf2 N ionic liquid [54]. All experiments were performed under an inert gas atmosphere. Unfortunately, we could not employ ferrocene as an internal reference, due to the complexity of voltammograms and the need to exclude any interference with ferrocene. Therefore, a Pt-wire had to be used as a quasi-reference electrode. In our experience Pt has a sufficiently stable electrode potential for a while under the applied conditions. The cyclic voltammogram of [BMP]Tf2 N containing 0.1 mol l−1 SeCl4 on a platinum substrate at 25 ◦ C is presented in Figure 6.14(a). At a potential of −750 mV vs. Pt a dark red deposit forms on the electrode surface, obviously passivating it. It is likely that this peak is correlated to the reduction of Se(IV) to the red phase of elemental Se. It cannot be excluded that the black phase is also formed. If the

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

6.9 Grey Selenium

Fig. 6.14 CVs of SeCl4 (0.1 M) in [BMP]Tf2 N ionic liquid on Pt substrate at 25 ◦ C (a) and 150 ◦ C (b). Scan rate is 10 mV s−1 . At about −750 mV the deposition of red amorphous Se occurs. The red colour of the deposit turns to grey at C4 .

The deposited Se is partly dissolved at A3 . The shoulders C1 and C2 and their anodic counterparts A1 and A2 might be related to two different UPD processes. The pair C5 and A5 are likely to be due to the reduction of the deposited Se to Se2− .

temperature is increased to 150 ◦ C (Figure 6.14(b)) five cathodic processes C1 –C5 and their corresponding anodic counterparts A1 –A5 are observed. C3 and C4 are correlated with the electrodeposition of selenium. Visually, first a red deposit forms at C3 , which turns to a grey colour at C4 . The peaks C1 and C2 are presumably correlated with different UPD processes. The peaks C5 and A5 are likely to be associated with the further reduction of the deposited selenium to Se2− as the selenium film can disappear completely at this electrode potential. As one example the SEM picture in Figure 6.15 shows an electrodeposited Se layer made potentiostatically at −1100 mV vs. Pt at 150 ◦ C. The XRD pattern of

Fig. 6.15 SEM image of Se layer, made potentiostatically at −1100 mV vs. Pt at 150 ◦ C. The XRD pattern of this layer shows that crystalline grey Se is electrodeposited.

163

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

164 6 Electrodeposition of Semiconductors in Ionic Liquids

the electrodeposit reveals the characteristic peaks of the crystalline grey selenium. From XRD and SEM alone it cannot be excluded that some red or black Se also form, thermodynamically, however, it is unlikely at these temperatures. In our opinion the electrodeposition of selenium is quite promising for a variety of applications. For example, the possibility to deposit grey selenium, indium, and copper in one ionic liquid at variable temperatures might be regarded as the first step in making selenium-containing compound semiconductors like CIS by electrochemical means.

6.10 Conclusions

In this chapter we have summarized selected literature data on the electrodeposition of semiconductors in ionic liquids. It has been demonstrated that elemental silicon, germanium, and selenium can be elecrodeposited in ionic liquids. Furthermore, it is shown that compound semiconductors like InSb, AlSb, CdTe and others can be made, especially at elevated temperatures where kinetic barriers are easier to overcome, even allowing the exclusive electrodeposition of grey selenium. In this context ionic liquids are very promising for semiconductor electrodeposition. Both wide electrochemical and thermal windows allow processes which are impossible in aqueous or organic solvents.

References 1 Gore, R.B., Pandey, R.K., and Kulkarni, S.K. (1989) J. Appl. Phys., 65, 2693. 2 Loizos, Z., Mitsis, A., Spyrellis, N., Froment, M., and Maurin, G. (1993) Thin Solid Films, 235, 51. 3 Razeghi, M. (2003) Eur. Phys. J.: Appl. Phys., 23, 149. 4 Baaziz, H., Charifi, Z., and Bouarissa, N. (2001) Mater. Chem. Phys., 68, 197. 5 Kelly, J.J., Rikken, J.M.G., Jacobs, J.W.M., and Valster, A. (1988) J. Vac. Sci. Technol. B, 6, 48. 6 Schock, H.W. (1996) Appl. Surf. Sci., 92, 606. 7 Leifeld, O., Beyer, A., M¨uller, E., Gr¨utzmacher, D., and Kern, K. (2000) Thin Solid Films, 380, 176. 8 Jaiswal, S.L., Simpson, J.T., Withrow, S.P., White, C.W., and Norris, P.M. (2003) Appl. Phys., A77, 57. 9 Pandey, R.K., Sahu, S.N., and Chandra, S. (1996) Handbook of Semiconductor

10 11

12

13

14

15

Electrodeposition, Marcel Dekker, New York, NY. Gregory, B.W. and Stickney, J.L. (1991) J. Electroanal. Chem., 300, 543. Colletti, L.P., Flowers, B.H., and Stickney, J.L. (1998) J. Electrochem. Soc., 145, 1442. Flowers, B.H. Jr., Wade, T.L., Garvey, J.W., Lay, M., Happek, U., and Stickney, J.L. (2002) J. Electroanal. Chem., 524–525, 273. Mathe, M.K., Cox, S.M., Venkatasamy, V., Happek, U., and Stickney, J.L. (2005) J. Electrochem. Soc., 152, C751. Venkatasamy, V., Mathe, M.K., Cox, S.M., Happek, U., and Stickney, J.L. (2005) Electrochim. Acta, 51, 4347 Venkatasamy, V., Jayaraju, N., Cox, S.M., Thambidurai, C., Happek, U., and Stickney, J.L. (2006) J. Appl. Electrochem., 36, 1223.

c06 (JWBG008-Endres)

December 24, 2007

17:31

Char Count=

References 16 Raza, A., Engelken, R., Kemp, B., Siddiqui, A., and Mustafa, O. (1995) Proc. Arkansas Acad. Sci., 49, 143. 17 Mengoli, G., Musiani, M.M., Paolucci, F., and Gazzano, M. (1991) J. Appl. Electrochem., 21, 863. 18 Kozlov, V.M., Agrigento, V., Bontempi, D., Canegallo, S., Moraitou, C., Toussimi, A., Bicelli, L.P., and Serravalle, G. (1997) J. Alloys Compds., 259, 234. 19 Wade, T.L., Vaidyanathan, R., Happek, U., and Stickney, J.L. (2001) J. Electroanal. Chem., 500, 322. 20 Kozlov, V.M., Bozzini, B., and Bicelli, L.P. (2004) J. Alloys Compds., 366, 152. 21 Fulop, T., Bekele, C., Landau, U., Angus, J., and Kash, K. (2004) Thin Solid Films, 449, 1. 22 Fink, C.G. and Dokras, V.M. (1949) J. Electrochem. Soc., 95, 80. 23 Szekely, G. (1951) J. Electrochem. Soc., 98, 318. 24 Monnier, R. and Tissot, P. (1964) Helv. Chim. Acta, 47, 2203. 25 Agrawal, A.K. and Austin, A.E. (1981) J. Electrochem. Soc., 128, 2292. 26 Gobet, J. and Tannenberger, H. (1986) J. Electrochem. Soc., 133, C322. 27 Gobet, J. and Tannenberger, H. (1988) J. Electrochem. Soc., 135, 109. 28 Matsuda, T., Nakamura, S., Ide, K., Nyudo, K., Yae, S.J., and Nakato, Y. (1996) Chem. Lett., 7, 569. 29 Nicholson, J.P. (2005) J. Electrochem. Soc., 152, C795. 30 Wicelinski, S.P. and Gale, R.J. (1986) in Fifth International Symposium on Molten Salts, (PV 86-1), (eds M.-L. Saboungi, D.S. Newman, K. Johnson, and D. Inman), the Electrochemical Society Softbound Proceedings Series, Pennington, NJ, p. 144 31 Carpenter, M.K. and Verbrugge, M.W. (1990) J. Electrochem. Soc., 137, 123. 32 Verbrugge, M.W. and Carpenter, M.K. (1990) AIChE J., 36, 1097. 33 Carpenter, M.K. and Verbrugge, M.W. (1994) J. Mater. Res., 9, 2584. 34 Yang, M.H., Yang, M.C., and Sun, I.W. (2003) J. Electrochem. Soc., 150, C544.

35 Aravinda, C.L. and Freyland, W. (2006) Chem. Commun., 16, 1703. 36 Mann, O., Aravinda, C.L., and Freyland, W. (2006) J. Phys. Chem. B, 110, 21521. 37 Lin, M.C., Chen, P.Y., and Sun, I.W. (2001) J. Electrochem. Soc., 148, C653. 38 Hsiu, S.I. and Sun, I.W. (2004) J. Appl. Electrochem., 34, 1057. 39 Endres, F. and Schrodt, C. (2000) Phys. Chem. Chem. Phys., 2, 5517. 40 Endres, F. (2001) Phys. Chem. Chem. Phys., 3, 3165. 41 Endres, F. and Zein El Abedin, S. (2002) Phys. Chem. Chem. Phys., 4, 1640. 42 Endres, F. and Zein El Abedin, S. (2002) Phys. Chem. Chem. Phys., 4, 1649. 43 Endres, F. (2002) Electrochem. Solid-State Lett., 5, C38. 44 Katayama, Y., Yokomizo, M., Miura, T., and Kishi, T. (2001) Electrochemistry, 69, 834. 45 Zein El Abedin, S., Borissenko, N., and Endres, F. (2004) Electrochem. Commun., 6, 510. 46 Borisenko, N., Zein El Abedin, S., and Endres, F. (2006) J. Phys. Chem. B, 110, 6250. 47 Howlett, P.C., Izgorodina, E.I., Forsyth, M., and MacFarlane, D.R. (2006) Z. Phys. Chem, 220, 1483. 48 Endres, F., Zein El Abedin, S., and Borissenko, N. (2006) Z. Phys. Chem., 220, 1377. 49 Bebensee, F., Borissenko, N., Frerichs, M., H¨offt, O., Maus-Friedrichs, W., Zein El Abedin, S., and Endres, F. (2008) Z. Phys. Chem., accepted. 50 Huang, B.M., Lister, T.E., and Stickney, J.L. (1997) Surf. Sci., 392, 27. 51 Sorenson, T.A., Lister, T.E., Huang, B.M., and Stickney, J.L. (1999) J. Electrochem. Soc., 146, 1019. 52 Alanyalioglu, M., Demir, U., and Shannon, C. (2004) J. Electroanal. Chem., 561, 21. 53 Zhang, X.Y., Cai, Y., Miao, J.Y., Ng, K.Y., Chan, Y.F., Zhang, X.X., and Wang, N. (2005) J. Cryst. Growth., 276, 674. 54 Zein El Abedin, S., Saad, A.Y., Farag, H.K., Borisenko, N., Liu, Q.X., and Endres, F. (2007) Electrochim. Acta, 52, 2746.

165

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

167

7 Conducting Polymers Jennifer M. Pringle, Maria Forsyth, and Douglas R. MacFarlane

7.1 Introduction

The utilization of ionic liquids for the synthesis and use of conducting polymers brings together two of the most exciting and promising areas of research from recent years. Conducting polymers are organic materials that can display electronic, magnetic and optical properties similar to metals, but that also have the mechanical properties and low density of a polymer. They have the potential to allow the design and fabrication of a vast number of electrochemical devices including photovoltaics, batteries, chemical sensors, supercapacitors, conducting textiles, electrochromics and electromechanical actuators [1–4]. In addition, these materials have the potential to impact in a major way on new biomedical processes such as controlled neural growth, which has significant application in spinal regeneration [5]. The use of electroactive polymers for the fabrication of electromechanical actuators, where the polymer can be made to bend and straighten on application of a small potential, has particular significance in the medical field, where they are being investigated as artificial muscles for a range of prosthetic and therapeutic uses. The use of conducting polymers in photovoltaic devices is another, extremely important area of research and potential applications range from simple solar cells to sunlight harvesting paints and fabrics, for the production of electricity from sunlight. Research into conducting polymers has been increasingly intense for the last 25 years, since MacDiarmid, Heeger and Shirakawa published their seminal work on polyacetylene, which demonstrated that the conductivity of these materials can be increased by several orders of magnitude by doping with anions [6, 7]. The importance of these materials and the progress made in this field is reflected in the award of the Nobel Prize for Chemistry in 2000 to these founding researchers in this area. However, to allow the widespread use of conducting polymers, more research is needed to improve their general performance, and one of their present limitations is the rapid degradation of key properties such as conductivity and electrochemical cyclability. This limitation is primarily a result of the electrolyte used in the Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

168 7 Conducting Polymers

preparation and cycling of the electroactive polymer, which is required as a source of dopants when the polymer is oxidized. These dopants have a significant influence on all the properties of the polymer, including conductivity, mechanical properties, electrochemical efficiency and stability, most likely through the control of the structure and morphology of the material. Ionic liquids offer a unique combination of chemical and physical properties that make them interesting as electrolyte and solvent in one. Interestingly, but not entirely of surprise at the fundamental level, many of the anions that are effective in producing high conductivities in conducting polymers are also the anions that commonly occur in ionic liquid compounds. Thus appropriate ionic liquids provide a superb source of the dopant anions. Conducting polymers, such as poly(aniline), poly(pyrrole) or poly(thiophene) (Figure 7.1) have a conjugated system of delocalized π-orbitals, which allows conduction to occur in the oxidized, or “doped” polymer. The technological interest in these materials lies in their redox behavior. When the films are oxidised in an appropriate electrolytic medium, positive charges are generated along the backbone and solvated counterions enter the polymer from the solution to effect charge balance. This results in an opening of the polymeric structure and an increase in volume. The opposite process occurs on reduction, when the incorporated anions are expelled back into solution and the film recovers its original volume. The size and nature of the dopant counterion incorporated during synthesis can have a dramatic effect on the ion movement occurring during redox processes. There is a competitive reaction between anion expulsion and cation incorporation (from the electrolyte) during the reduction cycle. These two reactions compete to achieve charge neutrality caused by the loss of charge on the polymer backbone (Figure 7.2) [4]. For example, polyelectrolyte dopants, being relatively immobile, tend to remain in the polymer and with such films cation incorporation processes dominate. The potential benefits of using ionic liquids as electrolytes in conducting polymer devices have been investigated by a number of authors in recent years, for applications such as actuators [8–17], supercapacitors [18–20], electrochromic devices [12, 21] and solar cells [22], with significant improvements in lifetimes and device performance reported. For example, in 2002, Lu et al. [12] reported significant improvements in device performance when the ionic liquids 1-ethyl-3-methylimidazolium hexafluorophosphate, [C2 mim][PF6 ], and 1-ethyl-3-methylimidazolium tetrafluoroborate, [C2 mim][BF4 ], were used as supporting electrolytes for poly(pyrrole) and poly(aniline) actuators and for polyethylenedioxythiophene (PEDOT) in electrochromic devices, respectively. For the PEDOT study, the ionic liquid was also used as the growth medium for the electropolymerization of PEDOT. Notably, in both the poly(aniline) in the [C2 mim][BF4 ] and the poly(pyrrole) in the [C2 mim][PF6 ] actuator systems, cation incorporation and expulsion was the predominant strain-generating mechanism during actuation, whereas in the propylene carbonate/tetrabutylammonium hexafluorophosphate (PC/[Bu4 N]PF6 ) system it was anion movement that was observed [12]. Thus, the linear displacement of this actuator is in the opposite direction, illustrating the importance of the nature of ion

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.1 Introduction 169

Fig. 7.1 The chemical structure of common conducting polymers, in their undoped, neutral form [4].

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

170 7 Conducting Polymers

Fig. 7.2 The redox cycling of poly(pyrrole) involving intercalation and expulsion of (a) the anion or (b) the cation from the electrolyte to effect charge balance.

incorporation in such ionic liquid systems. The poly(aniline) actuator maintained electromechanical actuation and electroactivity for >10 000 cycles, without any significant decrease in stress or strain (2000 cycles). 7.3.1.4 Poly(aniline) Osteryoung and coworkers have also investigated the use of chloroaluminate ionic liquids for the synthesis of polyaniline [62, 63]. Unlike pyrrole and thiophene, aniline was successfully polymerized in acidic, neutral and basic chloroaluminate melts, although the best results were obtained using the neutral composition. The oxidation potential of aniline is significantly affected by the composition of the melt, probably as a result of the formation of an adduct between aniline and AlCl3 , similar to that observed for pyrrole. The oxidation potential of the aniline was also found to be influenced by the nature of the electrode used (platinum or glassy carbon), and further shifted upon deposition of polymer onto the working electrode, in a direction that again depended on the composition of the molten salt used [62]. Further, using molten salts of different acidity is reported to result in changes to the backbone structure of the resultant poly(aniline). In comparison to poly(aniline) films formed from aqueous and organic solvents, those prepared and cycled in basic chloroaluminate melts were reportedly very stable, retaining more than 90% of their electrochemical activity after 30 000 cycles at 100 mV s−1 [63]. They also observed what was believed to be the influence of the viscosity of the molten salt medium on the electrochemical behavior of the conducting polymer; when the poly(aniline) electrode, in its insulating state, was placed in a basic or neutral melt, it required around 30 potentiodynamic cycles before the maximum electroactivity was obtained. This kinetic limitation is most likely related to the ability of the anions to permeate the film (solvent swelling), which can be expected to be significantly different in this, relatively viscous, molten salt media compared to molecular solvent systems. It is also dependent on the size of the anion present in the molten salt. This observation has subsequently been corroborated by researchers using air and water stable ionic liquids [64]. 7.3.2 Synthesis in Air- and Water-stable Ionic Liquids 7.3.2.1 Poly(pyrrole) Poly(pyrrole) is one of the most popular conducting polymers as it can be highly conducting, quite environmentally stable and relatively easy to synthesize. Sekiguchi

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

180 7 Conducting Polymers

et al. [46] have studied synthesis of this polymer in ionic liquids utilizing the [C2 mim] cation and the [BF4 ]− , [PF6 ]− and [OTf]− anions. Comparison of the growth CVs of poly(pyrrole) in these ionic liquids showed the highest polymer oxidation and reduction currents were obtained during growth in [C2 mim][OTf], and the lowest from using the [C2 mim][PF6 ]. The authors thus suggest that the higher viscosity of the former species results in a faster polymerization rate, because the polymerization reaction products are accumulated near the electrode surface and thus can undergo further radical coupling, oxidation and deposition onto the electrode rather than diffusing away into the bulk. However, this is not always seen and therefore other factors also influence these processes. The authors concluded that the [OTf]− anion was the superior choice for the electropolymerization of pyrrole. When comparing the performance of this neat ionic liquid to that of solutions of the ionic liquid in water or acetonitrile it was noted that the undiluted ionic liquid gave greater polymer redox peak currents and smaller redox peak separation. The film grown in the neat ionic liquid was reportedly thinner than that from the dilute solutions but more electrochemically active and more highly doped. The films grown from the aqueous and acetonitrile solutions displayed a granular morphology (larger from the water solution) whereas the film from the ionic liquid appeared to be very smooth. The origin of the cauliflower-like features observed in poly(pyrrole) from molecular solvent systems has been investigated by a number of researchers, who suggest that their appearance can be significantly altered depending on the nature of the anion used, the electrode material and polishing techniques used, the synthesis temperature and so on [4]. Sekiguchi et al. [65] also report the use of poly(pyrrole) films as a matrix for hosting palladium particles and show that these are considerably more dispersed when deposited from the ionic liquid than from aqueous solutions of the salt. Significantly smoother film morphologies have also been observed for poly(pyrrole) grown from the ionic liquids [C2 mim][NTf2 ] and [C4 mpyr] [NTf2 ] compared to those grown under the same experimental conditions from PC/Bu4 NPF6 (Figure 7.5) [51]. However, lower polymer redox currents were observed in the more viscous, less conductive pyrrolidinium ionic liquid compared to the imidazolium species (85 cP and 2.2 mS cm−1 , respectively, at 25 ◦ C for [C4 mpyr] [NTf2 ] [66] compared to 34 cP and 8.8 mS cm−1 at 20 ◦ C for [C2 mim][NTf2 ] [67]).

Fig. 7.5 Poly(pyrrole) films grown in [C4 mpyr][NTf2 ] (a), [C2 mim][NTf2 ] (b) and PC/Bu4 NPF6 (c), constant potential onto Pt.

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.3 Synthesis of Conducting Polymers 181

The electrochemical synthesis of poly(pyrrole) from [C4 mim][PF6 ] has also been studied in some detail [29, 32, 68]. Fenelon and Breslin used this ionic liquid to deposit poly(pyrrole) onto an iron electrode [29], an example of the electrochemical deposition of a conducting polymer from an ionic liquid onto a corrosionsusceptible electrode rather than the inert species used in the other studies. The polymer was deposited onto iron using a relatively high potential (1.3 V vs. a Ag wire quasi-reference electrode) but was electroactive and conducting, indicating negligible polymer overoxidation problems compared to those associated with using aqueous systems or even those observed using a platinum working electrode at these potentials in this ionic liquid. The authors determined a critical growth concentration of 0.2 mol dm−3 , above which the rate of electropolymerization does not significantly increase, and reported that the poly(pyrrole) layers grown onto this substrate were also smoother than those grown from aqueous systems. The poly(pyrrole)/iron electrode was stable for periods greater that 16 h in this ionic liquid, maintaining its low charge-transfer resistance and with no dissolution of iron through the pores on the polymer coating detected, as has been observed using aqueous systems. In this investigation the authors performed the polymerization in a dry nitrogen atmosphere, with an ionic liquid water content of ca. 10 ppm, below which it is reported that the polymerization rate decreases. Mazurkiewicz et al. [32] have also reported an influence of water content on the polymerization of pyrrole in this ionic liquid. They demonstrated that the redox response of the polymer growth CV was much more defined when the polymer was grown in [C4 mim][PF6 ] that had been purged with dry nitrogen (Figure 7.6(a)) compared to the response using ionic liquid that had been equilibrated in air (Figure 7.6(b)), and also showed the difference in growth CVs in [C4 mim][PF6 ] compared to that in deoxygenated 0.25 M Bu4 NPF6 in PC (Figure 7.6(c)) [32]. Figure 7.6 demonstrates the typical appearance of growth CVs of conducting polymers; the potential is repeatedly cycled in a positive and negative direction and at potentials above the oxidation potential of the monomer, polymer deposition onto the electrode occurs. The polymer oxidation and reduction peaks show an increase in current with successive cycles (the arrows show the direction of peak progression) indicating the deposition of increasing amounts of electroactive polymer. The peak position may also shift as the film becomes thicker, which may be attributed to factors such as heterogeneous electron-transfer kinetics or a decrease in conductivity, counter-ion mobility or conjugation length. A significant improvement in cycle life was also demonstrated for the poly(pyrrole) in the [C4 mim][PF6 ] (>900 cycles) compared to cycling in 0.25 M PC/Bu4 NPF6 (300 cycles) [32]. Finally, during the synthesis of poly(pyrrole) in [C2 mim][NTf2 ] an unusual mechanism of growth was observed, with the polymer growing along the surface of the ionic liquid [69]. When the working electrode is a thin platinum wire, and the reaction is performed in air but using nitrogen-purged ionic liquid, the polymer grows along the surface of the ionic liquid after forming an initial thin layer on the submerged body of the electrode (Figure 7.7(a)). We believe that the presence of some water is necessary for this “solution-surface electropolymerization”, to react

December 25, 2007 13:32

Fig. 7.6 Growth of poly(pyrrole) in (a) air-equilibrated [C4 mim][PF6 ], (b) [C4 mim][PF6 ] after N2(g) purging and (c) deoxygenated 0.25 M Bu4 NPF6 in PC, 100 mV s−1 , 30 cycles [32].

c07 (JWBG008-Endres) Char Count=

182 7 Conducting Polymers

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.3 Synthesis of Conducting Polymers 183

Fig. 7.7 Poly(pyrrole) grown along the surface of [C2 mim][NTf2 ]: (a) constant potential, (b) with a circular auxiliary electrode, (c) using voltage pulses [69].

with the H+ produced in the reaction (water being a stronger base than the [NTf2 ]− anion) and when a dry ionic liquid is used this is provided by absorption from the atmosphere. This phenomenon can be encouraged using an auxiliary electrode that circles the working electrode, to give directional growth (Figure 7.7(b)). Further, using a pulsed voltage the polymer forms first as a series of fibrils that can extend over a significant portion of the film. (Figure 7.7(c)). This fine structure imparts a larger surface area to the polymer than would be present in a solid, homogeneous film. 7.3.2.2 Thiophenes Shi et al. [70] were the first to demonstrate the use of an air and moisture stable ionic liquid, [C4 mim][PF6 ], for the electrochemical synthesis of poly(thiophene), grown onto a platinum working electrode by potentiodynamic, constant potential or constant current techniques. The use of growth potentials between 1.7 and 1.9 V (vs. Ag/AgCl) reportedly gave smooth, blue–green electroactive films, whereas potentials above 2 V resulted in film destruction by overoxidation. Sekiguchi et al. [65] used [C2 mim][OTf] for the polymerization of thiophene and demonstrated larger polymer redox currents during potentiodynamic growth in this ionic liquid than were observed in a 0.1 M solution of the ionic liquid in acetonitrile. Thus, as for poly(pyrrole), the authors concluded that this ionic liquid gave a higher growth rate, as well as smoother films and improved electrochemical capacity. The growth of poly(thiophene), poly(bithiophene) and poly(terthiophene) in [C2 mim][NTf2 ] and [C4 mpyr][NTf2 ] has also been studied [27]. The oxidation potential of these monomers decreases with increasing chain length, consistent with their behavior in conventional electrolyte/solvent systems. This is the primary advantage of using such materials – the high potential required to oxidise thiophene can result in side-reactions and overoxidation of the poly(thiophene) polymer film; in other words poly(thiophene) is not stable at the potentials required for its synthesis [71]. The use of dimers or oligomers of thiophene is one way of overcoming this “polythiophene paradox” [72], and increasing the stereoregularity of the polymer by reducing the number of β,β or α,β mis-linkages. Ideally the conjugation length

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

184 7 Conducting Polymers

Fig. 7.8 Cyclic voltammograms of thiophene polymerization (0.2 M, 50 mV s−1 ) onto a Pt working electrode: (a) growth and (b) post-growth in [C2 mim][NTf2 ], (c) growth and (d) post-growth in [C4 mpyr][NTf2 ], vs. a Ag pseudo-reference electrode. Arrows indicate the peak development with successive scans [27].

of the polymer would also be increased, although in reality the opposite may occur [73, 74]. The oxidation potential for all of the monomers appeared to be higher in the pyrrolidinium ionic liquid by approximately 0.1 V, which may be due to the higher viscosity and lower conductivity of this ionic liquid compared to the imidazolium species. There are also significant differences in the growth CVs of the polymers in the two different ionic liquids (Figure 7.8). For each monomer and ionic liquid, measurement of the total cathodic charge passed during reduction of the polymers in the final post-polymerization CVs, compared to the peak polymer oxidation currents from the final growth cycles, allows comparison of the film electrochemical activities while taking into account the relative amounts of the polymer. The former value is often used as an indication of the amount of polymer grown, but this assumes that the electrochemical activities of the films are identical.

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.3 Synthesis of Conducting Polymers 185

The peak oxidation current during the final growth cycle of poly(thiophene) is slightly higher in the imidazolium ionic liquid (Figure 7.8(a)), whereas the film from the pyrrolidinium species exhibits a larger total reduction charge in the postpolymerization CV, suggesting better electrochemical activity, possibly as a result of slower, more ordered film growth. Alternatively, this may indicate the superiority of the pyrrolidinium ionic liquid as a cycling solvent, but this is probably less likely given its high viscosity; differences in the post-polymerization CVs of such polymer films in the ionic liquids probably reflect both an influence of the nature of the ionic liquids on the polymer growth and the effect of using the different ionic liquids as the solvent for the post-polymerization cycling. Thus, for a direct assessment of the influence of the ionic liquid on polymer growth, post-polymerization cycling may be best performed in the same solvent, possibly a molecular solvent/electrolyte system. However, the proven benefits of using ionic liquids as the supporting media for the electrochemical cycling of conducting polymers, such as improved stability, suggest that assessment of post-polymerization cycling of the polymers in the ionic liquids is of more interest to researchers considering utilization of these materials in electrochemical devices. For the electropolymerization of bithiophene [27], which is adequately soluble in both ionic liquids, under the same conditions the growth CVs (Figure 7.9) suggest a stronger influence of the nature of the ionic liquid than was observed for the thiophene monomer, with the polymerization of bithiophene appearing to be four times faster in the [C2 mim][NTf2 ] than in the [C4 mpyr][NTf2 ], and the redox currents in the post-polymerization CVs also proportionally larger. Further, there is only one distinct reduction peak evident during growth and cycling of the film in the imidazolium ionic liquid but there are two distinct peaks evident during growth in the pyrrolidinium species, although in the post-polymerization cycles these are considerably broadened. Terthiophene is less soluble than thiophene or bithiophene, and is insoluble in most organic solvents, but concentrations of 0.05 M are attainable in both [C2 mim][NTf2 ] and [C4 mpyr][NTf2 ] with gentle (50 ◦ C) heating. This is an adequate concentration for the electrosynthesis of coherent poly(terthiophene) films onto either small Pt working electrodes, for electrochemical analysis, or onto ITO glass of a few square centimeters for analysis by UV–Vis and any photovoltaic testing required [75]. In the growth CVs of poly(terthiophene) in these ionic liquids (Figure 7.10) the influence of the nature of the ionic liquid on the growth rate of the polymer films is again evident, with significantly faster growth in the imidazolium species, and proportionally larger redox currents in the post-polymerization CVs (Figure 7.10 (b,d)). The films display multiple redox peaks during postpolymerization cycling in the ionic liquids, more clearly defined in the imidazolium ionic liquid. The poly(terthiophene) films exhibit good reversibility and the redox currents appear relatively stable over the 15 cycles recorded. Murray et al. [76] have demonstrated that an ionic liquid can be used as both the growth medium for poly(terthiophene) and also as a route to incorporation of anionic dyes into the polymer, for use in photovoltaic devices. Again, the solubility of terthiophene in [NTf2 ]-based ionic liquids is demonstrated; 1.6 × 10−2 M in

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

186 7 Conducting Polymers

Fig. 7.9 Cyclic voltammograms of bithiophene polymerization (0.1 M, 50 mV s−1 ): (a) growth and (b) post-growth in [C2 mim][NTf2 ], (c) growth and (d) post-growth in [C4 mpyr][NTf2 ], vs. a Ag pseudoreference electrode. Arrows indicate the peak development with successive scans [27].

[C4 mim][NTf2 ]. Electropolymerization of terthiophene, by potentiodynamic cycling with an ITO glass working electrode, from a solution of the ionic liquid containing terthiophene and the anionic dye Erioglaucine, resulted in the formation of a thick, mechanically strong polymer film with the dye incorporated as a dopant. The films produced were more robust than those obtained using dimethylformamide as a solvent, and these could then be reduced in an acetonitrile solution of a cationic dye (brilliant green) to yield polymer films containing both anionic and cationic dopants, resulting in a significantly improved photovoltaic performance. Naudin et al. [26] have studied the electrochemical synthesis of poly(3(4-fluorophenyl)thiophene) in two alternative [NTf2 ]− ionic liquids, utilizing the 1-ethyl-2,3-dimethylimidazolium and 1,3-diethyl-5-methylimidazolium cations. These ionic liquids have melting points of 20 and −22 ◦ C, respectively, and viscosities of 88 and 36 cP, respectively, at 20 ◦ C [67]. The authors report that the oxidation potential of the monomer is higher in these ionic liquids (1.16 and 1.22 V, respectively, for galvanostatic growth at 12.7 mA cm−2 ) compared to growth of this polymer in propylene carbonate or acetonitrile (0.98 and 1.1 V, respectively), which

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.3 Synthesis of Conducting Polymers 187

Fig. 7.10 Cyclic voltammograms of terthiophene polymerization (0.01 M, 50 mV s−1 ): (a) growth and (b) post-growth in [C2 mim][NTf2 ], (c) growth and (d) post-growth in [C4 mpyr][NTf2 ], vs. a Ag pseudoreference electrode [27].

they attribute to the lower stability of the monomer cation radical in the ionic liquids (a lower stability of the cation that is formed during the electrochemical polymerization would make its formation less energetically favorable and therefore require higher potentials). Electrochemical analysis of the films by cycling in the ionic liquids indicated slower redox processes than those observed for the films grown and cycled in 1 M acetonitrile/[Et4 N]BF4 , as evidenced by a larger separation between the anodic and cathodic peaks. This peak separation is only partly attributed to the lower ionic conductivity of the ionic liquids (3.2 and 6.6 mS cm−1 , respectively) compared to the acetonitrile solution (43 mS cm−1 ); it also reflects differences in the polymer films. The slower redox processes are also consistent with the smooth morphology of the films. The doping levels, determined electrochemically, appeared to be little influenced by the growth media and unfortunately the n-dedoping wave of this

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

188 7 Conducting Polymers

polymer overlaps with the negative limit of electrochemical stability of these ionic liquids (−2.1 V), which somewhat limits the redox switching. In general, the films exhibited poorer electrochemical activity in the ionic liquids, which was attributed to poorer swelling and slower ion transport kinetics, although the kinetics could be improved by growth of a thinner film. The doping of the polymers was also studied using X-ray photoelectron spectroscopy (see Section 7.4.3), which also indicated the presence of residual ionic liquid in the films that was hard to remove by washing. The electrochemical activity of the films was assessed by scanning at 100 mV s−1 over the complete p-doping and n-doping range for the polymer and a rapid decrease in activity was observed (75% after 50 cycles). The authors suggest that this may be a result of gradual loss of ionic liquid from the polymer (deswelling) during cycling. An alternative explanation is that ions are trapped in the polymer, as evidenced by a significant charge imbalance between the n-doping and n-dedoping charges in the CV. Interestingly, the authors note that a polymer film cycled in the n-doping region in the ionic liquid could be reactivated by cycling in the p-doping region in the same ionic liquid, or by transferal to the acetonitrile solution to re-swell the polymer. 7.3.2.3 Poly(3,4-ethylenedioxythiophene) Poly(3,4-ethylenedioxythiophene) (PEDOT) is a particularly popular conducting polymer as it can have good conductivity and stability and has a low band gap, which is pertinent to its use in photovoltaic devices. A number of authors have now studied the electrochemical synthesis of this polymer in different ionic liquids. Lu et al. [77] first demonstrated the use of [C4 mim][BF4 ] to electrodeposit PEDOT onto ITO, and its application in an electrochromic numeric display. Randriamahazaka et al. [64, 78, 79] have studied the synthesis and behavior of PEDOT in [C2 mim][NTf2 ] in detail. In their primary report, the authors electrodeposited a PEDOT film from an acetonitrile/LiClO4 solution and studied its electrochemical behavior when cycled in the ionic liquid [79]. In their subsequent paper [64], they reported the electrochemical response of PEDOT that was grown in the ionic liquid, and cycled in the ionic liquid that also contained lithium bis(trifluoromethanesulfonyl)amide (LiNTf2 ), and contrasted this with the behavior of a PEDOT film prepared in acetonitrile. PEDOT grown from acetonitrile and cycled in the ionic liquid displayed two oxidation and reduction peaks, the less anodic of which decreased in peak potential but increased in current upon increasing the concentration of LiNTf2 in the ionic liquid. In contrast, PEDOT that was prepared in the ionic liquid itself displayed only one anodic and one cathodic peak (scan rate 100–500 mV s−1 ), at the same position as the second oxidation and reduction peak that was observed in the CV of PEDOT grown from acetonitrile, and the presence of LiNTf2 in the ionic liquid had little effect on the electrochemical behavior of the film. In both cases it was concluded that the imidazolium cation of the ionic liquid was the primary intercalating/de-intercalating species. It is also interesting to note that when the PEDOT film from acetonitrile was cycled in the ionic liquid, the authors observed a continuous change in the shape of the CV and increases in the redox current (up to 20 cycles), which was attributed to the uptake of the

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.3 Synthesis of Conducting Polymers 189

ionic liquid into the film. This effect has also been observed during the cycling of PEDOT in the pyrrolidinium analogue [80]. Damlin et al. [81] have reported the synthesis and p-doping and n-doping of PEDOT in [C4 mim][BF4 ] and [C4 mim][PF6 ] and characterized the resultant films by CV, in situ UV–Vis spectroelectrochemistry and ATR-FTIR (see also Section 7.4). Here, two oxidation peaks were observed in the first few growth cycles in the [C4 mim][BF4 ] (at 50 mV s−1 ), which merged into one as the film became thicker, and two reduction peaks were also seen. In the [C4 mim][PF6 ] three oxidation peaks were observed at first, again merging into one with successive cycles, thus indicating an influence of the anion. In this ionic liquid, two reduction peaks are again evident. The authors report that the shapes of the CVs, and the oxidation potential of the monomer, are similar in the ionic liquids to those in organic solvents using Bu4 N PF6 or Et4 N BF4 electrolytes. The synthesis of PEDOT in [C2 mim][NTf2 ] and [C4 mpyr][NTf2 ] has also been studied, and the multiple redox peaks observed were influenced by the choice of ionic liquid [80]. The current increase during potentiodynamic growth of the film in the pyrrolidinium species was less than in the imidazolium analogue, suggesting a slower film growth due to the higher viscosity and lower conductivity of this medium that limits ion/molecule transport kinetics. This is as observed for growth of poly(pyrrole) and poly(thiophene)s. Post-polymerization CVs of these films were recorded in both acetonitrile/Bu4 NClO4 and the ionic liquid, and compared to those of PEDOT grown from an acetonitrile solution. For both films grown from ionic liquid there was an increase in the electrochemical activity upon cycling in the acetonitrile solution, suggesting better swelling of the polymer and thus faster transport of ionic species into and out of the polymer during cycling. Thus, the observed activity reflects the electrochemical accessibility of the polymer to the electrolyte, which may suggest that the electrochemistry of the polymer is a surface-dominated phenomenon. On return to the [C4 mpyr][NTf2 ] there was a rapid return to the lower charge capacity regime, which was ascribed to structural changes. However, there was a progressive increase in current with cycling in the [C4 mpyr][NTf2 ], as observed by Randriamahazaka et al. [64] and Wagner et al. [80] for the cycling of PEDOT in [C2 mim][NTf2 ] after growth or cycling in acetonitrile, which is likely to be linked to the slow dissolution of entrained acetonitrile out of the film and/or the slow uptake of ionic liquid into the film. In agreement with other authors [79], no memory effect upon cycling the films in these different solvents was observed. The growth and post-polymerization CVs of PEDOT from [C2 mim][NTf2 ] and from an acetonitrile solution are shown in Figure 7.11. There is a decrease in the electrochemical activity of the film grown in the acetonitrile solution when cycled in the ionic liquid, whereas the activity of the film grown in the ionic liquid was less affected by the nature of the cycling solvent. Comparison of the chronoamperograms recorded during EDOT electropolymerization in the two different ionic liquids and two conventional acetonitrile-based electrolytes allows some conclusions to be drawn about the mechanism of polymer deposition of PEDOT from these different media (Figure 7.12) [80]. The current transients suggest that the process is initially much slower in the solution

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

190 7 Conducting Polymers

Fig. 7.11 Cyclic voltammograms of PEDOT (0.1 M, 20 cycles, every third shown, 100 mV s−1 ): (a) Film I growth in acetonitrile/[Bu4 NClO4 ], (b) post- growth of films I and II in acetonitrile/[Bu4 NClO4 ], (c) Film II growth in [C2 mim][NTf2 ] and (d) post-growth of films I and II in [C2 mim][NTf2 ], vs. a Ag pseudo-reference electrode [80].

containing Bu4 NClO4 as the electrolyte than in the other cases. Moreover, the different shape of the curve suggests a different mechanism of deposition; the current transient in acetonitrile/Bu4 NClO4 is indicative of progressive nucleation, with a slower growth rate and thus lower currents, whereas the current transients in the ionic liquids and the acetonitrile/LiNTf2 solution (Figure 7.12(b–d)) suggest

Fig. 7.12 Current–time responses to potential step from 0 to 1.4 V for the electropolymerization of 0.1 M EDOT onto ITO electrodes in different media: (a) 0.1 M Bu4 NClO4 in acetonitrile, (b) [C4 mpyr][NTf2 ], (c) 0.1 M LiNTf2 in acetonitrile and (d) [C2 mim][NTf2 ] [80].

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.4 Characterization

instantaneous nucleation, thus indicating a strong influence of the anion on polymer growth. The spectroelectrochemistry of the films from these solutions was also studied – see Section 7.4.3. Danielsson et al. [25] have studied the synthesis of PEDOT in ionic liquids that utilize bulky organic anions, 1-butyl-3-methylimidazolium diethylene glycol monomethyl ether sulfate and 1-butyl-3-methylimidazolium octyl sulfate, the latter of which is a solid at room temperature and thus requires the addition of either monomer or solvent (in this case water) to form a liquid at room temperature. Polymerization in a water-free ionic liquid was only possible in the octyl sulfate species, but the polymerization of EDOT was successful in aqueous solutions of both the ionic liquids (0.1 M). The ionic liquid anions appear to be mobile within the polymer, exchangeable with chloride ions at a polymer/KCl(aq) interface, but it is interesting that when the PEDOT is in aqueous solutions of the ionic liquid, at higher concentrations (0.01–0.1 M) the imidazolium cation can suppress this anion response. The ion mobility in both the ionic liquid and in the polymer film in contact with the solution is significantly increased by addition of water. 7.3.2.4 Poly(para-phenylene) Endres et al. [82] have demonstrated the suitability of an air- and water-stable ionic liquid for the electropolymerization of benzene. This synthesis is normally restricted to media such as concentrated sulfuric acid, liquid SO2 or liquid HF as the solution must be completely anhydrous. The ionic liquid used, 1-hexyl-3methylimidazolium tris(pentafluoroethyl)trifluorophosphate, can be dried to below 3 ppm water, and this ionic liquid is also exceptionally stable, particularly in the anodic regime. Using this ionic liquid, poly(para-phenylene) was successfully deposited onto platinum as a coherent, electroactive film. Electrochemical quartz crystal microbalance techniques were also used to study the deposition and redox behavior of the polymer from this ionic liquid (Section 7.4.1) [83].

7.4 Characterization

It is particularly important to fully characterize the nature of the materials produced when different growth media and methodologies are being investigated, and this is perhaps one area of conducting polymer research that has generally lacked sufficient attention. This is partly due to the insolubility of these materials, but there is still a wide range of analytical techniques suitable, and a number of these have been utilized for the analysis of conducting polymers synthesized in ionic liquids. 7.4.1 Electrochemical Characterization

Cyclic voltammetry of the conducting polymers allows assessment of the electrochemical activity of the films and the redox processes involved. The electrochemical

191

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

192 7 Conducting Polymers

response of the polymers can be strongly dependent on the solvent/electrolyte system used as the cycling medium. This is related to the extent to which the solvent swells the polymer, which affects the size of response and capacitance measured. Thus, post-polymerization CVs of the polymers reflect not only the properties of the polymer film and any influence of the growth solvent used, but are also affected by the nature of the solvent used for the post-polymerization cycling. It has been noted that the post-polymerization CVs of polymers grown in ionic liquids can be significantly different depending on whether they are cycled either in the ionic liquid or in a molecular solvent/electrolyte system, or that the films may take time to equilibrate on transferal to a new medium [64, 80]. This effect has primarily been attributed to the different degrees of polymer swelling in the different media. Electrochemical analysis can provide valuable information relating to the structure of the polymer. For example, Randriamahazaka et al. [64, 78, 79] have studied the behavior of PEDOT in [C2 mim][NTf2 ], using a PEDOT film electrodeposited from an acetonitrile/LiClO4 solution and cycled in the ionic liquid [79]. Using potential step experiments, the authors showed that the redox switching dynamics of the polymer consisted of two different processes, with different kinetics. Both a fast and a slow process were identified for the polymer oxidation and reduction, and the authors note that this is consistent with the proposal that the structure of the PEDOT films is composed of two coexisting zones, one rigid and compact zone containing polymer chains that are long and highly conjugated, and one zone with a more open polymer configuration with chains of a shorter conjugation length [84], as also proposed by other authors [85, 86]. The presence of these two zones is a suggested explanation for the presence of the two oxidation and reduction peaks that are observed in the cyclic voltammogram of this polymer, with the lower potential peak arising from oxidation of the compact, highly conjugated polymer and the more anodic peak ascribed to oxidation of the more open polymer with shorter conjugation. There is extensive and somewhat inconclusive discussion in the literature regarding the possible origin of the multiple redox peaks sometimes observed for poly(thiophene) species. It has been proposed that these peaks are due to the transitions between the neutral, polaron, bipolaron and metallic states of the polymer [87], which may also be influenced by the rate of counterion transport [88], the effect of “charge-trapping” [84], conformational changes accompanying radical cation formation [89], consideration of the mechanical strain on the polymer that results from the forced intrusion of anions into the film [71], as well as the aforementioned reduction of different areas of the polymer film or of polymer chains of significantly different lengths [86, 90, 91]. We have observed multiple redox peaks in a number of conducting polymer films synthesized in ionic liquids and an additional possible explanation for this, when an ionic liquid is used as the supporting electrolyte, is the presence of additional redox processes involving the intercalation and expulsion of the ionic liquid cation. For further electrochemical characterization of the conducting polymer films, the charge passed during the electrochemical synthesis can also be used to estimate the film thickness and level of doping, and thus the capacitance of the

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.4 Characterization

film if required. This may be more accurately achieved using an electrochemical quartz crystal microbalance (EQCM), to study both polymer growth and the intercalation/expulsion of anions/cations from the film during cycling. In this technique, the polymer is deposited onto a metal-coated quartz crystal (as the working electrode, typically gold- or platinum-coated) whose oscillating resonance frequency changes according to the mass of the polymer. Thus, both growth of the polymer and its redox cycling (with the associated mass changes due to ion incorporation/expulsion) may be studied. Endres et al. [83] have demonstrated the applicability of this technique for the analysis of a conducting polymer in ionic liquids, through the synthesis and cycling of poly(p-phenylene) in 1-hexyl3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate. It should be noted that when using this technique in combination with a viscous medium such as an ionic liquid, the damping effect that the ionic liquid has on the frequency of the crystal must be considered [83]. However, if this damping does not change significantly during the growth and cycling of the polymer, whereas the resonance frequency does (by at least one order of magnitude more than the change in damping) then under these conditions it is valid to convert the frequency change to changes in the mass of the polymer using the Sauerbrey equation. Using this technique, Schneider et al. [83] showed that, as a result of the large size of the anion used, cycling of the polymer in the ionic liquid involved a significant amount of cation exchange, which was particularly prominent at higher scan rates. At low scan rates there is sufficient time for movement of the large anion and this dominates, especially in the cathodic peak region, whereas in faster scans anion removal or insertion into the polymer film becomes more difficult. Outside the cathodic peak region, at low scan rates anion and cation movement are approximately equal but, as the scan rate is increased, movement of the ionic liquid cation becomes the dominant process over the whole potential range. Cyclic voltammetry is also an ideal analytical tool for assessing the electrochemical stability of the polymer films. This is a fundamental requirement for any conducting polymer to be considered for long-term use in electrochemical devices. The use of ionic liquids for the electrochemical cycling of poly(aniline) has been reported to enhance lifetimes to over a million cycles [12], and significant improvements in the cycling stability of poly(pyrrole) have also been reported [32]. Electrochemical impedance spectroscopy is a powerful tool for elucidating the diffusion and conduction parameters of the film, particularly when used in combination with computer modeling. Danielsson et al. [25] utilized this technique to study parameters such as solution resistance, charge transfer resistance, double layer capacitance and bulk redox capacitance for PEDOT grown in ionic liquids composed of imidazolium cations and bulky anions, and assessed how these were influenced by dilution of the ionic liquid to different concentrations in water. Naudin et al. [26] also used this technique for the analysis of poly(3-(4fluorophenyl)thiophene) films grown in [1-ethyl-2,3-dimethylimidazolium][NTf2 ] and [1,3-diethyl-5-methylimidazolium][NTf2 ]. In both these studies, the polymers in the neat ionic liquids displayed slower ion transport compared to those in molecular solvents or diluted ionic liquids. Fenelon and Breslin [29] also used

193

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

194 7 Conducting Polymers

electrochemical impedance measurements to demonstrate the conductivity and stability of poly(pyrrole) deposited onto iron from [C4 mim][PF6 ]. The switching potential of the polymer – the potential at which there is a transition between conducting and insulating states – can be significantly influenced by the nature of the solvent and electrolyte used for growth and cycling in molecular solvents [4], and thus this will also be influenced by the use of ionic liquids, and the nature of the cations and anions used. Boxall and Oseryoung [68] determined the switching potentials of poly(pyrrole) and poly(N-methylpyrrole) from [C4 mim][PF6 ], using rotating ring-disk voltammetry, to be 0.63 ± 0.04 and 1.07 ± 0.03 V, respectively, vs. the cobaltocenium/cobaltocene couple. This technique was also used to study the potential-dependant conductivity of the polymers. The DC conductivity of the polymer films is clearly an important property of conducting polymer films and this can be accurately measured using a four-point probe conductivity apparatus, or may be extrapolated from EIS measurements. There are few reports in the literature documenting the conductivity of polymers synthesized in ionic liquids. Sekiguchi et al. [65] measured the conductivities of poly(pyrrole) and poly(thiophene) synthesized in [C2 mim][OTf], and compared this to polymers synthesized in a dilute solution of this ionic liquid in acetonitrile or water. The conductivities were measured using a two-probe method, with the polymers collected as powders by scraping off the electrode. This technique gives relatively low values but nonetheless showed that the conductivities of the polymers from the neat ionic liquids were significantly higher than those from the acetonitrile or water solutions. Consistent with this, the doping level of the poly(pyrrole) from the ionic liquid was significantly higher. In our preliminary investigations into the synthesis of poly(pyrrole) from [NTf2 ]-based ionic liquids, without any optimization of the synthesis technique, we have measured conductivities up to ca. 100 S cm−1 . 7.4.2 Morphological Characterization

Very striking differences between films grown in conventional solvents and those grown in ionic liquids have been observed using scanning electron microscopy (SEM) [51, 65, 80, 92, 93]. Generally, the films grown from ionic liquids appear to be considerably smoother, which may also result in improved conductivities. SEM analysis of poly(thiophene) grown from [C2 mim][NTf2 ] and [C4 mpyr][NTf2 ] reveals a slightly smoother morphology for the poly(thiophene) films from the pyrrolidinium ionic liquid [27]. The influence of the nature of the ionic liquid on the film morphology is consistent for poly(thiophene), poly(pyrrole), poly (bithiophene) and poly(terthiophene), although the difference is less marked in the latter two. The polythiophene film grown in the [C2 mim][NTf2 ] (Figure 7.13) displays a ‘packed grain’ structure commonly observed in polythiophene films, and is very similar to that reported by Sekiguchi et al. [65], who noted that the grain size was smaller than that obtained when acetonitrile was used as the solvent. This

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.4 Characterization

Fig. 7.13 SEMs of poly(thiophene) films grown from [C2 mim][NTf2 ] (a)–(c) and [C4 mpyr][NTf2 ] (d)–(f), viewed from above (a), (b), (d) and (e) and edge-view (c) and (f) [27].

grain morphology is typical of 3D nucleation and growth [94]. The film grown from the [C4 mpyr][NTf2 ] (Figure 7.13) was slightly smoother, suggesting a more ordered film, which is consistent with slightly slower film growth (Figure 7.8). The edge-on views of the poly(thiophene) films (Figure 7.13) also show a dense film on the ITO electrode below the granular polymer, clearly suggesting that a different nucleation and growth mechanism operates at the onset of film growth, as has been proposed by Schrebler et al. [95]. Thus, initial film growth is consistent with an instantaneous 2D mechanism, involving soluble oligomer growth at the ITO electrode and subsequent deposition at some critical chain length, to form a compact film. This is then followed by progressive 3D nucleation and growth giving the granular morphology, which is probably a result of the formation of a more branched poly(thiophene). This is not unique to growth of polythiophene in ionic liquids – similar behavior is reported in other solvents and electrolytes [95–97] and for poly(pyrrole) grown by solution-surface electropolymerization from this ionic liquid [69]. Poly(bithiophene) films from these two ionic liquids are morphologically similar (Figure 7.14), even though the redox behavior (Figure 7.9) is markedly different, suggesting that the dominant differences in the films produced are on an atomic or sub-micron rather than macroscopic level. The morphology of the poly(bithiophene) films appears to be similar to that described by Roncali et al. [74] who reported a thin film on the surface of the electrode, covered by a thick brittle powdery deposit, from the galvanostatic polymerization of bithiophene in acetonitrile. The nodular structures are smaller in the poly(bithiophene) films than in the poly(thiophene), which is consistent with the formation of shorter chain polymers [73], but this does not

195

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

196 7 Conducting Polymers

Fig. 7.14 SEM images of poly(bithiophene) ((a) and (b), 10 :m, (c) and (d) 2 :m edge-view) and poly(terthiophene) ((e) and (f), 10 :m) grown from [C2 mim][NTf2 ] (a), (c), (e) and [C4 mpyr][NTf2 ] (b), (d), (f).

appear to result in inferior electrochemical activity (Figure 7.9). The edge-on views (Figure 7.14) again shed some light on the polymer growth mechanism. In contrast to the polythiophene films, the initial growth layer of both poly(bithiophene) films is granular, suggestive of a 3D nucleation and growth mechanism, and this appears to be followed by a second granular 3D nucleation and growth phase [96], reflecting an influence of the starting monomer. Poly(terthiophene) films from these ionic liquids are very smooth, with a spongy morphology at the micron level (Figure 7.14(e, f)) similar to that described by Sezai

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.4 Characterization

Fig. 7.15 SEM images of reduced PEDOT films on ITO, prepared by cyclic voltammetry (20 cycles, 100 mV s−1 ) from (a) 0.1 M LiNTf2 in acetonitrile, (b) [C4 mpyr][NTf2 ], (c) [C2 mim][NTf2 ] and (d) 0.1 M Bu4 NClO4 in acetonitrile [80].

Sarac et al. [98] for poly(terthiophene) grown in acetonitrile, but without the large amount of powdery deposit observed in the poly(bithiophene) films (Figure 7.14(a, b)). Again, the film from the pyrrolidinium ionic liquid appears to be slightly more compact. SEM analysis of PEDOT films grown from these ionic liquids compared to films grown from acetonitrile/ Bu4 NClO4 (Figure 7.15) [80] suggests that the effect of changing the dopant anion can be a more significant influence than the use of the ionic liquids as the film from acetonitrile/ Bu4 NClO4 is markedly different from those grown in either the ionic liquids or the acetonitrile/LiNTf2 . Reduction of a conducting polymer, with the simultaneous expulsion of anions, is generally expected to result in films becoming more compact, and this can be studied by SEM. For these PEDOT films, the morphology of the polymer grown in acetonitrile/ Bu4 NClO4 became more compact upon reduction, but this effect was not clearly observed in the other films [99].

197

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

198 7 Conducting Polymers

7.4.3 Spectroscopic Characterization

Raman and FTIR Spectroscopy are important techniques for analysing the level of doping of polymer films, both across the entire surface and by depth profiling using a cross-section of film [100]. Raman spectroscopy measures the vibrational energies of molecules by irradiating the sample with monochromatic laser light and measuring the scattered radiation, which differs in frequency from the incident light by an amount determined by the Stokes shift of the molecule. Raman spectra of conducting polymers can provide information about the identity of the polymer (for example Figure 7.16), the degree of oxidation of the polymer, and the identification and quantification of any Raman-active dopants within the film. Similar but complementary information can be obtained using infrared spectroscopy. In situ Fourier transform infrared spectroscopy (FTIR) with attenuated total reflection (ATR) allows changes in the structural and electronic properties of polymers in the near-IR region to be studied during electrochemical reactions. Damlin et al. [81] demonstrated the use of this technique to study the p-doping and n-doping of PEDOT synthesized and cycled in [C4 mim][BF4 ] and [C4 mim][PF6 ]. Few conducting polymers can be both p-doped (by oxidation) and n-doped (by reduction). The poor stability of the reduced form of PEDOT in organic electrolytes under atmospheric conditions has not allowed the investigation of the n-doping of this polymer in organic solvents. However, the PEDOT films were successfully n-doped (−0.9 to −1.95 V) and p-doped (−0.9 to +0.8 V) in the ionic liquids, as evidenced by the doping-induced infrared active vibrations. In the tetrafluoroborate ionic liquid the current response of the n-doping cycles appeared to stabilize after about five scans and remained stable for the next nine cycles. At a low degree of doping (< 0.1 V) the changes in the doping induced infrared active vibrations of the polymers were the same, irrespective of growth solvent, whereas during p-doping at more positive potentials the polymers from the ionic liquids showed further increases in the intensities of these bands that were not observed in the polymer from

Fig. 7.16 Examples of Raman spectra of (a) poly(pyrrole) and (b) poly(terthiophene), from [C2 mim][NTf2 ].

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.4 Characterization

acetonitrile. The in situ FTIR-ATR spectra of the polymer from the ionic liquids were similar to those from 0.1 M Bu4 NClO4 in acetonitrile, with only slight shifts in peak position. However, the relative intensities of the bands of the n-doped and p-doped polymer are higher in the polymer from the [C4 mim][BF4 ], suggesting that it is easier to create charge carriers in this film compared with those grown from the [C4 mim][PF6 ] or the acetonitrile solution [81]. The p-doping and n-doping of conducting polymers, and the resultant incorporation of cations/anion, can also be studied by X-ray photoelectron spectroscopy (XPS). This is a relatively surface-specific technique, where the photoelectrons from the solid (which have quite a short range) give rise to spectra with peaks of binding energies specific to the elements present. For each element, the binding energies are influenced by chemical bonding or oxidation state (allowing, for example, the nitrogen of the poly(pyrrole) to be identified separately from the negative nitrogen of the [NTf2 ]− anion, as shown in Figure 7.17), thus deconvolution of the peaks allows structural analysis of the polymer, and the relative intensities of the peaks allows complete compositional analysis of the sample. However, for analysis of conducting polymers grown in ionic liquids the spectra may be complicated by the presence of elements common to both polymer and growth medium. For example, poly(pyrrole) grown in an imidazolium NTf2 ionic liquid will give peaks in the N 1s core level spectra from the imidazolium cation

Fig. 7.17 Example of the deconvolution of peaks from XPS analysis of poly(pyrrole) containing the [NTf2 ] anion and the [C2 mim] cation.

199

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

200 7 Conducting Polymers

and the anion as well as the polymer itself. Further, there may be residual ionic liquid on the polymer surface not removed by washing, as observed by Naudin et al. [26], in addition to ionic liquid cations and anions that are actually intercalated into the polymer film. However, some of this complication may be avoided by judicious choice of ionic liquid, such as the use of the [BF4 ]− or [PF6 ]− anion or a phosphonium cation. Naudin et al. [26] demonstrated the use of XPS for analysis of poly(3-(4-fluorophenyl)thiophene) grown and cycled in ionic liquids and its use for determining the nature and quantity of the dopant species in the p-doped and n-doped polymer. UV–Vis spectroscopy is a cheap and readily available technique that is ideal for studying conducting polymer films at each stage of growth and reduction/oxidation (when deposited onto ITO glass). As a result of electronic transitions between the fundamental levels and polaronic or bipolaronic levels, which involve photons with energy in the visible region of the spectrum, the UV–Vis spectra of conducting polymer films are significantly altered on oxidation (films change color). The wavelength maxima position changes as a function of the degree of oxidation/reduction, hence this is a valuable analytical technique for probing the electronic structure of the films. Damlin et al. [81] utilized this technique to study the p-doping and n-doping of PEDOT synthesized and cycled in [C4 mim][BF4 ] and [C4 mim][PF6 ]. They reported that the n-doping of PEDOT in the ionic liquids has little effect on the UV–Vis spectrum, which can also be the case when organic solvents are used, whereas the p-doping results in significant changes, suggesting the existence of different types of charge carriers in the different doping regimes (polarons vs. bipolarons). The UV–Vis spectra also suggested little effect on the average conjugation length of the polymers from using the ionic liquid, which is indicated by the position of the π–π * transition that occurs around 560 nm. Fenelon and Breslin [29] demonstrated the use of spectroelectrochemistry to monitor the growth of poly(pyrrole) from [C4 mim][PF6 ] onto an iron working electrode by plotting the natural logarithm of the absorbance of the 450 nm transition (attributed to the electronic transition from the valence to the antipolaron band of the polymer) against time. The authors thereby demonstrated that the growth of the poly(pyrrole) was a two-step process, starting with a faster nucleation and growth step (for approximately 8 min) followed by a steady growth phase (approximately 30 min), both of which obeyed first-order kinetics. Using this technique a difficulty in fully reducing conducting polymer films in ionic liquids has been observed [80], as indicated by the presence of a significant free carrier electron band in the spectrum at high wavelengths, which may be due to poor solvent swelling or a result of cation incorporation rather than anion expulsion during reduction (Figure 7.18). Spectroelectrochemistry (Figure 7.18(c)) of the PEDOT film electrodeposited from [C4 mpyr][NTf2 ] also shows incomplete reduction, and the reduced spectra do not change, even with reduction potentials down to −1.6 V. The appearance of the shoulder at around 900 nm, in both Figure 7.18 (b) and (c), is consistent with an incompletely reduced material [101].

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.4 Characterization

Fig. 7.18 UV–Vis spectra of PEDOT films (a) deposited in [C2 mim][NTf2 ], oxidized 20 min at 0.8 V and reduced 20 min at −1 V in the ionic liquid, (b) grown in [C2 mim][NTf2 ], oxidized 20 min at 0.8 V and

reduced 20 min at −1 V in acetonitrile/0.1 M Bu4 NClO4 solution, (c) spectroelectrochemistry of PEDOT in 0.1 M LiNTf2 in acetonitrile after deposition from [C4 mpyr][NTf2 ] from −1 to 0.9 V [80].

7.4.4 Solid-state NMR

One of the primary hindrances to improving understanding of the influence of the ionic liquid on conducting polymers, and studying the incorporation of ions from these media into polymers, is the difficulty in analyzing such insoluble materials. Understanding the ion incorporation processes is paramount to the development of these systems in, for example, actuator devices. The use of nuclear magnetic resonance (NMR) spectroscopy to study the nature of the polymers produced is a valuable but under-utilized technique. This can not only provide information on the polymer backbone but can also allow the study of intercalated ions within the film [102–104]. Judicious choice of ionic liquid cation and anion can allow them to be detected independently within the polymer film. For example, utilization of an ionic liquid comprised of a phosphonium cation, to allow its detection by 31 P NMR, and the NTf2 anion, for study by 19 F NMR spectroscopy [105]. The intercalation of

201

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

202 7 Conducting Polymers

the bulky phosphonium cation is potentially of interest for use in actuator devices. In addition, the polymer backbone can be studied by 13 C NMR [103, 104]. Poly(pyrrole) films grown by constant potential in the ionic liquid tri(hexyl) (tetradecyl) phosphonium bis(trifluoromethanesulfonyl)amide, [P6,6,6,14 ][NTf2 ], and subsequently cycled 100 times and reduced in the monomer-free ionic liquid, showed the presence of both ionic liquid cations and anions within the film. Interestingly, poly(pyrrole) films grown by constant potential in the ionic liquid but with no subsequent electrochemical cycling also contained both ionic liquid cation and anion within the film (Figure 7.19). The 31 P NMR (Figure 7.19(a)) indicates incorporation of the phosphonium cation into the film. The dominant signal appears at −12 ppm, with a minor peak at 32 ppm. This latter peak is assigned to the phosphonium cation of residual ionic liquid on the surface of the film, which was not removed by washing but is not actually intercalated into the film. This shift is consistent with that reported by Bradaric et al. [106] and with our own analysis of the neat ionic liquid. The large peak at −12 ppm is assigned to phosphonium cations intercalated into the polypyrrole film, and this shift is consistent with a P–N interaction. The dramatic change in the chemical shift of the phosphorus compared to the neat ionic liquid indicates a significant change in environment on incorporation into the polymer film.

Fig. 7.19 (a) 31 P (b) 19 F and (c) 13 C NMR spectra of polypyrrole grown at constant potential in [P6,6,6,14 ][NTf2 ].

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.5 Future Directions 203

However, the DC conductivity of the pyrrole film was sufficiently high to indicate that the polymer chain was intact and this precludes the possibility of any chemical reaction between the cation and the polymer. The successful expulsion of cations from the film (see below) also eliminates this possibility. The 19 F NMR spectrum (Figure 7.19(b)) shows the presence of intercalated anions within the film, as indicated by the peak at −82 ppm that is consistent with [NTf2 ]− . The 13 C NMR spectrum of the film (Figure 7.19(c)) shows a broad resonance from the poly(pyrrole) centered at ca. 120 ppm, and also significant intensity between 0 and 30 ppm from the alkyl chains of phosphonium cations within the film [106]. NMR analysis of a poly(pyrrole) grown by constant potential from a 0.1 M solution of the [P6,6,6,14 ][NTf2 ] in acetonitrile also showed the presence of both the ionic liquid cation and anion within the film. However, although the phosphonium cations appear to be easily incorporated into the poly(pyrrole) during synthesis, initial results suggest that they are not easily incorporated during cycling. Poly(pyrrole) films grown at constant potential from a 0.1 M solution of LiNTf2 in acetonitrile and then cycled 100 times in the neat [P6,6,6,14 ][NTf2 ] contained no detectable phosphonium cations. This may be related to a lack of solvent swelling of the film in the ionic liquid, which restricts ion movement. This technique can also be used to investigate the ease of expulsion of anions from the polymer. It was initially thought that given the large size of the phosphonium cations they would remain incorporated in the poly(pyrrole), as is observed for bulky anions such as polyelectrolyte dopants, but films grown in the ionic liquid and then oxidized overnight in the ionic liquid showed no detectable phosphonium cations. However, poly(pyrrole) oxidized for only 1 h still contained significant amounts of phosphonium cations, thus it takes considerable time for the films to fully oxidize through incorporation of NTf2 anions and expulsion of the phosphonium cations (the current becomes negligible after about 5 h). However, poly(pyrrole) films grown in the ionic liquid then oxidized in a 0.1 M solution of LiNTf2 in acetonitrile for 4 h resulted in expulsion of all of the phosphonium cations from the film (the current was reduced to a background level in less than 1 h). This was attributed to the effect of solvent swelling of the polymer, which enables increased ion movement in and out of the film. This may also explain the observed lower electrochemical activity of the films in the ionic liquid compared with their activity when cycled in molecular solvent systems. It is important to note, however, that the importance of solvent swelling would be significantly reduced for thinner polymer films.

7.5 Future Directions 7.5.1 Chiral Ionic Liquids

There is significant interest in the formation of chiral conducting polymers, such as chiral poly(aniline), as a result of their potential applications in chiral sensors,

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

204 7 Conducting Polymers

for chiral separations and so on. One route to these materials is by incorporating a chiral dopant anion during the electrochemical polymerization [107]. Thus, if the polymer was synthesized in a chiral ionic liquid, of which there are a growing number [108], then formation of an optically active conducting polymer might result. 7.5.2 Protic Ionic Liquids

One family of ionic liquids that has to date only been sparsely investigated for use with conducting polymers is the protic ionic liquids, where the cation has one or more mobile hydrogen atoms. A recent manuscript by Bic¸ak detailed the synthesis of 2-hydroxy ethylammonium formate [109], which melts at −82 ◦ C and has a room-temperature ionic conductivity of 3.3 mS cm−1 , and reported the ability of this protic ionic liquid to dissolve poly(aniline) (17 g mL−1 ) and poly(pyrrole) (no concentration specified). The dissolution of conducting polymers into any solvent is of significant interest for a variety of reasons, such as improving their processability and ease of incorporation into different devices. Li et al. [93] have used 1-ethylimidazolium trifluoroacetate, which is a Brønsted acidic ionic liquid, as a medium for the electropolymerization of aniline. They report that in this ionic liquid the oxidation potential of aniline is lower (0.58 V compared to 0.83 V in 0.5 M H2 SO4 ) and that the growth rate of the polymer is increased. Further, the resultant films are smooth, strongly adhered to the Pt working electrode and are very electrochemically stable. Similar results have been reported by Liu et al. [92], who found that this was the best ionic liquid for the polymerization of aniline, compared to the unsatisfactory results observed in other protic ionic liquids 1-butylimidazolium tetrafluoroborate, 1-butylimidazolium nitrate and 1-butylimidazolium p-toluenesulfonate, as well as the 1-butyl-3-methylimidazolium hydrogen sulfate and 1-butyl-3-methyimidazolium dihydrogen phosphate. However, this family of ionic liquids holds an additional attraction, which is the potential to create “distillable ionic liquids.” Earle et al. [110] have recently reported that a range of ionic liquids that are commonly perceived as nonvolatile, including various imidazolium NTf2 − salts, can actually be distilled at low pressure and temperatures of 200–300 ◦ C without significant decomposition. In such a process the ionic liquids are transferred into the gas phase as ionic species. However, in the case of protic ionic liquids, hydrogen transfer between the cation and anion can allow distillation of the neutral components and then subsequent recombination to reform the ionic liquid as the distillate [111]. The primary amine reported by Bicak [112], 2-hydroxy ethylammonium formate, decomposes on heating to give a formamide, but the combination of the formate anion with tertiary amine cations such as N-methylpyrrolidinium can give an ionic liquid that is distillable at modest temperatures and pressures [111]. The use of distillable ionic liquids for the synthesis and use of conducting polymers may provide additional advantages in terms of easy removal of the ionic liquid and isolation of any oligomeric species from the growth solutions. Further, as this area of ionic liquid research is relatively new,

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.5 Future Directions 205

there is a considerable range of protic ionic liquids as yet undiscovered or still to be investigated for their use for the growth and synthesis of different conducting polymers, and the different acidities of these ionic liquids may be of particularly interest for the synthesis of poly(aniline). 7.5.3 Nano-dimensional Polymers

The increasing interest in the use of conducting polymers [2] is fuelling a continued need for materials with improved physical and chemical properties. In particular, there is a recent drive towards nanostructured and reduced dimensionality materials, such as thin films, nanotubes, wires, particles and so on, which can exhibit markedly different properties from those of the bulk materials [113, 114]. There is already a small number of reports of the use of ionic liquids for the electrochemical synthesis of nanostructured conducting polymers and research in this area is predicted to increase significantly in the coming years. Koo et al. [31] have reported the polymerization of pyrrole using a nanoporous aluminum oxide template in [C4 mim][BF4 ], which yields poly(pyrrole) nanotubes and nanowires. Poly(aniline) nanotubules have been made by electrochemical polymerization onto an ITO glass electrode from [C4 mim][PF6 ] containing 1 M trifluoroacetic acid [115]. 7.5.4 Chemical Polymerization

The choice of synthetic route for the production of conducting polymers, either through electrochemical or chemical oxidation of the monomer or, in rare cases, photopolymerization or enzyme-catalysed polymerization, is primarily dictated by the final application of the polymer. While electrochemical polymerization is widely used for the controlled synthesis of polymer films, chemical polymerization results in the formation of powders or colloidal dispersions, is much more amenable to scale-up and can be used as a means to coat nonconducting substrates with conducting polymers. Although outside the scope of this discussion, it is interesting to note that despite the increased interest in the use of ionic liquids for the electrochemical synthesis of conducting polymers, there is a significant dearth of investigations into the potential benefits of using ionic liquids for the synthesis of these materials via alternative routes. The good electrochemical stability and solubilising properties of ionic liquids should allow access to a plethora of monomers and chemical oxidants at significant concentrations, some of which are either insoluble in, or outside the electrochemical stability of, molecular solvents. Gao et al. [116] have reported the chemical synthesis of poly(aniline) at a water/ionic liquid interface, in which the polymer grows into the water phase as nanoparticles, and Bic¸ak et al. [109] have also reported the use of their aforementioned protic ionic liquid, 2-hydroxyethyl ammonium formate, for the chemical synthesis of organo-soluble poly(aniline). An initial investigation into the use of [C2 mim][NTf2 ] for the chemical synthesis of poly(pyrrole), poly(terthiophene) and nano-dimensional poly(thiophene) has also

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

206 7 Conducting Polymers

been reported [117]. However, the range of different monomers, oxidants and ionic liquids available, as well as different experimental techniques that may be employed, for example with the ionic liquid as one phase or in a biphasic system in combination with a second solvent, suggests that this may be a particularly fruitful area of future investigation for conducting polymer researchers.

7.5.5 Remaining Challenges

The potential improvements that ionic liquids may impart to conducting polymers have been widely discussed – increased doping levels, smoother films, increased conductivity, decreased over-oxidation and improved electrochemical stability and so on. However, the research to date in this area has only just begun to investigate these hypotheses and demonstrate any material advantages in the use of ionic liquids; future directions in this area must focus on some of these issues in addition to simply demonstrating the use of new ionic liquids for conducting polymer synthesis. The influence of the ionic liquid on the polymerization process itself is yet to be clarified. For example, does the ionic nature of these media stabilize the radicals/ cations that are formed during the polymerization and, if so, how does this impact on the mechanism of polymer growth, the resultant chain length, conjugation length and so on? The solubility of the oligomers that are formed during the polymerization in the ionic liquid, and the extent to which they diffuse away from the electrode (which may be less in viscous ionic liquids) influences the length at which they precipitate onto the electrode. This will impact on polymer properties such as conjugation length, morphology, conductivity and so on, and therefore warrants some investigation. The solubility of the oligomers formed during the polymerization is also of interest not just from a mechanistic point of view but because the solubilization of conducting polymers, even if they are relatively short chained, is desirable for a number of applications. It would also be interesting to measure the degree of solvent swelling of the polymer films in ionic liquids compared to molecular solvents, as this will impact significantly on the ion mobility and thus the electrochemical activity of the polymer. It is expected that this will be lower in ionic liquids compared to molecular solvents – although the extent of this difference may also depend on the thickness of the film – but this is yet to be quantified. The extent of the ionic liquid cation and anion intercalation into the film during growth and cycling, and the structural features of the ions that influence this, requires more investigation. It has been postulated that the higher ion concentration of ionic liquids compared to a traditional electrolyte/molecular solvent system will result in a higher polymer doping level, but this has not been extensively demonstrated. Anion and cation intercalation not only influences the properties of the polymer but is also important for understanding and developing conducting polymer–ionic liquid actuator devices.

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

7.6 Conclusions 207

The smoother film morphology imparted by use of ionic liquids has been attributed to slower film growth in this medium but film growth rate has not been quantified or compared to growth in molecular solvent systems. EQCM would be an ideal technique for such a study. The high viscosity of ionic liquids could, if desired, be reduced by using higher temperatures, mixtures of ionic liquids or ionic liquids diluted with either a molecular solvent or the monomer itself (if liquid), and this will, in turn, influence the film morphology. The addition of a second salt component, such as Bu4 N PF6 , to the ionic liquid may also be beneficial for film growth.

7.6 Conclusions

It would be most interesting at this point to be able to compile a list of the benefits of using ionic liquids for the synthesis and use of conducting polymers, and weigh these against the disadvantages of these new media. However, such a clear idea of the pros and cons of using ionic liquids is still some time away. The benefits of using ionic liquids as the supporting electrolyte in conducting polymer devices has been clearly demonstrated and predominantly concerns extended lifetimes and the potential to reduce problems such as solvent evaporation that are associated with the use of molecular solvents in these devices. However, at this point even this research is predominantly at a lab-based scale rather than a larger or commercial scale. Concerns such as ionic liquid cost, toxicity and large-scale availability will no doubt come to the fore as ionic liquid-containing conducting polymer devices are developed on a larger scale, whereas it may take longer for benefits such as improved lifetimes or performance to be fully realized. When the ionic liquid is used as the growth medium for these materials, financial concerns may be minimized by efficient recycling after use, and toxicity concerns will be confined to their use in the production process rather than their widespread use in devices for public use, for example in solar cells, batteries or sensors. Some properties of ionic liquids, such as their higher viscosity and lower conductivity compared to molecular solvents, may be a disadvantage for larger-scale conducting polymer synthesis, but the potential benefits of using ionic liquids lie not in the production process but in their ability to improve the conducting polymer itself, for example by improved conductivity, electrochemical activity, morphology and so on. This will, in turn, give better device performance. Thus, it is paramount that the efficacy of ionic liquids as a route to conducting polymers with improved material properties is demonstrated. At this point, the application of ionic liquids for the electrosynthesis of conducting polymers has been demonstrated by a number of authors and some differences between this and the use of molecular solvents reported. In a number of these cases an improvement in properties was reported (most extensively a smoother polymer morphology and increased cycle life) but the full range of benefits of using ionic liquids is yet to be fully realized or amply demonstrated. There is clearly

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

208 7 Conducting Polymers

massive scope for future investigation, utilizing the vast range of ionic liquids presently available either commercially or through reported synthetic routes. Thus, while the field of ionic liquids is immense, their potential use for the synthesis of the different types of conducting polymers is even more extensive. A number of common ionic liquid anions have already been proven to be beneficial dopants for conducting polymers but there are also an extensive number that are still to be tested and maybe one of these new ionic liquids will prove to be the key to synthesizing conducting polymers with all of the physical and electrochemical properties that we desire. References 1 Sandman, D.J. (1998) Handbook of Conducting Polymers, Mol. Cryst. Liq. Cryst. Sci. Technol, Section A: Mol. Cryst. Liq. Cryst., 325, 260. 2 Chandrasekhar, Prasanna (1999) Conducting Polymers, Fundamentals and Applications : a Practical Approach, Kluwer Academic, Boston, London. 3 Sonmez, G. (2005) Chem. Commun., 5251. 4 Wallace, G.G., Spinks, G.M., Kane-Maguire, L.A.P., and Teasdale, P.R. (2003) Conductive Electroactive Polymers: Intelligent Materials Systems, 2nd edn, CRC Press, Boca Raton, FL. 5 Wallace, G.G. and Kane-Maguire, L.A.P. (2002) Adv. Mater., 14, 953. 6 Shirakawa, H., Louis, E.J., MacDiarmid, A.G., Chiang, C.K., and Heeger, A.J. (1977) J. Chem. Soc., Chem. Commun., 578. 7 Chiang, C.K., Druy, M.A., Gau, S.C., Heeger, A.J., Louis, E.J., MacDiarmid, A.G., Park, Y.W., and Shirakawa, H. (1978) J. Am. Chem. Soc., 100, 1013. 8 Cho, M.S., Seo, H.J., Nam, J.D., Choi, H.R., Koo, J.C., and Lee, Y. (2006) Proc. SPIE-Int. Soc. Opt. Eng., 6168, 61682E–61682E/7. 9 Ding, J., Zhou, D., Spinks, G., Wallace, G., Forsyth, S., Forsyth, M., and MacFarlane, D. (2003) Chem. Mater., 15, 2392. 10 Lu, W., Norris, I.D., and Mattes, B.R. (2005) Aust. J. Chem., 58, 263. 11 Lu, W. and Mattes, B.R. (2005) Synth. Met., 152, 53. 12 Lu, W., Fadeev, A.G., Qi, B., Smela, E., Mattes, B.R., Ding, J., Spinks, G.M.,

13

14

15 16 17

18

19

20

21

22 23

24

Mazurkiewicz, J., Zhou, D., Wallace, G.G., MacFarlane, D.R., Forsyth, S.A., and Forsyth, M. (2002) Science, 297, 983. Zhou, D., Spinks, G.M., Wallace, G.G., Tiyapiboonchaiya, C., MacFarlane, D.R., Forsyth, M., and Sun, J. (2003) Electrochim. Acta, 48, 2355. Vidal, F., Plesse, C., Randriamahazaka, H., Teyssie, D., and Chevrot, C. (2006) Mol. Cryst. Liq. Cryst., 448, 697. Vidal, F., Plesse, C., Teyssie, D., and Chevrot, C. (2004) Synth. Met., 142, 287. Akle, B.J., Bennett, M.D., and Leo, D.J. (2006) Sens. Actuators, A, 126, 173. Fukushima, T., Asaka, K., Kosaka, A., and Aida, T. (2005) Angew. Chem., Int. Ed., 44, 2410. Stenger-Smith, J.D., Webber, C.K., Anderson, N., Chafin, A.P., Zong, K., and Reynolds, J.R. (2002) J. Electrochem. Soc., 149, A973–A977. Balducci, A., Soavi, F., and Mastragostino, M. (2006) Appl. Phys. A, 82, 627. Balducci, A., Bardi, U., Caporali, S., Mastragostino, M., and Soavi, F. (2004) Electrochem. Commun., 6, 566. Lu, W., Fadeev, A.G., Qi, B., and Mattes, B.R. (2004) J. Electrochem. Soc., 151, H33–H39. Yang, C., Sun, Q., Qiao, J., and Li, Y. (2003) J. Phys. Chem. B, 107, 12981. John, R. and Wallace, G.G. (1991) J. Electroanal. Chem. Interfacial Electrochem., 306, 157. Ogasawara, M., Funahashi, K., and Iwata, K. (1985) Mol. Cryst. Liq. Cryst., 118, 159.

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

References 25 Danielsson, P., Bobacka, J., and Ivaska, A. (2004) J. Solid State Electrochem., 8, 809. 26 Naudin, E., Ho, H.A., Branchaud, S., Breau, L., and Belanger, D. (2002) J. Phys. Chem. B, 106, 10585. 27 Pringle, J.M., Forsyth, M., MacFarlane, D.R., Wagner, K., Hall, S.B., and Officer, D.L. (2005) Polymer, 46, 2047. 28 Snook, G.A., Best, A.S., Pandolfo, A.G., and Hollenkamp, A.F. (2006) Electrochem. Commun., 8, 1405. 29 Fenelon, A.M. and Breslin, C.B. (2005) J. Electrochem. Soc., 152, D6–D11. 30 Yoon, C.O., Sung, H.K., Kim, J.H., Barsoukov, E., Kim, J.H., and Lee, H. (1999) Synth. Met., 99, 201. 31 Koo, Y.K., Kim, B.H., Park, D.H., and Joo, J. (2004) Mol. Cryst. Liq. Cryst., 425, 55. 32 Mazurkiewicz, J.H., Innis, P.C., Wallace, G.G., MacFarlane, D.R., and Forsyth, M. (2003) Synth. Met., 135–136, 31. 33 Mermilliod, N., and Tanguy, J. (1986) J. Electrochem. Soc., 133, 1073. 34 Huddleston, J.G., Visser, A.E., Reichert, W.M., Willauer, H.D., Broker, G.A., and Rogers, R.D. (2001) Green Chem., 3, 156. 35 Seddon, K.R., Stark, A., and Torres, M.-J. (2000) Pure Appl. Chem., 72, 2275. 36 Jacquemin, J., Husson, P., Padua, A.A.H., and Majer, V. (2006) Green Chem., 8, 172. 37 Favre, C., Abello, L., and Delabouglise, D. (1997) Adv. Mater., 9, 722. 38 Nogami, Y., Pouget, J.-P., and Ishiguro, T. (1994) Synth. Met., 62, 257. 39 Lee, K., Miller, E.K., Aleshin, A.L., Menon, R., Heeger, A.J., Kim, J.H., Yoon, C.O., and Lee, H. (1998) Adv. Mater., 10, 456. 40 Kiani, M.S. and Mitchell, G.R. (1992) Synth. Met., 48, 203. 41 Zhao, H., Price, W.E., and Wallace, G.G. (1994) J. Membrane Sci., 87, 47. 42 Garcia, B. and Belanger, D. (1998) Synth. Met., 98, 135. 43 Villareal, I., Morales, E., Otero, T.F., and Acosta, J.L. (2000) Synth. Met., 108, 57. 44 Ye, F., Noftle, R.E., and DesMarteau, D.D. (1993) Synth. Met., 60, 141.

45 Ciprelli, J.-L., Clarisse, C., and Delabouglise, D. (1995) Synth. Met., 74, 217. 46 Sekiguchi, K., Atobe, M., and Fuchigami, T. (2002) Electrochem. Commun., 4, 881. 47 Zawodzinski, T.A. Jr. and Osteryoung, R.A. (1988) Inorg. Chem., 27, 4383. 48 Pickup, P.G. and Osteryoung, R.A. (1984) J. Am. Chem. Soc., 106, 2294. 49 Pickup, P.G. and Osteryoung, R.A. (1985) J. Electroanal. Chem. Interfacial Electrochem., 195, 271. 50 Zawodzinski, T.A. Jr., Janiszewska, L., and Osteryoung, R.A. (1988) J. Electroanal. Chem. Interfacial Electrochem., 255, 111. 51 Pringle, J.M., Efthimiadis, J., Howlett, P.C., Efthimiadis, J., MacFarlane, D.R., Chaplin, A.B., Hall, S.B., Officer, D.L., Wallace, G.G., and Forsyth, M. (2004) Polymer, 45, 1447. 52 Geetha, S. and Trivedi, D.C. (2004) Mater. Chem. Phys., 88, 388. 53 Geetha, S. and Trivedi, D.C. (2005) Synth. Met., 148, 187. 54 Trivedi, D.C. (1989) J. Chem. Soc., Chem. Commun., 544. 55 Koch, V.R., Miller, L.L., and Osteryoung, R.A. (1976) J. Am. Chem. Soc., 98, 5277. 56 Goldenberg, L.M. and Osteryoung, R.A. (1994) Synth. Met., 64, 63. 57 Kobryanskii, V.M. and Arnautov, S.A. (1992) Makromol. Chem., 193, 455. 58 Goldenberg, L.M., Pelekh, A.E., Krinichnyi, V.I., Roshchupkina, O.S., Zueva, A.F., Lyubovskaya, R.N., and Efimov, O.N. (1990) Synth. Met., 36, 217. 59 Lere-Porte, J.P., Radi, M., Chorro, C., Petrissans, J., Sauvajol, J.L., Gonbeau, D., Pfister-Guillouzo, G., Louarn, G., and Lefrant, S. (1993) Synth. Met., 59, 141. 60 Janiszewska, L. and Osteryoung, R.A. (1987) J. Electrochem. Soc., 134, 2787. 61 Janiszewska, L. and Osteryoung, R.A. (1988) J. Electrochem. Soc., 135, 116. 62 Tang, J. and Osteryoung, R.A. (1991) Synth. Met., 45, 1. 63 Tang, J. and Osteryoung, R.A. (1991) Synth. Met., 44, 307. 64 Randriamahazaka, H., Plesse, C., Teyssie, D., and Chevrot, C. (2004) Electrochem. Commun., 6, 299.

209

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

210 7 Conducting Polymers

65 Sekiguchi, K., Atobe, M., and Fuchigami, T. (2003) J. Electroanal. Chem., 557, 1. 66 MacFarlane, D.R., Meakin, P., Sun, J., Amini, N., and Forsyth, M. (1999) J. Phys. Chem. B, 103, 4164. 67 Bonhote, P., Dias, A.-P., Papageorgiou, N., Kalyanasundaram, K., and Graetzel, M. (1996) Inorg. Chem., 35, 1168. 68 Boxall, D.L. and Osteryoung, R.A. (2004) J. Electrochem. Soc., 151, E41–E45. 69 Pringle, J.M., Forsyth, M., Wallace, G.G., and MacFarlane, D.R. (2006) Macromolecules, 39, 7193. 70 Shi, W., Ge, D., Wang, J., Jiang, Z., Ren, L., and Zhang, Q. (2006) Macromol. Rapid Commun., 27, 926. 71 Marque, P. and Roncali, J. (1990) J. Phys. Chem., 94, 8614. 72 Krische, B. and Zagorska, M. (1989) Synth. Met., 28, C263–C268. 73 Roncali, J. (1992) Chem. Rev., 92, 711. 74 Roncali, J., Garnier, F., Lemaire, M., and Garreau, R. (1986) Synth. Met., 15, 323. 75 Chen, J., Officer, D.L., Pringle, J.M., MacFarlane, D.R., Too, C.O., and Wallace, G.G. (2005) Electrochem. Solid-State Lett., 8, A528–A530. 76 Murray, P.S., Ralph, S.F., Too, C.O., and Wallace, G.G. (2006) Electrochim. Acta, 51, 2471. 77 Lu, W., Fadeev, A.G., Qi, B., and Mattes, B.R. (2003) Synth. Met., 135–136, 139. 78 Randriamahazaka, H., Plesse, C., Teyssie, D., and Chevrot, C. (2005) Electrochim. Acta, 50, 4222. 79 Randriamahazaka, H., Plesse, C., Teyssie, D., and Chevrot, C. (2003) Electrochem. Commun., 5, 613. 80 Wagner, K., Pringle, J.M., Hall, S.B., Forsyth, M., MacFarlane, D.R., and Officer, D.L. (2005) Synth. Met., 153, 257. 81 Damlin, P., Kvarnstrom, C., and Ivaska, A. (2004) J. Electroanal. Chem., 570, 113. 82 Zein El Abedin, S., Borissenko, N., and Endres, F. (2004) Electrochem. Commun., 6, 422. 83 Schneider, O., Bund, A., Ispas, A., Borissenko, N., El Abedin, S.Z., and Endres, F. (2005) J. Phys. Chem. B, 109, 7159. 84 Zotti, G., Schiavon, G., and Zecchin, S. (1995) Synth. Met., 72, 275.

85 Wallace, G.G. and Innis, P.C. (2002) J. Nanosci. Nanotechnol., 2, 441. 86 Skompska, M. (1998) Electrochim. Acta, 44, 357. 87 Chen, X. and Inganaes, O. (1996) J. Phys. Chem., 100, 15202. 88 Zotti, G. and Schiavon, G. (1989) Synth. Met., 31, 347. 89 Domagala, W., Lapkowski, M., Guillerez, S., and Bidan, G. (2003) Electrochim. Acta, 48, 2379. 90 Sato, M., Tanaka, S., and Kaeriyama, K. (1987) Makromol. Chem., 188, 1763. 91 Hillman, A.R., Efimov, I., and Skompska, M. (2002) Faraday Discuss., 121, 423. 92 Liu, B.-Y., Xu, D.-Q., and Xu, Z.-Y. (2005) Chin. J. Chem., 23, 803. 93 Li, M.C., Ma, C.A., Liu, B.Y., and Jin, Z.M. (2005) Electrochem. Commun., 7, 209. 94 Li, F.B. and Albery, W.J. (1992) Langmuir, 8, 1645. 95 Schrebler, R., Grez, P., Cury, P., Veas, C., Merino, M., Gomez, H., Cordova, R., and del Valle, M.A. (1997) J. Electroanal. Chem., 430, 77. 96 del del Valle, M.A., Ugalde, L., Diaz, F.R., Bodini, M.E., Bernede, J.C., and Chaillou, A. (2003) Polym. Bull., 51, 55. 97 Ugalde, L., Bernede, J.C., Del Valle, M.A., Diaz, F.R., and Leray, P. (2002) J. Appl. Polym. Sci., 84, 1799. 98 Sezai Sarac, A., Evans, U., Serantoni, M., Clohessy, J., and Cunnane, V.J. (2004) Surf. Coatings Technol., 182, 7. 99 Wagner, K., Pringle, J.M., Hall, S.B., Forsyth, M., MacFarlane, D.R., and Officer, D.L. (2005) Synth. Met., 153, 257. 100 Furukawa, Y. (1996) J. Phys. Chem., 100, 15644. 101 Groenendaal, L.B., Zotti, G., Aubert, P.-H., Waybright, S.M., and Reynolds, J.R. (2003) Adv. Mater., 15, 855. 102 Meng, H., Perepichka, D.F., Bendikov, M., Wudl, F., Pan, G.Z., Yu, W., Dong, W., and Brown, S. (2003) J. Am. Chem. Soc., 125, 15151. 103 Forsyth, M., Truong, Van Tan, and Smith, M.E. (1994) Polymer, 35, 1593. 104 Forsyth, M. and Smith, M.E. (1993) Synth. Met., 55, 714.

c07 (JWBG008-Endres)

December 25, 2007

13:32

Char Count=

References 105 Pringle, J.M., MacFarlane, D.R., and Forsyth, M. (2005) Synth. Met., 155, 684. 106 Bradaric, C.J., Downard, A., Kennedy, C., Robertson, A.J., and Zhou, Y. (2003) Green Chem., 5, 143. 107 Pornputtkul, Y., Kane-Maguire, L.A.P., and Wallace, G.G. (2006) Macromolecules, 39, 5604. 108 Baudequin, C., Bregeon, D., Levillain, J., Guillen, F., Plaquevent, J.-C., and Gaumont, A.-C. (2005) Tetrahedron: Asym., 16, 3921. 109 Bicak, N., Senkal, B.F., and Sezer, E. (2005) Synth. Met., 155, 105. 110 Earle, M.J., Esperanca, J.M.S.S., Gilea, M.A., Canongia Lopes, J.N., Rebelo, L.P.N., Magee, J.W., Seddon, K.R., and Widegren, J.A. (2006) Nature, 439, 831. 111 MacFarlane, D.R., Pringle, J.M., Johansson, K.M., Forsyth, S.A., and

112 113

114 115

116

117

Forsyth, M. (2006) Chem. Commun., 1905. Bicak, N. (2004) J. Mol. Liq., 116, 15. Leclere, Ph., Surin, M., Viville, P., Lazzaroni, R., Kilbinger, A.F.M., Henze, O., Feast, W.J., Cavallini, M., Biscarini, F., Schenning, A.P.H.J., and Meijer, E.W. (2004) Chem. Mater., 16, 4452. Aleshin, A.N. (2006) Adv. Mater., 18, 17. Wei, D., Kvarnstroem, C., Lindfors, T., and Ivaska, A. (2006) Electrochem. Commun., 8, 1563. Gao, H., Jiang, T., Han, B., Wang, Y., Du, J., Liu, Z., and Zhang, J. (2004) Polymer, 45, 3017. Pringle, J.M., Ngamna, O., Chen, J., Wallace, G.G., Forsyth, M., and MacFarlane, D.R. (2006) Synth. Met., 156, 979.

211

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

213

8 Nanostructured Metals and Alloys Deposited from Ionic Liquids Rolf Hempelmann, and Harald Natter

8.1 Introduction

Nanomaterials are composed of structural entities – isotropic grains or particles, rods, wires, platelets, layers – of size, at least in one dimension, between 1 and 100 nm [1–3]. Larger particles are called submicron particles, smaller ones are known as clusters. Some physical properties of nanomaterials [4–6] differ from those of coarse-grained materials of the same chemical composition due to two essential features: 1. Large specific surface area and concomitantly large specific surface energy; hence surface sensitive properties (like catalytic activity) are enhanced [7–10] and processes where the surface energy is the driving force (sintering, grain growth) are facilitated [11–13]. 2. Quantum size effects; famous examples are the color shift upon size reduction of semiconductor nanoparticles like CdSe [14–16], and the surface plasmon resonance of metallic nanoparticles like gold [17, 18].

The term nanomaterial includes materials consisting of or containing individual nanoparticles, nanorods, nanowires or nanoplatelets (for instance as composites), materials in the form of thin layers or coatings, and compact polycrystalline materials with grain sizes below 100 nm; the latter contain an appreciable volume fraction of interphase or grain boundary regions (analogously to the surface region of nanoparticles) and correspondingly a large specific interphase energy, such that these bulk nanomaterials are said to be dominated by their interphases [19]. Magnetic and mechanical properties are examples of properties which, even for coarse-grained materials, depend on grain boundaries and other lattice defects, in spite of their tiny volume fraction [20–22]. In bulk nanomaterials, with their large volume fraction of grain boundaries and, particularly with respect to the above-mentioned properties, pronounced nanoeffects occur which is the reason for the academic and industrial interest in bulk nanostructured materials. Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

214 8 Nanostructured Metals and Alloys Deposited from Ionic Liquids

Often the starting point of the route to bulk nanomaterials is a powder of nanoparticles which can be prepared either “top-down” by milling or “bottom-up” by controlled chemical synthesis [23]. This powder has to be compacted/densified in order to get bulk nanomaterial. An example of such a route is the famous inert gas condensation (IGC) introduced by Gleiter and Birringer [24]: a metal, for instance Pd, is evaporated under He gas, in the gas atmosphere condensation starts with the formation of clusters and small nanoparticles; these are deposited on a cold finger (77 K), scraped off, collected, and uniaxially cold-pressed into the form of a tablet. Isostatic hot-pressing of metal oxide nanoparticles is a route to nanoceramics [25–28]. In all these routes the densification step is crucial. Efficient densification (sintering) involves diffusion processes and requires elevated temperatures. However, at elevated temperatures grain growth takes place and after a short time the sample is no longer nanocrystalline. At lower temperatures the sample remains nanocrystalline but the densification is incomplete; the resulting porosity is the main disadvantage of all routes to bulk nanomaterials which start from powders. This critical compaction step is avoided in the case of the electrochemical route of pulsed electrodeposition (PED) [29] which transforms cations, i.e. atomic species, directly into nanomaterials without the detour via nanoparticles. In this way densities up to 99% of the theoretical value can be achieved, such that these materials exhibit, for instance, intrinsic mechanical properties and not those dominated by voids. Furthermore, this technique allows variation of the grain size [30–33]; this is important because many chemical and physical properties of nanostructured materials depend on the grain size. Only by variation of the crystallite size – this is a novel aspect in materials science and technology [34] – is it possible to tune and hopefully improve certain physical properties of one and the same material: for example, the enhanced hardness of nano-Au, the toughness of nano-Ni/P alloys [35], the soft magnetic properties of nano-Ni [36] and the resistance of nanostructured materials [37, 38] promise industrial applications [39–41]. The production of such “tailor-made” nanomaterials by electrochemical procedures is advantageous because the two crucial steps in nanocrystal formation – nucleation and growth of nuclei – can be controlled by physical (current, voltage, time, temperature) and chemical (grain refiners, complex formers) parameters during the deposition process [42, 43]. In Section 8.2 the basics of pulsed electrodeposition (PED) will be described for the case of aqueous electrolytes which allow the deposition of comparatively noble metals like Cu, Ni, Pd, or Au; less noble metals like Fe or Zn can still be electrodeposited from aqueous electrolytes because they exhibit a comparatively large overpotential for hydrogen evolution. However, the main limitation of aqueous electrolytes, of course, is their narrow electrochemical window which adversely affects the electrodeposition of metals like Al or Ta. Therefore, recently, the PED technique has been extended to ionic liquids as electrolytes. General electrochemical aspects of ionic liquids can be found in Ref. [44]; here, in Section 8.3, we will only address the technical aspects with respect to PED. Examples of nanometals and nanoalloys electrodeposited from chloroaluminate-based ionic liquids are given in

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

8.2 Pulsed Electrodeposition from Aqueous Electrolytes 215

Section 8.4. The main disadvantage of these electrolytes is their extreme sensitivity to moisture [45]; so-called ionic liquids of the third generation are water stable [46], and examples of electrodeposition from these water-stable ionic liquids are presented in Section 8.5. Finally, a short summary and outlook for possible future developments is given in Section 8.6.

8.2 Pulsed Electrodeposition from Aqueous Electrolytes

The pulsed electrodeposition technique (PED) is a versatile method for the preparation of nanostructured metals and alloys [47]. In the last two decades PED has received much attention worldwide because it allows the preparation of large bulk samples with high purity, low porosity and enhanced thermal stability. 8.2.1 Fundamental Aspects

The electrochemical deposition of nanostructured metals and alloys is a two-step process: 1. The formation of a large number of nuclei 2. The controlled growth of the deposited nuclei.

These two conditions can be realized by the proper choice of the chemical and physical process parameters. The size and the number of nuclei can be controlled by the overvoltage (η): r =

2σ V z e 0 |η|

(8.1)

In this electrochemical version of the Kelvin equation [48] r is the critical nucleation radius, σ the specific surface energy, V the atomic volume in the crystal and z the number of elementary charges e0 . The message of Eq. (8.1) is: the higher the overvoltage the smaller the formed nuclei. A large overvoltage brings about a large current density and thus a large rate of formation of nuclei. So PED is a deposition process at large overpotential, a rare type of study in electrocrystallization because, in most cases, underpotential deposition is studied [49–54]. This high overpotential and the concomitant high deposition rate, however, can be maintained only for a few milliseconds (ton -time) because the metal ion concentration in the vicinity of the cathode decreases drastically and the process threatens to become diffusioncontrolled. In order to avoid this and avoid losing control over the process via Eq. (8.1), the current has to be switched off, one has to wait for 20–100 ms (toff -time) in order that the metal ions can diffuse from the bulk electrolyte to the cathode and

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

216 8 Nanostructured Metals and Alloys Deposited from Ionic Liquids

compensate the metal ion depletion there. Then the next current and voltage pulse is applied, and so on: this is the reason for the pulsing. In the electroless break between two pulses exchange currents flow, an undesired electrochemical phenomenon because it induces the so-called Ostwald ripening [55]: the larger crystallites (with the lower surface energy) grow at the expense of the smaller ones (with the higher surface energy), and the crystallite size distribution broadens and shifts to larger size values, i.e., deteriorates. Another origin of Ostwald ripening could be surface diffusion of metal adatoms. In view of this deposition mechanism several measures have to be taken to control the crystallite size: 1. The voltage/current density in the peak has to be large; a reasonable value is 1 A m−2 in the more common galvanostatic mode. 2. The ton -time has to be as short as is necessary to avoid the diffusion control regime and as long as possible in order to reach sufficient deposition efficiency; a reasonable value is 3 ms. 3. The toff -time has to be as long as necessary for the material transport in the electrolyte to take place, but not longer in order to minimize Ostwald ripening and reach sufficient deposition efficiency. 4. The use of organic additives (grain refiners): they have to adsorb on the fresh nuclei, with a Gibbs enthalpy of adsorption just sufficient to suppress the exchange current in the toff -time but insufficient to impede the metal ion deposition in the ton -time, i.e., in the next pulse. This enables one to control the crystallization process during the toff -time because these molecules are adsorbed reversibly on the electrode surface and hinder the surface diffusion of the adatoms. For different metals different additives are most suited, but a typical and widespread example is thiourea. The use of additives is common practice in the galvanic industry, the selection of additive is a matter of experience, i.e., purely empirical [41]. 5. Temperature influences all diffusion processes (cation diffusion in the electrolyte, surface diffusion of the adatoms). If small crystallite sizes are desired the deposition should be performed at ambient temperatures or below in order to slow down the kinetics of recrystallisation of the nuclei.

The bath composition, the pH-value, the hydrodynamic conditions and also the use of special current pulse shapes are further possibilities to influence the deposition process. It is advantageous to perform the pulsed electrodeposition in the galvanostatic mode because the average deposition rate can be simply derived from Iav =

Ipulse ton ton + toff

(8.2)

In addition, there is better control of the current efficiency and the alloy composition. The potentiostatic mode would be, on the basis of Eq. (8.1), more desirable but is experimentally more difficult to realize because a three-electrode set-up is

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

8.2 Pulsed Electrodeposition from Aqueous Electrolytes 217

necessary and, upon the voltage jump, the current should start theoretically from an infinite value, which is not feasible due to electronic limitations. Also, in the potentiostatic mode a short reverse pulse would be necessary to fix the initial potential, which is not generally desirable because passivation may occur during this reverse potential pulse. Suitable substrates for the electrodeposition are stainless steel or titanium electrodes (20 × 20 mm, distance between the electrodes 25 mm); because of the poor adhesion the resulting deposits can be removed mechanically from the electrode. Tuning the grain size must be accompanied by measuring the grain size. The most convenient method is the line width or line shape analysis of X-ray diffraction peaks according to Scherrer [56] (estimate of the grain size), or according to the Williamson and Hall procedure [57] which allows one to separate grain size and microstrain effects on the line width. Fourier transform techniques like the Warren–Averbach technique [58–60] or certain full profile fit routines [61] even allow determination of the crystallite size distribution. Details can be found in Ref. [62]. 8.2.2 Nanometal Deposition with Nano-Gold as an Example

For n-gold, deposited from a commercial gold(I)sulfite bath, the influence of the deposition parameters on the nanostructure has been demonstrated [63]. For a better comparison of the different results the experiments were performed at an average current density (Iav ) of 3 mA cm−2 , see Eq. (8.2). If ton and toff are kept at constant values (for details see Figure 8.1), the smallest crystallite size of 12 nm is found for Ipulse = 0.5 A cm−2 (see Figure 8.1); for higher current values powder

Fig. 8.1 The crystallite size dependence of the pulsed current density for gold deposits deposited from a commercial sulfite bath without any additives [29].

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

218 8 Nanostructured Metals and Alloys Deposited from Ionic Liquids

Fig. 8.2 The effect of the toff time on the nanostructure of gold deposits. An average current density of 3 mA cm−2 was used for all experiments [29].

formation on the electrode surface is observed. The effect of the toff -time on the nanostructure of the deposits is shown in Figure 8.2. As expected the smallest crystallites can be found for the shortest toff -times, in accordance with the deposition mechanism outlined above (recrystallization process of the nuclei during the toff time). To study the effects of organic additives on the nanostructure of the deposit, butanediamine (a molecule with free amino groups), diammonium-EDTA (a chelating complex former) and saccharin (a molecule with a sulfur group) were compared. The electrolytes free of additives yield crystallite sizes in the range of 100 nm. With very small amounts of additives a strong reduction (by a factor of about 2) in the crystallite size is observed, see Figure 8.3. Increasing the concentration of the grain refiners does not cause a substantial further decrease in the crystallite size. The three additives show the same effect, but the degree of grain refining is substance-specific: as is well known, gold reacts strongly with sulfur groups. Similar results were obtained for the system nano-copper/citric acid [64], nano-nickel/saccharin [61, 65] and nano-Pd/Na2 EDTA [42]. A detailed description of the influence of organic additives on the microstructure of metal deposits is given by Fischer [66]. Further nanometals with crystallite sizes between 10 and 100 nm prepared by pulsed electrodeposition from aqueous electrolytes are: Pd [42, 67], Fe [61], Co [68], and Cr [69]. Detailed information on the preparation and physical properties are given in the cited references. 8.2.3 Nanoalloy Deposition with Fex Ni1–x Alloys as an Example

The deposition of Fex Ni1−x alloys is of industrial interest because these materials find applications in electronic devices (e.g. computes hard disk). The most popular

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

8.2 Pulsed Electrodeposition from Aqueous Electrolytes 219

Fig. 8.3 The activity of different grain refiners (butanediamine, ammonium ethylenediamine tetra-acetic acid, benzosulfimide) on the nanostructure of gold deposits [29].

alloys are Permalloy (soft magnetic properties) and Invar (very low thermal expansion). The polycrystalline compounds can be prepared by melting processes or by direct current plating [70]. The magnetic and mechanical properties of these alloys can be designed by nanostructuring. In the case of alloy deposition the bath composition is an additional process parameter which can influence the nanostructure of the deposit. An electrochemical DC current procedure was reported by Cheung et al. [71]. One condition for preparing homogeneous alloys by electrochemical methods is nearly equal electrode potentials of the components. Fe and Ni exhibit standard reduction potentials of −0.44 and −0.22 V; in view of these similar values alloy formation can be expected. Ni exhibits a more positive standard reduction potential than iron and therefore the Ni/Fe ratio in the deposited alloys should be higher than in the electrolyte. Actually, the literature reports opposite experimental results [70]. This anomalous codeposition (ACD) was also observed for CoFe, ZnNi, ZnFe and CuPb. Since the composition of the alloy depends strongly on the pH value [72] the electrolyte has to contain a buffer system. For different concentrations of iron salts (Figure 8.4) alloys (crystallite size: 16–19 nm) with iron contents up to 71 mol% have been obtained. Hessami et al. [73] explain the ACD by an increased dissociation rate of the FeOH+ complex compared to that of the NiOH+ complex. For this reason the concentration of free iron ions in the electrolyte increases and therefore the alloys exhibit an increased iron content. There is the risk (or chance) of preparing gradient materials in which the iron concentration increases with deposition depth whereas laterally the deposited alloy is perfectly homogeneous. Varying the crystallite size by varying the pulse parameters or the temperature is impossible because these factors also change the alloy composition (Figure 8.5). The best way to vary the crystallite size without changing the alloy composition is by the addition of different amounts of grain refiners.

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

220 8 Nanostructured Metals and Alloys Deposited from Ionic Liquids

Fig. 8.4 The influence of the Fe3+ concentration of the electrolyte on the alloy composition [29].

8.3 Special Features of Ionic Liquids as Electrolytes

The electrochemistry of ionic liquids is different in some essential features from the electrochemistry of aqueous electrolytes. Particularly for electrodeposition, which involves charge transfer from the electrolyte to the electrode, the double layer on the electrode is of great importance. In general the cathode is negatively charged for the electrodeposition of metals and therefore coated with a (Helmholtz-) layer of cations at least 0.5 nm thick; but the metal species in most ionic liquids is anionic (for instance AlCl4 − ). This makes the metal deposition process complicated, for more details we refer to Chapter 2.

Fig. 8.5 The influence of the temperature (left axis) and the pulse current density (right axis) on the alloy composition of FeNi alloys [29].

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

8.3 Special Features of Ionic Liquids as Electrolytes 221

In ionic liquids the coordination chemistry and concentration of metal complexes are also substantially different from those in aqueous electrolytes, with consequent effects on both the thermodynamics, i.e., the redox potential, and the kinetics of the deposition process. For details we again refer to Chapter 2. Since the viscosity of ionic liquids is large in some cases and concomitantly the diffusion is slow, ionic liquids generally exhibit a lower conductivity than aqueous electrolytes. To improve the mass transport it has been suggested to add diluents like benzene, toluene or acetonitrile. Water may also be a suitable diluent in some cases, acting as both a ligand and a viscosity improver. Brighteners are common additives in the electroplating industry and are mostly based on empirical recipes. A well-known inorganic example is arsenic acid. Here only the influence of organic compounds on the electrodeposition from ionic liquids is discussed. The main effect of brighteners is adsorption on the electrode surface thus impeding nucleation and influencing growth. To determine the effectiveness of the brighteners Natter et al. [74] have examined organic molecules with different structures (see Table 8.1). The first group comprises aromatic and aliphatic carboxylic acids with one or two carboxylic groups. The second group consists of aromatic carboxylic acids with chlorine substituents in different positions (2-, 3- and 4-chlorobenzoic acid). The third group comprises carboxylic acids with one or two hydroxy substituents and the last group are substances with a sulfur-containing functional group (benzoic acid sulfimide, sodium butanesulfonate, sodium dodecyl sulfate). Aliphatic carboxylic acids and also their hydroxy substituted derivatives (tartaric, malonic, malic and salicylic acids) show hardly any grain refining effect, maybe because these substances do not have free electron pairs which are important for adsorption at many metals. The aromatic salicylic acid reduces the crystallite size down to 54 nm. For this reason the aromatic

Table 8.1 Effect of different additives on the crystallite size (electrolyte:

63 mol% absolute dry AlCl3 , 37 mol% [EMIM]Cl, DC: 5 mA cm−2 , additive concentration: 4 wt.%).

Additive

benzoic acid 3-chlorobenzoic acid 2-chlorobenzoic acid 4-chlorobenzoic acid benzoic acid sulfimide phthalic acid anhydride sodium dodecyl sulfate sodium butane sulfonate tartaric acid salicylic acid malonic acid malic acid

D/nm

19 13 18 24 12 16 21 132 99 54 61 133

c08 (JWBG008-Endres)

January 11, 2008

9:13

Char Count=

222 8 Nanostructured Metals and Alloys Deposited from Ionic Liquids

carboxylic acids seem to be the better grain refiners. The electronic interaction of the aromatic ring can be enhanced with halogen substituents. The use of benzoic acid chloroderivatives shows a strong crystallite refining effect which depends on the position of the chlorine group. 3-Chlorobenzoic acid interacts strongly with the metal surface and works as a good grain refiner. Saccharine (benzoic acid sulfimide) has additional nitrogen and sulfur atoms which increase the surface activity and decrease the crystallite size. The same effect can be observed for long chain sulfonates and phthalic acid anhydride.

8.4 Nanocrystalline Metals and Alloys from Chlorometallate-based Ionic Liquids

Chlorometallate-based ionic liquids are eutectic mixtures of an organic chloride, RCl, and a metal chloride MClx , mostly with M = Al or Zn; other possible but less commonly used metals are Sn, Ga, Fe, or Ge, i.e., chlorostannate, chlorogallate, etc. Chloroborate (BCl4 − ) or bromoaluminate (AlBr4 − ) systems are also possible. In chloroaluminate-based ionic liquids the AlIII exists in complexes like AlCl4 − , Al2 Cl7 − or Al3 Cl10 − , in chlorozincate-based ionic liquids the ZnII exists in complexes like ZnCl2 − , Zn2 Cl5 − or Zn3 Cl7 − , depending on the concentration. The mole fraction of AlCl3 can be denoted as r = [AlCl3 ]/([AlCl3 ] + [RCl]). According to this ratio the chloroaluminate systems are categorized as follows: r > 0.5 called Lewis-acidic, r = 0.5 called Lewis-neutral and r < 0.5 called Lewisbasic melts. These systems have the serious drawback of being extremely sensitive to humidity/water which causes hydrolysis and the formation of HCl: for that reason extreme care has to be applied (controlled inert gas atmosphere with a water content below 1 ppm). However, under these demanding conditions, which might be too difficult in industrial applications, the chloroaluminate-based ionic liquids have attractive electrochemical properties. They allow the electrodeposition not only of Al but also of a large number of other metals: salts or oxides of other metals can be dissolved in chloroaluminate systems, and then the result is dependent on the ratio r mentioned above (and of course on the potential): from Lewis-acidic systems, with r > 0.5, Al alloys are deposited, and from Lewis-basic systems, with r 1.5 V. In situ STM measurements under potentiostatic conditions were performed on Au(111) in the ionic liquid [Py1,4 ] TFSA containing 0.5 M TaF5 . The STM picture of Figure 9.10(a) shows the surface morphology of a deposited tantalum layer obtained at a potential of −1.25 V (vs. Pt). As seen, the deposited layer is rough and some triangularly shaped islands with heights of several nanometers, Figure 9.10(b), grow above the deposited layer. With time, these islands grow vertically and laterally and finally merge together to a thick layer, Figure 9.10(c) and (d). The thickness of the deposited layer obtained from the change of z-position of the piezo, was found to be about 300 nm. In order to investigate whether the in situ deposit is metallic or not, current–voltage (I–U) tunneling spectroscopy was conducted. A typical in situ tunneling spectrum of the 300 nm thick layer of the electrodeposit at different positions is shown in Figure 9.10(e). As seen, the I–U spectrum exhibits metallic behavior with an exponential-like rise in the current, indicating the formation of elemental Ta. There are approaches in the literature to determine the apparent electron work function from such I–U spectra. As several simplifications are required and as the influence of the ionic liquid on the distancedependent tunneling spectra are completely unknown we rather restrict ourselves to a qualitative description. In the light of the above results, it can be concluded that ionic liquids – due to their wide electrochemical windows – not only give access to many elements like e.g. Al and Ta and many others, they also show unexpected cation/anion effects, both at the interface electrode/ionic liquid and on the electrodeposition of metals. Due to the extremely high number of ionic liquids there is an enormous potential to perform original electrochemical studies on both the microscale and the nanoscale. But, there is a price to pay: as it is tough to purify ionic liquids – hitherto they can neither be recrystallized nor distilled nor sublimed without remarkable decomposition – only ultrapure liquids can be recommended for fundamental studies, and even these liquids can contain very low amounts of inorganic impurities. The experimentalist has to find first which impurities are present in a liquid of interest, then if they cause any problems at all and if there is a certain level of impurities which maybe can be tolerated. Thus, there is no quick (and dirty) electrochemistry in ionic liquids, and the comparatively slow progress, especially with the in situ STM, is maybe one of the greatest challenges for the experimentalist.

9.5 Electrodeposition of Poly(p-phenylene)

Conducting polymers have been extensively investigated due to their potential applications in supercapacitors, sensors, batteries, electrochromic devices and light

c09 (JWBG008-Endres)

December 24, 2007

17:48

Char Count=

9.5 Electrodeposition of Poly(p-phenylene) 253

emitting diodes. Polypyrrole and polythiophene are the most studied conducting polymers due to their high stability and simple preparation [17–19]. Among all conducting polymers, poly(p-phenylene) (PPP) is very interesting because it is suitable for the fabrication of blue polymer light emitting diodes (PLED) [20– 22]. However, the electrochemical polymerization of benzene to PPP is still a challenge as water in the solution has to be strictly avoided. Therefore, in the past only solvents like concentrated sulfuric acid [23], liquid SO2 [24] or liquid HF were feasible for the electropolymerization of benzene. In 1993, ionic liquids based on AlCl3 were employed for the first time for the electropolymerization of benzene [25]. Because of side reactions due to chlorine co-evolution during the electropolymerization the quality of the deposits was not satisfactory. Recently, we have reported for the first time that modern air- and water-stable ionic liquids are also well suited for the electropolymerization of benzene [26, 27]. In contrast to the above-mentioned solvents, ionic liquids deliver much milder chemical conditions. The electropolymerization of benzene in the ionic liquid 1-hexyl3-methylimidazolium tris(pentafluoroethyl)trifluorophosphate [HMIM]FAP is well reproducible and gives as a result poly(p-phenylene) of spherulitic morphology with grain sizes as small as 500 nm [26]. Figure 9.11(a) shows three successive cyclic voltammograms for the oxidation of benzene in the ionic liquid [HMIM]FAP. In the anodic scan of the first cycle (curve 1) an anodic current starts at an electrode potential of 1.8 V (vs. Pt-quasi reference). The current rises strongly, indicating the oxidation of benzene to form a polymer. It is worth mentioning that the rise of this current occurs at an electrode potential that is about 1 V below the anodic decomposition limit of the liquid, see Ref. [26]. In the back scan a reduction process at a potential of 0.75 V is recorded. In the second anodic cycle (curve 2) an oxidation peak at an electrode potential of 1.25 V is obtained, and in the second cathodic scan with reference to the first cycle a more pronounced current flows. In the third cycle (curve 3) both anodic and cathodic currents increase. Such cyclic voltammetric behavior is typical for the electrosynthesis of conducting polymers. Visual inspection during the deposition of the polymer reveals that a yellowish film has formed on the platinum electrode after the second cycle. With subsequent cycles the color changes from yellowish to brownish, and finally a black polymer film is obtained on the electrode surface. Quite a similar observation was reported by Kowalski et al. during the electrosynthesis of poly(p-phenylene) by polymerization of benzene in a mixture of glacial acetic acid and concentrated sulfuric acid [28]. Here, a thick and well adhering polymer film was also obtained by applying, for a sufficiently long time, an electrode potential of 2.2 V vs. the quasi-reference to the platinum working electrode. The surface morphology of an electrodeposited polymer film on the platinum electrode is shown in Figure 9.11(a). As seen, the film consists of small spherical and globular grains with an average diameter of about 3 µm. The smallest grains that can be observed with the selected resolution of the SEM have sizes of around 500 nm. IR spectra of the deposited polymer film reveal the formation of poly(p-phenylene) as a result of the polymerization of benzene in the employed ionic liquid [26]. In order to gain further insight into the growth and characterization of the deposited polymer film we acquired in situ STM and STS measurements.

c09 (JWBG008-Endres)

December 24, 2007

17:48

Char Count=

254 9 Electrodeposition on the Nanometer Scale: In Si tu Scanning Tunneling Microscopy

Fig. 9.11 (a) Cyclic voltammogram of 0.2 M benzene in [HMIM]FAP on platinum. The numbers refer to the respective cycle. Scan rate: 10 mV s−1 . (b) SEM micrograph of the electropolymerized film on the platinum electrode after synthesis in [HMIM]FAP.

Figure 9.12 shows a set of in situ STM images obtained on Au(111) in the ionic liquid [HMIM]FAP containing 0.2 M benzene, as well as an I–U spectrum of a deposited PPP layer. As shown in the STM image of Figure 9.12(a), obtained at a potential of 0.9 V, the gold surface is subject to slight oxidation at such anodic potential and the gold terraces are still distinct. At 1.3 V, a number of randomly distributed 2D-islands are formed on the gold surface, as manifested in the STM image of Figure 9.12(b). This indicated the start of the formation of PPP. By setting the potential at 1.9 V, a relatively thick polymer layer of PPP is obtained, Figure 9.12(c). In order to measure the band gap of the deposited polymer layer, current–voltage tunneling spectroscopy was performed. It has already been shown by us that I–U

c09 (JWBG008-Endres)

December 24, 2007

17:48

Char Count=

9.5 Electrodeposition of Poly(p-phenylene) 255

Fig. 9.12 In situ STM images of Au(111) in the ionic liquid [HMIM]FAP containing 0.2 M benzene at (a) 0.9 V, (b) 1.3 V and (c) 1.9 V vs. Pt-quasi ref.(d) In situ I–U spectrum of a PPP layer obtained at a potential of 1.9 V (vs. Pt).

tunneling spectroscopy is a valuable technique for in situ characterization of electrodeposited semiconductors [29–31] and metals [15]. We could show with in situ I–U tunneling spectroscopy that germanium with a layer thickness of 20 nm and more is semiconducting with a symmetric band gap of 0.7 ± 0.1 eV. On the other hand, we have found that very thin layers of germanium with thicknesses of several monolayers clearly exhibit metallic behavior [30, 31]. Figure 9.12(d) represents an in situ current–voltage tunneling spectrum of the deposited PPP layer shown in Figure 9.12(c). As seen in the spectrum, a band gap of 2.1 ± 0.2 eV is recorded. This value approaches the value reported in the literature for PPP, 2.7 eV, [32]. A detailed in situ STM and STS study on the film formation stages during the electropolymerization of benzene in the ionic liquid [HMIM]FAP is now in preparation and will be published elsewhere [33].

c09 (JWBG008-Endres)

December 24, 2007

17:48

Char Count=

256 9 Electrodeposition on the Nanometer Scale: In Si tu Scanning Tunneling Microscopy

9.6 Summary

Ionic liquids represent a promising class of solvents with unprecedented properties for electrochemistry. They can have wide electrochemical windows of 6 V, they have – in most cases - practically no vapor pressure around room temperature, they have wide thermal windows of 300 ◦ C and they show unusual cation/anion interaction which can influence chemical and electrochemical reactions. Ultrapure ionic liquids allow in situ STM experiments at deeply negative electrode potentials with high quality and, consequently, give insight into the initial electrodeposition of reactive elements like Si, Al, Ta and presumably many others. Some ionic liquids like those with the bis(trifluoromethylsulfonyl)amide anion are, however, subject to an unexpected anion breakdown which can alter the nanoscale processes. Varying the cation or the anion of an ionic liquid might have a dramatic influence on the surface electrochemistry, as shown with the example of Al deposition. This opens the door to many fundamental studies, not only with classical electrochemistry but also with in situ STM. As shown in the example of benzene polymerization, the growth and in situ characterization of conducting polymers can be probed on the nanoscale with in situ STM.

References 1 Aravinda, C.L., Burger, B., and Freyland, W. (2007) Chem. Phys. Lett., 434, 271. 2 Mann, O., Aravinda, C.L., and Freyland, W. (2006) J. Phys. Chem. B, 110, 21521. 3 Aravinda, C.L., Mukhopadhyay, I., and Freyland, W. (2004) Phys. Chem. Chem. Phys., 6, 5225. 4 Bonnell, D. (ed.) (2001) Scanning Probe Microscopy and Spectroscopy, Wiley-VCH, Verlag GmbH. 5 Endres, F., Zein El Abedin, S., and Borissenko, N. (2006) Z. Phys. Chem., 220, 1377. 6 Boressinko, N., Zein El Abedin, S., and Endres, F. (2006) J. Phys. Chem. B., 110, 6250. 7 Lin, L.G., Wang, Y., Yan, J.W., Yuan, Y.Z., Xiang, J., and Mao, B.W. (2003) Electrochem. Commun., 5, 995. 8 Howlett, P.C., Izgorodina, E., Forsyth, M., and MacFarlane, D.R. (2006) Z. Phys. Chem., 220, 1483. 9 Zein El Abedin, S., Saad, A.Y., Farag, H.K., Borisenko, N., Liu, Q.X., and Endres, F. (2007) Electrochim. Acta, 52, 2746.

10 Zein El Abedin, S., Moustafa, E.M., Hempelmann, R., Natter, H., and Endres, F. (2006) Chem. Phys. Chem., 7, 1535. 11 Moustafa, E.M., Zein El Abedin, S., Shkurankov, A., Zschippang, E., Saad, A.Y., Bund, A., and Endres, F. (in press) J. Phys. Chem. B. 12 Zell, C.A., Endres, F., and Freyland, W. (1999) Phys. Chem. Chem. Phys., 1, 697. 13 Endres, F. (2003) in Ionic Liquids in Synthesis, (eds P. Wasserscheid and T. Welton), Wiley-VCH, Verlag GmbH. 14 Staikov, G., Lorenz, W.J., and Budevski, E. (eds.) (1996) Electrochemical Phase Formation and Growth, Wiley-VCH, Verlag GmbH. 15 Zein El Abedin, S., Farag, H.K., Moustafa, E.M., Welz-Biermann, U., and Endres, F. (2005) Phys. Chem. Chem. Phys., 7, 2333. 16 Zein El Abedin, S., Welz-Biermann, U., and Endres, F. (2005) Electrochem. Commun., 7, 941. 17 Skotheim, T.A., Elsenbaumer, R.L., and Reynolds, J.R. (1998) Handbook of

c09 (JWBG008-Endres)

December 24, 2007

17:48

Char Count=

References

18

19

20 21 22 23 24

25

Conducting Polymers, 2 edn, Marcel Dekker, New York. Kobayashi, T., Yoneyama, H., and Tamura, H. (1984) J. Electroanal. Chem., 161, 419. Inzelt, G., Pineri, M., Schultze, J.W., and Vorotyntsev, M.A. (2000) Electrochim. Acta, 45, 2403. Grem, G., Leditzky, G., Ullrich, B., and Leising, G. (1992) Adv. Mater., 4, 36. Grem, G., Leditzky, G., Ullrich, B., and Leising, G. (1992) Synth. Met., 51, 383. Leising, G. (1993) Phys. Bl¨atter, 49, 510. Shepard, A.F. and Dannels, B.F. (1966) J. Polym. Sci., Polym. Chem., 4, 511. Aeiyach, S., Soubiran, P., Lacaze, P.C., Froyer, G., and Pelous, Y. (1995) Synth. Met., 68, 213. Lere-Porte, J.P., Radi, M., Chorro, C., ` Petrissans, J., Sauvajol, J.L., Gonbeau, D., Pfister-Guillouzo, G., Louarn, G., and Lefrant, S. (1993) Synth. Met., 59, 141.

26 Zein El Abedin, S., Boressinko, N., and Endres, F. (2004) Electrochem. Commun., 6, 422. 27 Schneider, O., Bund, A., Ispas, A., Boressinko, N., Zein El Abedin, S., and Endres, F. (2005) J. Phys. Chem. B., 109, 7159. 28 Kowalski, J., Ploszynska, J., and Sobkowiak, A. (2002) Synth. Met., 130, 149. 29 Boressinko, N., Zein El Abedin, S., and Endres, F. (2006) J. Phys. Chem. B., 110, 6250. 30 Endres, F. and Zein El Abedin, S. (2002) Phys. Chem. Chem. Phys., 4, 1640. 31 Endres, F. and Zein El Abedin, S. (2002) Phys. Chem. Chem. Phys., 4, 1649. 32 Leising, G., Pichler, K., and Stelzer, F. (1988) in Electronic Properties of Conjugated Polymers III (eds H. Kuzmany, M. Mehring, and S. Roth), Springer, Heidelberg, p. 100. 33 Carsten, T., Zein El Abedin, S., and Endres, F. (to be published).

257

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

259

10 Plasma Electrochemistry with Ionic Liquids J¨urgen Janek, Marcus Rohnke, Manuel P¨olleth, and Sebastian A. Meiss

10.1 Introduction

Electrochemical reactions occur at the interface between two phases with sufficiently different conduction behavior, i.e. a predominantly ion-conducting electrolyte phase and an electrode phase with predominantly electronic conduction. Among all possible types of interfaces the most intensively applied are solid metal|liquid electrolyte and solid metal|solid electrolyte. Electrode systems which have been much less studied are those formed by combining either a solid or liquid conducting phase with a low-temperature gas discharge (plasma). This chapter aims to discuss and summarize theoretical and practical aspects of such plasma interfaces, presenting the existing examples from our own recent work on plasma electrochemical reactions between typical ionic liquids and plasmas. First, we address the plasma state and essential properties with respect to its application in electrochemistry. Today, low temperature plasmas – mostly in the form of radiofrequency or microwave plasmas – play an important role in the treatment or modification of solid surfaces. However, as plasma chemistry is usually not an element of chemistry curricula, we include a very brief introduction but refer the reader to the literature for more detailed information. Plasma electrochemical reactions have been studied by chemists for a surprisingly long time, with the first report on cathodic metal deposition at the free surface of a liquid electrolyte with free electrons from a plasma dating back to 1887 [1], long before the plasma state had been named by Langmuir in 1928 [2]. A short survey of past work with more conventional liquid electrolytes is also included in this chapter. Typical low-temperature plasmas are usually only weakly ionized and quasineutral but are thermally in a non-equilibrium state, i.e. the different plasma species (molecules, atoms, ions and electrons) possess different kinetic energy distributions. Because of their small mass electrons acquire much more kinetic energy than atomic or molecular species and thus show an energy distribution which corresponds to a much higher temperature than in the case of much heavier particles. The stability of ionic liquids towards reduction by these “hot” electrons Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

260 10 Plasma Electrochemistry with Ionic Liquids

and also towards reactions with other reactive plasma species is considered in Section 10.4. The central part of this chapter comprises a summary of our recent attempts to deposit metals from ionic liquids by plasma cathodic reduction. The concluding part then analyzes the plasma electrochemical approach with respect to possible applications.

10.2 Concepts and Principles

Like other salt melts ionic liquids are characterized by a specific combination of physicochemical properties: high ionic conductivity, low viscosity, high thermal stability compared to conventional liquid solvents, wide electrochemical windows of up to 7 V and – in most cases – extremely low vapor pressures. Due to their low vapor pressure ionic liquids are not only well suited for the application of UHV-based analytical techniques (e.g. photoelectron spectroscopy [3]), but also for use in plasma reactors with typical pressures of the order of 1 Pa up to 10 kPa. Moreover, due to their high electrical conductivity, ionic liquids may even be used as “electrodes” for plasmas. To date there are just a few reports on the combination of low-temperature plasmas and ionic liquids available in the literature [4–6]. Therefore, the essential aspects of experiments with ionic liquids in typical plasma reactors are discussed in this section. 10.2.1 Plasma Electrochemistry

As plasma chemistry deals with charged particles, there is no doubt that it can be considered as plasma electrochemistry a priori. However, a comparison of electrochemistry and plasma chemistry [4–9] in more detail is instructive. Electrochemistry deals with the interplay of electric fields (potential differences) and chemical reactions. Once electric fields get strong enough, e.g. at interfaces, electron transfer can be enforced leading to reduction or oxidation of chemical compounds. Two routes exist within this interplay: (i) external electric potential differences can be used to control chemical processes; (ii) chemical processes can be used to generate electric potential differences. The first (synthetic/charging) route is the basis of electrochemical synthesis and galvanic technology. The second (analytical/discharging) route is the basis for batteries, fuel cell and sensor technology. Obviously, plasmas can be used very efficiently within the synthetic approach (i), and all examples given in this paper are assigned to the synthetic approach. It is much less obvious whether plasmas can be used also in the counter-direction. In order to measure a stable and reproducible electromotive force (EMF) the corresponding electrochemical (galvanic) cell must be in (local) thermodynamic equilibrium. Low-temperature plasmas represent non-equilibrium states and are highly inhomogeneous systems from a thermodynamic point of view, often not

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.2 Concepts and Principles 261

fulfilling the conditions of local equilibrium (at least at interfaces). Therefore EMF measurements in these plasmas will not provide easily accessible thermodynamic information. Flames can, in small volumes, be considered as equilibrium systems under certain conditions, and almost a century ago the first attempts to measure EMF signals in flame plasmas were reported [10–12]. The theoretical analysis of the EMF measurements in flames is complicated by complex electrode processes and, to date, no adequate treatment has been published. First steps towards a correct and comprehensive description have been published by Caruana et al. [13, 14]. Several other researchers have tried to measure EMF signals in flames but got no stable signal (e.g. Lorenz et al. [15]). 10.2.2 Low-temperature Plasmas: Electrodes or Electrolytes?

From the electrochemical point of view, the answer to this question appears, at first glance, to be simple: electrolytes are usually defined as electrically conductive media with a negligible electronic conductivity, for example as purely ionic conductors. In contrast, electrode materials have to be predominantly electronic conductors (mostly metals). This definition originates from electrochemistry in the liquid state, where an electronic contribution to the bulk charge transport in electrolytes is a rare phenomenon (except in some well-known cases, e.g. sodium/ammonia solutions). In solid state electrochemistry we deal mostly with mixed ionic/electronic conductors, as electronic charge carriers are a priori always present in solids, either as intrinsic defects (electron–hole pairs due to a sufficiently small band gap) or as a consequence of non-stoichiometry (metal excess: n-type doping, non-metal excess: p-type doping). The mixed character of conduction in the solid state is the basis for chemical diffusion of crystal components and, among other effects, is responsible for the occurrence of diffusion potentials. It becomes obvious in the Nernst equation for the EMF of solid state galvanic cells, which contains the electronic transference number as a factor. A plasma is always a mixed ionic and electronic conductor, mostly with a small ionic transference number. The majority of charge carriers in plasmas of electropositive gases (e.g. noble gases) are cations and electrons. The charge carriers in electronegative gases (e.g. halogens) are cations, anions and electrons. In addition to their two orders of magnitude higher mean energy, the mobility of free electrons is two orders of magnitude higher than the mobility of free ions and, thus, the electronic partial conductivity is a priori much larger than the ionic partial conductivity. Following these arguments, plasmas should rather be regarded as electrodes than as electrolytes. However, this simple analysis completely neglects the inhomogeneity of most plasmas, in particular in the boundary regions in front of walls. Low-temperature plasmas are non-equilibrium systems which only exist in stationary states driven by a continuous conversion of electric energy into heat. Diffusion of the charged plasma species from the plasma bulk to the plasma boundaries establishes extended space charges with large diffusion potentials and leads to

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

262 10 Plasma Electrochemistry with Ionic Liquids

a negative charging of the plasma walls. The resulting (electron-poor) positive space charge region acts as a rectifying element and leads to strongly non-linear current–voltage characteristics. As a consequence, the total resistance of a plasma is rather controlled by the mobility of ions in the plasma sheaths than by the mobility of electrons (see below). Depending on the applied potential to an electrode in a plasma, the carrier concentration in the plasma sheath changes considerably. Under negative polarization the plasma sheath is particularly poor in electrons and therefore will act as an electrolyte rather than an electrode. In the present context, we suggest considering low-temperature plasmas as fluid mixed conductors with a small ionic transference number in principle – in particular for positive (anodic) polarization of the electrode in the ionic liquid and corresponding negative (cathodic) polarization of the electrode in the plasma. As exemplified below, plasmas are more often used as gaseous electrodes than as electrolytes. One of the most important application of electrolytes, i.e. their use as “electron filters” in galvanic cells, is hampered by the large electronic contribution to the bulk conductivity and by the relatively large diffusion potential within the plasma sheaths. There is also a practical aspect that further complicates quantitative potential measurements in plasmas: the local plasma density and the corresponding charge carrier density depend on the boundary conditions set by the reactor walls and the electrode arrangement. Often the wall and electrode surfaces are slowly covered with thin sputtered films. Once these are electrically conducting the local plasma state may change considerably with time. 10.2.3 The Plasma|Electrolyte Interface

One of the characteristic features of plasmas is their inhomogeneity at boundaries. The faster electrons charge any wall of a plasma reactor negative and leave a positive space charge in front of the wall, the so-called “sheath” [16]. This space charge can formally be treated like a diffusion potential in conventional electrolytes. Depending on the degree of ionization, the Debye length can take large values (up to several millimeters) for dilute plasmas with large potential drops within this region. The experimental study of these inhomogeneous plasma boundaries is hampered by the fact that most methods interfere with the electric fields within the plasma. Therefore, only a few methods can be used for the investigation. However, for a qualitative discussion we can restrict ourselves to the general picture of plasma|solid boundaries with a positive space charge in the plasma in front of the solid, extending to the order of the Debye length. The situation at the plasma|wall interface is depicted schematically in Figure 10.1. The negatively charged wall does not affect neutral particles, while positive particles are accelerated towards it and negative particles are repelled from the wall so that only highly energetic negative particles can reach it. In a stationary state the fluxes of highly energetic negative and positive particles towards the wall compensate each other. The consequence of the plasma sheaths is two-fold: first, significant voltages have to be applied to plasma electrochemical cells in order to draw sufficiently

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.2 Concepts and Principles 263

Fig. 10.1 Positive space charge layer at the interface between a plasma and (a) a dielectric, (b) a metallic, (c) an electrolytic wall with floating potential W . Due to the negative surface charge mainly neutral, positive and only high energetic negative plasma particles reach the wall ( je− : flow of electrons, jK+ : flow of cations, jK : flow of neutralized particles, jA− : flow of anions). The potential difference between the zero poten-

tial 0 and the potential of the plasma p is denoted as plasma potential p . If the wall’s potential is not “free floating” (without an applied external potential) as shown above, the characteristics of the potential in the wall, the surface charge at the interface and hence the characteristics of the sheath and pre-sheath potential can be influenced by the applied external potential.

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

264 10 Plasma Electrochemistry with Ionic Liquids

large currents for electrochemical reactions, see below. Only by a careful design of the electrodes can we overcome the electric current limitations of the space charge regions. Secondly, the large electric fields within the plasma sheaths lead to a significant acceleration of ionic charge carriers and the resulting sputter effects. 10.2.4 Types of Plasmas and Reactors

The properties of plasmas vary strongly with gas composition, pressure and the method and parameters of the plasma generation process. The charge carrier concentration depends on the pressure and the fractional ionization of the plasma, for instance basically on the power density. The mobility of the electrons depends on the electron temperature, which is typically several orders of magnitudes greater than the gas temperature or the temperature of the ionized species in non-thermal low temperature plasmas used for electrochemical purposes. Both parameters, as well as the gas composition, can usually be changed relatively easy within the limits of the given experimental set-up. When employing the plasma merely as a gaseous but chemically inert electrode, one will choose a noble gas, typically argon. For other purposes, reactive gases might be added to the noble gas or replace it. Depending on the reaction at the plasma|electrolyte interface, gaseous reaction products may emerge into the plasma, or components may disappear from it due to reactions with the electrolyte or substances dissolved in it, therefore changing its composition. As the pressure of a typical non-thermal low-temperature plasma is two to four orders of magnitude smaller than atmospheric pressure, even small absolute amounts of emerging or disappearing gas components can have a relatively high effect on the gas composition of the plasma. Therefore, it is desirable to design plasma electrochemical experiments in such a way that allows a high gas throughput, especially near the interface, to retain a well defined gas composition. The method of the plasma generation has a strong influence on the parameters of the plasma, including the distribution of the various ionized species of the gas components. Feasible plasma reactors are direct current (DC) discharge reactors and inductively or capacitively coupled radio frequency (RF) discharge reactors, the latter not being discussed here further. Microwave (µW) discharges are spatially much more concentrated, due to the much smaller wavelength, and lead to a considerable increase in temperature. They are often applied in the form of “remote plasma sources”, i.e. using a gas flux with particles excited during their passage through the plasma source rather than working in the center of the plasma source. Figure 10.2(a) shows the set-up for a DC discharge, where the electrolyte (ionic liquid) is part of the serial electric circuit. This set-up has the advantage that it is possible to gain information about the formation rate of a product at the plasma/electrolyte interface directly by measuring the electric current and having information about the relevant transference numbers. However, it is not possible to freely choose the applied voltage. It has a lower limit, given by the voltage of the order of 100 –1000 V needed to sustain the discharge. Figure 10.2(b) shows a DC discharge set-up where the electrolyte is not necessarily part of the electric circuit.

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.3 Early Studies 265

Fig. 10.2 Different types of plasma reactors employing the use of an IL: (a) DC discharge with the IL as an integral part of a serial set-up, (b) DC discharge with the IL as optional part of a parallel set-up, (c) in-

ductively coupled RF discharge with an electric circuit for electrochemical experiments independent from the plasma generating process.

It, rather, represents an ion-conducting wall of the plasma at a floating potential and reactions are motivated by the plasma–wall interactions described earlier. It is feasible to introduce a third electrode to the system, placing it in contact with the electrolyte, but not with the plasma, and therefore gaining some control over the potential difference between the electrolyte and the plasma. In the case of purely ion-conducting electrodes, the electric current offers information about the reaction rate at the plasma/electrolyte interfaces. In an inductively coupled RF discharge (Figure 10.2(c)) the plasma is not in contact with the external RF coil (“electrode-free discharge”). Again the ionic liquid acts as a “wall” to the plasma, with the effects described earlier. Its floating potential will be negative, due to the collected electrons, and a positive space charge is found above the surface. Introducing an electrode to the electrolyte allows one to influence its then no longer “floating” potential. A second electrode can be placed in the gas phase, but often metallic parts of the reactor itself are used as the second electrode. This set-up has been applied successfully in experiments with solid electrolytes and typical I–U curves are reported by Vennekamp [17].

10.3 Early Studies

The use of gas discharges for electrochemical processes has been investigated for more than 100 years, and a full account is beyond the scope of this chapter. We will focus on a few innovative and seminal studies which can be regarded as major advances. The first plasma electrochemical experiments were already reported in 1887 by Gubkin [1], in the same year when Arrhenius published his most influential paper on electrolytic dissociation of salts in water [18]. Gubkin investigated

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

266 10 Plasma Electrochemistry with Ionic Liquids

Fig. 10.3 Set-up of the reproduced Gubkin experiment: silver is dissolved at the anode inside of the liquid electrolyte and reduced at the plasma|electrolyte interface; and photograph of the laboratory experiment.

the plasma-assisted cathodic deposition of silver, platinum and zinc oxide. For this purpose an aqueous metal salt solution was put into a round flask fitted with two platinum electrodes. The anode was coated with a layer of the same metal that was dissolved in the form of its salt in the electrolyte. It was immersed in the liquid electrolyte and a vacuum was generated above the electrolyte by cooling the bulb with the boiling electrolyte after sealing the bulb. A glow discharge over the surface of the liquid electrolyte was produced by applying a high voltage between both electrodes. Gubkin observed the deposition of clearly visible metal particles, formed by reduction of the metal cations with free electrons from the plasma at the interface between the plasma and the liquid. The plasma electrochemical cell can be summarised as: metal Me (anode) | salt solution (Mez+ ) | plasma | inert metal Pt (cathode) Figure 10.3 shows Gubkin’s original experiment, as it was reproduced in our laboratory. A sketch of the experimental set-up and a photograph of the experiment are depicted. It has to be mentioned that Gubkin was not the first person, who reduced metal ions or metal compounds by using plasmas. Trasatti [19] reports on experiments performed by Father Beccaria as early as 1750, who seemingly observed the reduction of zinc oxide to zinc metal by an electric discharge. In the 1920s the phenomenon of electrostenolysis was investigated by S¨ollner [20]. When a voltage higher than required for electrostenolysis (U > 20 V) was applied to the cell anode | transition metal salt solution || membrane || heavy metal salt solution | cathode the deposition of metal was observed inside the membrane. In addition a light emission was observed at higher voltages, which indicates the occurrence of micro gas discharges. At still higher voltages spark discharges were observed. This

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.3 Early Studies 267

phenomenon is explained by the electrolytic formation of gas in pores and cracks of the glass membrane. Due to the high electric resistance of the gas bubbles, the main electrical potential decay is assumed at the gas bubbles inside the pores. In consequence, the electric fields at these pores are very high and micro-plasmas are generated inside the bubbles. At the gas|electrolyte interface metal deposition takes place. Similar phenomena of plasma formation in gas bubbles were observed in electrolytic commutators and capacitors [21]. The study of micro-, spark or arc discharges in liquid electrolytes (usually referred to as plasma electrolysis) has been continued by other groups [22], depositing either metals or metal oxides. Here the metal or metal oxide is deposited cathodically or anodically, respectively in the presence of a gas discharge in front of the electrode. Shen et al. reported that the resulting metal or metal oxide layers are comparatively dense and show better corrosion protection than conventionally deposited coatings [23]. They propose this plasma-assisted deposition as a method with high potential for industrial application in corrosion protection. The processes within micro arc discharges in liquid electrolytes are complex and not yet fully understood. As the properties of these discharges themselves cannot be controlled directly by well adjustable experimental parameters, we exclude them at this point from further consideration. However, ionic liquids may provide new opportunities for the further development of spark electrolysis. Gubkin’s simple plasma electrochemical experiment was reproduced and improved in the 1950s and 60s, mainly by Klemenc and Brenner [24–29]. A typical experimental set-up of Klemenc is depicted in Figure 10.4. The process was named glow discharge electrolysis or electrode-less electrolysis (which is a misnomer a priori, as electrolysis always requires electrodes) and an attempt was made to explain the phenomena occurring at the surfaces of the electrolytes. Surprisingly the observed yields of oxidation or reduction products were often higher than expected by Faraday’s law (positive deviation), e.g. as reported by Klemenc in the case of the oxidation of hydrochloric acid [24]. This positive deviation from Faraday’s law caused

Fig. 10.4 Experimental set-ups for different plasma (electro)chemical experiments: (a) DC set-up of Klemenc, (b) set-up for the recovery of metal from slags, (c) vapor-phase electrolytic deposition set-up of Ogumi et al.

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

268 10 Plasma Electrochemistry with Ionic Liquids

temporarily strong interest in glow discharge electrolysis. The effect was attributed to reactions driven by local high-temperature spots (hot spots) or by additional reactions caused by UV emission from the plasma. Negative deviations are usually caused by a partial electronic conductivity in the electrolyte or by sputter loss of the product. Klemenc concluded that the interface between a liquid electrolyte and a plasma is comparable to the metal|plasma interface [24]. The electrolytic decomposition of organic compounds at plasma electrodes was investigated both as an important side reaction and as a possible application itself. Compared to Gubkin’s original experimental arrangement the setups were improved, e.g. the electrode areas were separated spatially and the vacuum generation was improved by using better vacuum pumps. In addition, by reversing the applied potential to the electrode immersed in the plasma, plasma-anodic experiments were performed [28], but the mechanism of plasma-anodic processes remained unclear. Brenner and other authors investigated the glow discharge electrolysis of metal salt melts [29, 30]. At the interface between the salt melt and the plasma they deposited dendrites of zinc, cobalt, copper, silver and nickel. Their investigations can be considered as the forerunner of our current studies of ionic liquids. Glow discharge electrolysis reappeared in the 1990s as plasma electrolysis. New types of plasma reactors and discharges were developed and introduced for the deposition of either metals or metal oxides. Ogumi focused on solid electrolytes and developed a method for the plasma electrochemical deposition of ion-conducting metal oxides without liquid electrolytes [31, 32]. For this purpose he injected metal-containing precursors (e.g. ZrCl4 and YCl3 ) into a capacitively coupled RFdischarge. The experimental set-up is shown in Figure 10.4. Yttria-stabilised zirconia was then formed on an oxygen-conducting substrate by electrolytic deposition applying direct current between the substrate and a counter electrode within the plasma. Vennekamp et al. studied this approach more systematically and quantitatively. They proved both the Faradaic character of plasma electrochemical processes and the specific surface morphologies of plasma electrochemically grown solid electrolyte films [33, 34]. Today, plasma electrolysis of liquid electrolytes is applied to waste water treatment [35]. In these applications ozone is formed in the discharge region, which then reacts with organic waste molecules in the liquid solution. During the last few years another method for plasma electrolysis has been developed. Thermal plasmas in the form of a plasma torch are used to melt metal oxides and salts [36]. By applying an additional DC voltage via the thermal plasma electrode (anode), the pure metal or metal alloys are deposited at the cathode which is located in the melt (see Figure 10.4). The advantage of this process is that, due to the high temperature of the plasma discharge, metal oxides can be used as electrolytes. The process allows the direct recovery of pure metals from a slag of metal oxides [37]. The electrochemical cell is: anode | metal oxide slag | cathode He et al. reported the production of noble metal nanoparticles (Ag, Au, Pd, Pt) by using plasmas [38], but no external voltage was applied, and the reduction was

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.4 The Stability of Ionic Liquids in Plasma Experiments 269

achieved with free electrons from the gas discharge under a floating potential. They incorporated noble metal cations into a titanium dioxide gel by ion exchange and reduced the cations by hydrogen low-temperature plasma treatment in a commercial plasma etcher. Inside the matrix nanoparticles of 2–10 nm in diameter were produced. Directly applying Gubkin’s concept of a plasma cathode, Koo et al. produced isolated metal nanoparticles by reduction of a platinum salt at the free surface of its aqueous solution [39]. The authors used an AC discharge as cathode over the surface of an aqueous solution of H2 PtCl6 . Platinum particles with a diameter of about 2 nm were deposited at the plasma|liquid electrolyte interface by reduction with free electrons from the discharge. metal anode | aqueous H2 PtCl6 solution | plasma | metal cathode As indicated by Koo et al. in their paper and as shown in Figure 10.3, the gas discharge over an aqueous solution is a localised corona discharge rather than an extended plasma. This leads to a spatially highly inhomogeneous reduction process. As demonstrated in Section 10.5 the use of ionic liquids leads to homogeneous and extended gas discharges, contacting the whole surface area of the electrolyte. To our knowledge, this type of spatially extended and homogeneous plasma/electrolyte interface has not been investigated before.

10.4 The Stability of Ionic Liquids in Plasma Experiments

The voltages which are applied in order to ignite a DC discharge or which exist across plasma sheaths are far beyond the electrochemical window limits of any ionic liquid. But only a small part of the applied voltage (several hundreds of volts) actually drops across a pure ionic liquid or an ionic liquid containing an arbitrary metal salt situated beneath the burning plasma. Nevertheless, one may expect severe decomposition reactions and a number of questions can be raised: first, does the possible decomposition of ionic liquids lead to impurities of the obtained particles? And if so, to what extent? Secondly, does it affect the deposition negatively in other ways, e.g. by inhibiting the desired reaction? Thirdly, does the decomposition reduce the solubility of the metal salts and restrict the reusability of the ionic liquid? This section discusses some of these questions on the basis of reports on ionic liquid decomposition reactions. Our key reaction is the reduction of a metal salt dissolved in the ionic liquid with free electrons from plasmas in order to obtain metal particles. Processes in lithium ion batteries which employ ionic liquids with dissolved lithium salts can be considered as close relatives. In both cases the dissolved metal salts crucially affect the stability of the ionic liquid. Only if the anion of the dissolved metal salt is more easily oxidised and its cation is more easily reduced in comparison to the ions of the ionic liquid, the decomposition of the ionic liquid would be negligible; of course

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

270 10 Plasma Electrochemistry with Ionic Liquids

this would be the ideal case. At present, the objective of numerous research groups is to avoid the electrochemical decomposition of the ionic liquid or at least to reduce the extent of this side reaction. Hence, some decomposition reaction pathways of ionic liquid ions are already well investigated. Most examples stem from the field of lithium ion batteries, where the electrochemical stability of the electrolyte is a crucial point, e.g. with regard to the rechargeability of the devices. These examples will be briefly reviewed in this section, as ionic liquids with good stability towards lithium will probably be suitable candidates for plasma electrochemical reactions. The decomposition of the trifluoromethanesulfonate anion, CF3 SO3 − (OTf), and its derivatives, the imidazolium and pyrrolidium cations, are primarily considered. At the end of this section some proposals are given for reduction of the decomposition of the ionic liquid or perhaps even to avoid it completely. In general the electrochemical stability of an electrolyte is experimentally evaluated by means of cyclic voltammetry. However, the determination of the electrochemical windows exhibits several problems. First, the electrochemical degradation or breakdown of an electrolyte is an irreversible reaction, thus there is no theoretical redox potential [40, 41]. Passivation of the electrodes often makes it difficult to identify the onset of the reaction due to inhibition of further reactions [40, 42]. Some of the already used electrolytes, and also future candidates for lithium ion batteries, are based on organic solvents like propylene carbonate (PC), vinylene carbonate (VC), 1,2-dimethoxyethane (DME), etc. containing a lithium salt instead of an ionic liquid including a lithium salt. The organic solvent molecules of the electrolyte decompose simultaneously beside the electrolyte ions [40, 43, 44]. Thus different orders of anion stabilities were obtained for different electrolyte compositions [40, 45, 46]. Of course, impurities can lead to a similar phenomenon. The connection between the disintegration reactions of electrolyte salt and electrolyte solvent as well as the influence of their composition ratio was demonstrated, for example, by Rahner [47]. Koch et al. tried to circumvent the problem using pure ionic liquids to investigate the stability of anions, in order to find the most suitable counterion for the lithium ion, without distortion by a solvent [45]. However, once some ions of the ionic liquid were reduced or oxidised they could form neutral organic molecules (as will be described below) acting as impurities and leading to similar problems as for PC, VC, DME, etc. lithium salt solutions. Another point is that the reduction and oxidation potential limits (electrochemical window) are defined as the potentials at which the current density reaches a predefined value that is arbitrarily chosen [40, 48]. Ue et al. also mention that the same problem arises in the choice of the sweep rate [40]. For example Egashira and coworkers obtained a log I–U line shifted to a higher position at a faster potential scan in comparison to a slower scan because of non-Faradaic currents such as the larger charging currents of the double-layer, and the decomposition of impurities [41]. The last factor affecting the electrochemical window is the electrode itself, its composition and its morphological surface structure, which defines the electrocatalytic properties [40]. Johansson compared several theoretical measures in order to find a theoretically calculable substance property which correlates to the oxidation potential of anions

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.4 The Stability of Ionic Liquids in Plasma Experiments 271

and thus allows the prediction of the anodic stability limits of anions from different anion families. First, he compared the highest occupied molecular orbital (HOMO) energy, which he converted as all the other energy changes to electrochemical potentials additionally corrected for the Li+ /Li0 electrode, with the experimental literature oxidation potentials. Secondly, he used the vertical transition energy, which is the energy difference between the anion and the corresponding unrelaxed neutral radical following the Frank–Condon principle. The first two quantities are gas phase energies by definition. In order to mimic real battery electrolyte species better he carried out additional single point calculations for the anions and their radicals using a self-consistent reaction field method to get the corresponding vertical free energy [49]. All of the considerations above are also valid for the case of a metal salt dissolved in an ionic liquid. Nakajima et al. considered the decomposition of the trifluoromethanesulfonate anion [OTf], in the context of aluminum corrosion in lithium ion batteries. As a result of the electrochemical oxidation of [OTf], C–F active species like CF2 emerge, which either lead directly to corrosion of the aluminum or to a disproportionation reaction that forms atomic carbon which also corrodes the aluminum. Additionally, S–O-containing species are created during the oxidation of [OTf]. The authors were able to confirm their suggested decomposition products by energy dispersive X-ray (EDX) spectra. These findings should also be valid for the corresponding bis(trifluoromethanesulfonyl)amide, (CF3 SO2 )2 N− [NTf2 ], and tris(trifluoromethanesulfonate)methide, (CF3 SO2 )3 C− (CTf3 ), respectively, since they consist of comparable building substructures [50]. Witkamp and coworkers investigated the reductive decomposition of 1-butyl1-methylpyrrolidinium bis(trifluoromethanesulfonyl)amide, [BMP][NTf2 ], and 1butyl-3-methylimidazolium tetrafluoroborate, [BMIM][BF4 ], by a combination of simple and inexpensive semi-empirical calculations (Spartan ‘04 modelling program, PM3) and experiments, where a voltage (8 V) larger than the electrochemical windows of the considered room-temperature ionic liquids was applied at room temperature for 3 h [51]. Subsequently, the degradation product of [BMP][NTf2 ] was analysed via gas chromatography and mass spectroscopy (GC-MS) as well as nuclear magnetic resonance (NMR) spectroscopy, whereas for [BMIm][BF4 ] only NMR spectroscopy was used. In general the cations were reduced more easily than the anions on the cathodic limit. One exception is the heptachloroaluminate (Al2 Cl7 − ) anion of the acidic chloroaluminate ionic liquid family. After electron transfer from the electrode to the cation the obtained radical can undergo several possible decomposition and rearrangement pathways. Witkamp et al. calculated the energies of all conceivable breakdown products. The main pathway was then found by comparison of the several product energies. After formation of the analogue radical of 1-butyl-1-methyl-pyrrolidinium, it can decompose into methylpyrrolidine and a butyl radical, whereupon the energy of the products amounts to −61 kJ mol−1 in vacuum (see Figure 10.5, Eq. (2)). A second possible product represents the dibutylmethylamine radical (E = −43 kJ mol−1 ) resulting from a ring opening reaction (see Figure 10.5, Eq. (1)). The third and least likely product combination is butylpyrrolidine and a methyl radical

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

272 10 Plasma Electrochemistry with Ionic Liquids

Fig. 10.5 Decomposition pathways of 1,1-butylmethylpyrrolidinium according to Ref. [51].

(E = −21 kJ mol−1 ) (see Figure 10.5, Eq. (3)). For all decomposition pathways Witkamp et al. found experimental evidence. In the case of the 1-butyl-3-methylimidazolium cation a stable radical is obtained [50], due to the stabilizing interaction of the singly occupied p orbital of the C2 carbon atom with the p orbitals of the free electron pairs of the two adjacent nitrogen atoms. That is why the decomposition pathway would need 75 kJ mol−1 [51]. A dimerization needs only an energy of 33 kJ mol−1 , whereupon two 1-butyl-3methylimidazolium radicals are coupled to each other via their C2 atoms of the ring system (see Figure 10.6, Eq. (4)). Another reaction could be a disproportionation, i.e. a hydrogen abstraction from one radical to another, leading to 1-butyl-3-methyl2,3-dihydro-1-H-imidazole and a zwitterionic structure (see Figure 10.6, Eq. (5)). Finally, they suggest that a radical addition of two imidazolium radicals to the C–C double bond of the respective partner radical could take place, forming a cage-like,

Fig. 10.6 Decomposition pathways of 1-butyl-3-methylimidazolium according to Ref. [51].

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.4 The Stability of Ionic Liquids in Plasma Experiments 273

Fig. 10.7 Decomposition pathways of 1-butyl-3-methylimidazolium via a biradical transition state.

neutral structure, where the former two independent imidazolium radicals are connected to each other via two new bonds (see Figure 10.6, Eq. (6)). But the two radicals cannot form a product like the suggested one where the two double bonds remain. After a radical addition of one imidazolium to the C–C double bond of a second one, a biradical results. This biradical could react with the remaining double bond to give a cage structure but with three bonds formed instead of only the two proposed by Witkamp et al. and a concomitant vanishing of the two double bonds (see Figure 10.7, (7)). Compound (7) would be highly strained, hence, very unlikely. An internal rearrangement followed by a recombination of the two radical centers of the biradical could be another possible product (see Figure 10.7, (8)) but, as it is known, a transfer of a saturated alkyl group is also very unlikely to occur. Thus, they found evidence only for the first two reaction pathways in the NMR spectra of the decomposition of 1-butyl-3-methylimidazolium tetrafluoroborate [51]. In the case of metal deposition at the ionic liquid|plasma interface two possible reduction processes can conceivably take place. First, the metal cations of the dissolved metal salt can be reduced. Secondly, the cations of the ionic liquid can be reduced to neutral radicals, which can further react as described by Witkamp and as summarized above. As a first guess of which process is preferred, the rate constants of the reaction for example of silver ions (k ≥ 3.2 × 1010 L mol−1 s−1 ) and imidazolium ions (k ≤ 4.3 ×109 L mol−1 s−1 ) with hydrated electrons, taken from the data collection of Buxton et al., can be considered [52]. Thus, as long as sufficient silver ions are still in solution the reduction of the imidazolium cations of the ionic liquid represent the minor reaction pathway and the ionic liquid should not decompose significantly. How can disintegration of the ionic liquids be avoided or reduced? The cations of the ionic liquid have to be stabilized, e.g. via delocalisation of the positive charge, so that they are less eager for electron uptake in comparison to the dissolved metal

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

274 10 Plasma Electrochemistry with Ionic Liquids

salt cations. Once a reduction of the cation of the ionic liquid occurred, it would be advantageous if the radical undergoes a quenching reaction with a metal ion, i.e. an electron transfer, instead of decomposing or forming a dimer. To prevent oxidation of the anions a metal salt should be used with an anion that is much easier to oxidise than the anion of the ionic liquid. A completely different approach might be the use of radio frequency plasma instead of a DC plasma. The ignition and sustainment of the plasma is decoupled from the application of voltages to the electrodes that are now used only for electrochemical reactions. Another method which has been proven to be quite successful is the application of an U-shaped tube in order to avoid an IR-drop over the ionic liquid (see Figure 10.2). Unfortunately, this set-up led to a large size distribution of the obtained particles but it showed that RF plasma could further improve the stability of the ionic liquids during the metal deposition process.

10.5 Plasma Electrochemical Metal Deposition in Ionic Liquids

Considering the different general reaction schemes for processes at plasma|ionic liquid interfaces, the plasma-cathodic reduction of compounds dissolved in an ionic liquid is the most obvious application. In fact, the plasma-cathodic reduction of dissolved metal salts has recently emerged as a first example of plasma electrochemical processes with ionic liquids [53–55]. Up to now deposition of the metals Ag [53, 54], Pt, Cu and Pd [55] from different ionic liquids has been tested. The experimental approach is based on previous work on processes at the interface between a solid ionic conductor and a plasma [33, 56, 57] but it can, in principle, also be directly traced back to original works on “glow discharge electrolysis” of aqueous solutions by Gubkin [1], Kl¨upfel [58] and Klemenc [59], as summarised in Section 10.3. As shown schematically in Figure 10.8, the prototype experiment represents basically a cathodic reduction of a precursor (starting material), dissolved in the ionic liquid, with free electrons from the plasma phase – driven by the external electric field. Electrons are generated in the cathode region of the plasma and are driven towards the surface of the ionic liquid, where they reduce the dissolved metal compounds. In essence, we use the free surface of the ionic liquid in contact with a plasma as the electrode interface, leading to the deposition of solid products dispersed in the ionic liquid at the surface. The minimal experimental set-up (Figure 10.9) for a DC glow discharge experiment consists of a glass tube with two electrodes, of which the bottom electrode is made of either an inert metal like Pt or of a consumable bulk metal or semiconductor, thus keeping the concentration of the electroactive cation in the ionic liquid constant (side reactions at the anode are neglected). The pressure in the reactor is controlled by adjusting the mass flow of a gas (mostly argon in the case of metal and semiconductor deposition) and by a vacuum pump. Deposition of silver metal: In order to exemplify the plasma electrochemical deposition (PECD) technique in ionic liquids we first deposited silver

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.5 Plasma Electrochemical Metal Deposition in Ionic Liquids 275

Fig. 10.8 Schematic experimental set-up for the deposition of metal nanoparticles by plasma electrochemical reduction of a metal salt dissolved in an ionic liquid at room temperature.

nanoparticles from both a AgNO3 and a AgCF3 SO3 solution in ultrapure 1-butyl3-methylimidazolium trifluoromethanesulfonate ionic liquid ([BMIM][TfO]) by the use of an argon plasma. Similar experiments with 1-ethyl-3-methylimidazolium trifluoromethanesulfonate ([EMIM][TfO]) and 1-butyl-1-methylpyrrolidinium trifluoromethanesulfonate ([BMP][TfO]) have also been successful. In the first

Fig. 10.9 DC discharge over [BMIM][TfO].

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

276 10 Plasma Electrochemistry with Ionic Liquids

experiments saturated solutions of silver nitrate in the ionic liquid were used, later we used silver trifluoromethanesulfonate (Aldrich, ≥99 %) due to its better solubility in the ionic liquids. The solutions typically contained 0.3 g of CF3 SO3 Ag in 10 ml [BMIM][TfO], that corresponds to a concentration of about 0.15 mol l−1 or a molar fraction of 0.026. A 1 × 1 cm2 platinum sheet was used as anode and a hollow platinum cylinder of about 1.5 cm height and 0.75 cm diameter was used as cathode, placed in the gas phase above the ionic liquid at a distance of typically 10 cm. The glass reactor (2.5 cm diameter) was filled with 10 ml of the silver salt solution, fully covering the anode with approximately 2 cm liquid phase. The reactor was then evacuated, and the pressure was controlled to 100 Pa (argon atmosphere). Ascending bubbles were usually observed for some minutes due to emerging gases originally dissolved in the ionic liquid. After this outgassing the electric voltage was switched on. Drawing a current of 10 mA under galvanostatic conditions (corresponding to 2 mA cm−2 ), the voltage stabilised typically at about 470 V. The formation of a small number of bubbles at the anode could be observed after some time during the electrochemical experiment. This effect was more distinct with [EMIM][TfO] as solvent. The solution of the silver salts in the ionic liquids was almost transparent and colorless before the deposition experiment was started (Figure 10.10 (a)). After the onset of the glow discharge process the following observations were made: (i) A homogeneous plasma burnt with a pale pink/blue optical emission between the

Fig. 10.10 Plasma electrochemical deposition of silver nanoparticles at the free surface of [BMIM][TfO].

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.5 Plasma Electrochemical Metal Deposition in Ionic Liquids 277

upper electrode and the surface of the ionic liquid (Figure 10.10 (b)). During the initial period of the reaction the optical emission of the plasma changed slightly, indicating a change in the plasma composition. (ii) Starting from the surface of the ionic liquid a dark cloud appeared reproducibly in the ionic liquid (see. Figure 10.10 (b) and (c)). Upon longer reaction time this dark region widened until a completely dark ionic liquid was obtained (see Figure 10.10 (d)). Thus, the product phase spreads completely across the ionic liquid. (iii) At the platinum anode we observed the formation of gas bubbles. The amount of gas bubbles corresponds to the electric current across the cell, and we assume at this point that either the NO3 − or the CF3 SO3 − anion is oxidised, liberating oxygen and/or the products suggested in Section 10.4. After 5 to 10 min of reaction time the plasma electrolysis was stopped. After some minutes the homogeneous product region started to disperse and later to sediment at the bottom of the ionic liquid. Using an ultracentrifuge, the sedimentation process could be accelerated and the liquid phase of the dispersion could be removed and replaced easily by distilled water. Using ultrasound, the sediment could be dispersed again. Several of these cleaning steps were used to remove fully any remnants of the ionic liquid, thus purifying the reaction product. Images obtained by high resolution scanning electron microscopy (HRSEM) and high resolution transmission electron microscopy (HRTEM) (Figures 10.11 and 10.12) show aggregates of particles with average sizes in the nanometer region. Energy dispersive X-ray (EDX) spectra were recorded in scanning and transmission mode, both confirming that the aggregates mainly consist of silver with traces of ionic liquid. In transmission mode, we were able to focus on single nanocrystals, thus evidencing that they consist of pure silver; in particular no oxygen impurities could be detected. This finding was supported by selected area electron diffraction (SAED) and HRTEM. The diffraction patterns recorded on the aggregates show Bragg reflections located on concentric rings. The d-values determined from the diameter of the rings are fully consistent with those of pure silver (Figure 10.12) the profile of the reflections underlines the high crystallinity of the silver particles.

Fig. 10.11 SEM image of the silver nanoparticles.

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

278 10 Plasma Electrochemistry with Ionic Liquids

Fig. 10.12 TEM image and size distribution of the silver nanoparticles.

The substance was found to consist exclusively of silver nanoparticles (for particle size distribution see Figure 10.12). Silica nanoparticles were frequently found in the product. We attribute this to sputter effects of the glass reactor walls. These sputter effects can easily be reduced by a more sophisticated design of the plasma reactor. Deposition of copper metal: Since Cu(II) is the preferred oxidation state of copper, Cu2+ salts are more stable and more available, hence, in a technical application it would be favorable to use them as starting material. We tried to reduce Cu(CF3 SO3 )2 dissolved in [EMIM][TfO], [BMP][TfO] and [BMIM][TfO] with an argon plasma (gas pressure 100 Pa) as well as with a nitrogen plasma (100 Pa), respectively. Additional experiments with Cu(CF3 SO3 )2 dissolved in [EMIM][TfO] and Ar/H2 plasmas were carried out, with the distance between the hollow cathode in the gas phase and the surface of the ionic liquid metal salt solution being 3, 45 and 100 mm. Moreover, for the 3 mm distance several experiments with different gas pressures from 50 to 500 Pa were carried out. Virtually all observed reactions proceeded in the same manner: A platinum electrode was located in the middle of the ionic liquid, and a platinum hollow electrode was placed in the gas phase above, as described for silver. After 5 minutes a light brown cloud appeared in the upper half of the ionic liquid – Cu(CF3 SO3 )2 solution. During the next 5 minutes the triple phase boundary between the glass wall, the ionic liquid and the blue–pink plasma grew darker and brown threads starting from this dark region spread down into the light brown area. Then black particles emerged at the triple phase boundary and some of them finally sank down to the bottom of the ionic liquid (see Figure 10.13). Later during the reaction, the

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.5 Plasma Electrochemical Metal Deposition in Ionic Liquids 279

Fig. 10.13 Ar/H2 plasma (3 parts Ar to 1 part H2 ) burning over Cu(CF3 SO3 )2 dissolved in [EMIM][TfO] with a brown cloud and black deposits. The distance amounts to only 4.5 cm, thus the plasma consists mostly of dark space (Faraday space).

lower half of the ionic liquid also became brown, but even after 1 h there remained a distinction between the upper and the lower half of the ionic liquid phase in terms of the brightness of the brown color. Subsequent investigation of the obtained deposit with EDX revealed that it indeed consisted mainly of carbon and the residues of decomposed ionic liquid. Only a small amount of copper was found so the question remains as to whether this is copper metal or merely enclosed Cu+ or Cu2+ . Hence, at this point we conclude that copper deposition from Cu(II) salts does not easily result in Cu(0) deposition. Why was the reaction not successful in the case of copper? Can the rate constants (k) for the reaction of ions with hydrated electrons tabulated by Buxton et al. be used again, as in Section 10.4 in the case of silver ions, to estimate whether this reduction is kinetically reasonable at all? In general the reaction of (metal) ions with hydrated electrons is significantly affected by the counter ions of the considered ions and their complexation ligands. Moreover, the rate constants are given only for the reaction of Cu(I) with a hydrated electron to Cu(0), where k amounts to 2.7 × 1010 L mol−1 s−1 , and the reaction of Cu(II) with a hydrated electron to Cu(I) with k ≥ 2.9 ×1010 L mol−1 s−1 in the neutral/acid pH range, but not for the reaction of Cu(II) to Cu(0). A two-electron process is much less likely to occur and one would expect that the rate constant of this process would be lower than the k values for the two single reduction steps mentioned. The first k value [Cu(I) → Cu(0)] suggests that Cu(I) salts could be a proper starting material. The disadvantage of Cu(I) salts is that the stable ones, like the Cu(I) halides, are inherently insoluble in ionic liquids due to their covalent bonding character which leads to a diamond analogue zinc blende (sphalerite) structure. Those Cu(I) salts which do not possess this highly polymeric structural character are very sensitive to air and moisture. However, if a Cu(I)-containing ionic liquid is made by electro-oxidation of metallic copper directly

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

280 10 Plasma Electrochemistry with Ionic Liquids

in the ionic liquid the plasma electrochemical reduction to elemental copper should be feasible. Deposition of platinum metal: In the case of platinum no solid product was found. The ionic liquid darkened more and faster the smaller the distance between the surface of the ionic liquid [EMIM][TfO] containing tetrabutylammonium hexachloroplatinate ([n-Bu4 N]2 [PtCl6 ]) and the Ar/H2 -plasma (3:1, overall pressure 100 Pa) was chosen. So far no other ionic liquid has been tested. The rate constant for the reduction of the tetrabutylammonium ion with a hydrated electron is only 1.4 × 106 L mol−1 s−1 , hence the main rival pathway for reduction of platinum(IV) is the reduction of the imidazolium ion of the ionic liquid. As in the case of copper, a suitable platinum salt – maybe made by electro-oxidation of metallic platinum in a suitable ionic liquid – has to be found. Deposition of palladium metal: It was possible to deposit palladium nanoparticles by reduction of ammonium tetrachloropalladate ([NH4 ]2 [PdCl4 ]) dissolved in [EMIM][TfO]. The product yield was only about 2.5% compared with the theoretical value and was reached after 25 min. A longer application of the plasma did not lead to a significant increase in the amount of product. The lower yield and the slower reaction process might be a sign of the much more difficult two-electron reduction process compared to the one-electron process in the case of silver. Koo et al. stated that they did not obtain any platinum nanoparticles using a plasma without H2 gas [39]. In the case of palladium, nanoparticles formed when a Ar/H2 plasma (3:1, 100 Pa) was applied and when a pure Ar plasma was used. The HRSEM (Figure 10.14) picture reveals the high homogeneity of the particles which are all about 5 nm in diameter (HRTEM, Figure 10.15), only a few are bigger but that can be neglected. In Figure 10.15 is an additional SAED picture that exhibits the diffuse rings which are typical for equally seized nanoparticles. The successful deposition of silver and palladium nanoparticles proves the applicability of the PECD concept in ionic liquids. We expect that the cathodic deposition of other elements can be run in the same way under comparable conditions from

Fig. 10.14 HRSEM picture of the obtained palladium nanoparticles.

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

10.5 Plasma Electrochemical Metal Deposition in Ionic Liquids 281

Fig. 10.15 HRTEM/SAED picture of the palladium nanoparticles. The letters in the SAED picture represent the lattice indices: a = 111, b = 200, c = 220, d = 311.

suitable starting materials. In the case of highly reactive materials like titanium a hydrogen plasma may be used in order to avoid immediate re-oxidation by residual oxygen in the plasma phase. We believe that PECD in ionic liquids represents a versatile method with a potentially broad field of applications in the synthesis of metal and semiconductor micro- and nanoparticles. It can be expected that the physical properties of different ionic liquids, electric current density, the temperature, the chemistry of the plasma phase and also the convection in the liquid phase will influence the morphology of the reaction product and, thus, may be used profitably as control parameters. Comparing this approach with previous work – except the studies on solid electrolytes – ionic liquids have two distinct advantages over aqueous or organic solvents: (i) Due to their extremely low vapor pressure ionic liquids can be used without any problem in standard plasma vacuum chambers, and the pressure and composition in the gas phase can be adjusted by mass flow controllers and vacuum pumps. As the typical DC or RF plasma requires gas pressures of the order of 1 to 100 Pa, this cannot be achieved with most of the conventional liquid solvents. If the solvent has a higher vapor pressure, the plasma will be a localised corona discharge rather than the desired extended plasma cloud. (ii) The wide electrochemical windows of ionic liquids allow, in principle, the electrodeposition of elements that cannot be obtained in aqueous solutions, such as Ge, Si, Se, Al and many others. Often this electrodeposition leads to nanoscale products, as shown e.g. by Endres and coworkers [60]. The development of methods for the reproducible and continuous production of metal and semiconductor particles with a typical size on the nanoscale is still an active field of research [61–65]. The existing synthetic methods for isolated nanoparticles can be categorised into two major groups: (i) Gas or plasma phase-based preparation from gaseous or liquid precursors, (ii) preparation of nanoparticles in

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

282 10 Plasma Electrochemistry with Ionic Liquids

liquid solution by reduction or by precipitation, often with the help of template molecules or micelles [66]. Electrochemical methods are hardly used for the preparation of isolated nanoparticles, mainly because the reaction products are usually deposited as compact materials at a solid electrode rather than as free particles. A particularly successful method is pulsed electrodeposition (PED), a well-known technique in galvano plating [67], which was introduced into nanoscience by Erb et al. [68], mainly for n-Ni deposition. This concept was expanded by Natter and Hempelmann to deposit n-Pd [69], n-Cu [70], n-Fe [71] and n-Cr [72]. They were also able to deposit alloys like for example Nix Fe1–x or Nix Cu1–x [73]. Thus all metals with E > 0 V (vs. NHE) can be electrodeposited in this way from aqueous electrolytes [74, 75]. The electrodeposition of nanocrystalline metals and nanoscale semiconductors in ionic liquids is summarized in Chapters 6 and 8 . Koo et al. have recently published results from PECD of Pt nanoparticles in an aqueous solution of H2 PtCl6 [39]. They also observe the formation of relatively small particles with a typical diameter of 2 nm. From the electrochemical point of view, water is not a suitable solvent for plasma electrochemical processes, due to its relatively high vapor pressure, even at low temperatures.

10.6 Conclusions and Outlook

The plasma|ionic liquid interface is interesting from both the fundamental and the practical point of view. From the more fundamental point of view, this interface allows direct reactions between free electrons from the gas phase without side reactions – once inert gases are used for the plasma generation. From the practical point of view, ionic liquids are vacuum-stable electrolytes that can favorably be used as solvents for compounds to be reduced or oxidised by plasmas. Plasma cathodic reduction may be used as a novel method for the generation of metal or semiconductor particles, if degradation reactions of the ionic liquid can be suppressed sufficiently. Plasma anodic oxidation with ionic liquids has yet to be explored. In this case the ionic liquid is cathodically polarized causing an enhanced plasma ion bombardment, that leads to secondary electron emission and fast decomposition of the ionic liquid. Currently only a few exploratory experimental studies have been reported, and much work has still to be done in order to explore fully the properties and characteristics of plasma|ionic liquid interfaces. Currently it is still too early to comment if technical applications will be found. From the economic point of view, both ionic liquids and plasmas are comparatively expensive media, therefore only applications which show significant advantages compared to more conventional routes will be successful. Introducing reactive gases to the plasma phase may even lead to the formation of metal or semiconductor compounds, extending the experimental possibilities even further. From the physicochemical point of view, plasma electrochemical deposition is a highly interesting interfacial phenomenon, linking plasma chemistry and

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

References

electrochemistry and utilizing nucleation under conditions far from equilibrium. A systematic investigation of this process is required in order to understand the nucleation and growth process in detail. Acknowledgments

Parts of the experimental work have been performed in close collaboration with Dr. Lorenz Kienle at MPI f¨ur Festk¨orperforschung Stuttgart and Professor Frank Endres at TU Clausthal-Zellerfeld, Germany. The support of the DFG (Priority program “Ionic Liquids”, projects Ja 648/13-1 and En 370/16-1) and the Funds of the Chemical Industry (FCI) is gratefully acknowledged.

References 1 Gubkin, J. (1887) Annal. Phys. N.F., 32, 114–115. 2 Langmuir, I. (1928) Proc. Natl. Acad. Sci. U.S.A., 14, 627–637. 3 H¨offt, O., Bahr, S., Himmerlich, M., Krischok, S., Schaefer, J.A., and Kempter, V. (2006) Langmuir, 22, 7120–7123. 4 Peppler, K., P¨olleth, M., Meiss, S., Rohnke, M., and Janek, J. (2006) Z. Phys. Chem., 220, 1507–1527. 5 Meiss, S.A., Rohnke, M., Kienle, L., Zein El Abedin, S., Endres, F., and Janek, J. (2007) Chem. Phys. Chem, 8, 50–53. 6 Torimoto, T., Okazaki, K., Kiyama, T., Hirahara, K., Tanaka, N., and Kuwabata, S. (2006) Appl. Phys. Lett., 89, 243117. 7 Boenig, H.V. (1988) Fundamentals of Plasma Chemistry and Technology, Technomic Publishing Co., Lancaster. 8 Lieberman, M.A. and Lichtenberg, A.J. (1994) Principles of Plasma Discharges and Materials Processing, John Wiley & Sons, Inc., New York. 9 Drost, H. (1978) Plasmachemie, Akademie-Verlag, Berlin. 10 Moreau, M.G. (1913) Compt. Rend., 157, 922–924. 11 Moreau, M.G. (1913) Compt. Rend., 157, 1070–1072. 12 Moreau, M.G. and Bouty, M.E. (1914) Compt. Rend., 158, 260–262. 13 Hadzifejzovic, E., Sanchez Galiani, J.A. and Caruana, D.J. (2006) Phys. Chem. Chem. Phys., 8, 2797–2809. 14 Caruana, D.J. and McCormack, S.P. (2002) Electrochem. Commun., 4, 780–786.

15 Lorenz, H., Tittmann, K., Sitzki, L., Trippler, S. and Rau, H. (1996) Fresenius J. Anal. Chem., 356, 215–220. 16 Tonks, L. and Langmuir, I. (1929) Phys. Rev., 34, 876–922. 17 Vennekamp, M. and Janek, J. (2001) Solid State Ionics, 141–142, 71–80. 18 Arrhenius, S. (1887) Z. Phys. Chem., 1, 631–648. 19 Trasatti, S. (1999) J. Electroanal. Chem., 476, 90–91. 20 S¨ollner, K. (1929) Z. Elektrochem., 35, 789–799. 21 G¨unterschulze, A. and Betz, H. (1937) Elektrolytkondensatoren: Ihre Entwicklung, wissenschaftliche Grundlagen, Herstellung, Messung und Verwendung, Krayn, Berlin. 22 Yerokhin, A.L., Nie, X., Leyland, A., Matthews, A., and Dowey, S.J. (1999) Surf. Coat. Technol., 122, 73–93. 23 Shen, I. and Chu, P.K. (2004) Trends Corrosion Res., 3, 41–57. 24 Klemenc, A. and Ofner, G. (1953) Z. Elektrochem., 57, 615–617. 25 Klemenc, A. and Kohl, W. (1953) Monatsh. Chem., 84, 498–511. 26 Couch, D. and Brenner, A. (1959) J. Electrochem. Soc., 106, 628–629. 27 Hickling, A. and Ingram, M. (1964) Trans. Faraday. Soc., 60, 783–793. 28 Klemenc, A. (1952) Chimia, 6, 177–200. 29 Brenner, A. and Sligh, J. (1970) J. Electrochem. Soc., 117, 602–608. 30 Hamilton, L.W. and Ingram, M.D. (1972) J. Chem. Soc., Faraday Trans. 1, 68, 785–796.

283

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

284 10 Plasma Electrochemistry with Ionic Liquids

31 Ogumi, Z., Uchimoto, Y., Tsuji, Y., and Takehara, Z. (1992) Solid State Ionics, 58, 345–350. 32 Ogumi, Z., Uchimoto, Y., and Takehara, Z. (1995) Adv. Mater., 7, 323–325. 33 Vennekamp, M. and Janek, J. (2005) Phys. Chem. Chem. Phys., 7, 666–677. 34 Vennekamp, M. and Janek, J. (2003) Z. Anor. Allg. Chem., 629, 1851–1862. 35 Locke, B.R., Sato, M., Sunka, P., Hoffmann, M.R., and Chang, J.-S. (2006) Ind. Eng. Chem. Res., 45, 882–905. 36 Taylor, P.R. and Wang, W. (2002) Plasma Chem. Plasma Phys., 22, 387–400. 37 Taylor, P.R., Wang, W. and Vidal, E.E. (2003) in Metallurgical and Materials Processing: Principles and Technologies, 1 (eds F. Kongoli, K. Itagaki, C. Yamauchi, and H.Y. Sohn), Minerals, Metals & Materials Society, Warrendale, pp. 977–990. 38 He, J., Ichinose, I., Kunitake, T., and Nakao, A. (2002) Langmuir, 18, 10005–10010. 39 Koo, I.G., Lee, M.S., Shim, J.H., Ahn, J.H., and Lee, W.M. (2005) J. Mater. Chem., 15, 4125–4128. 40 Ue, M., Murakami, A., and Nakamura, S. (2002) J. Electrochem. Soc., 149, A1572–A1577. 41 Egashira, M., Takahashi, H., Okada, S., and Yamaki, J. (2001) J. Power Sources, 92, 267–271. 42 Barthel, J., Buestrich, R., Gores, H.J., Schmidt, M., and W¨uhr, M. (1997) J. Electrochem. Soc., 144, 3866–3870. 43 Ue, M., Takeda, M., Takehara, M., and Mori, S. (1997) J. Electrochem. Soc., 144, 2684–2688. 44 Kanamura, K., Umegaki, T., Ohashi, M., Toriyama, S., Shiraishi, S., and Takehara, Z. (2001) Electrochem. Acta, 47, 433–439. 45 Koch, V.R., Dominey, L.A., Nanjundiah, C., and Ondrechen, M.J. (1996) J. Electrochem. Soc., 143, 798–803. 46 Guyomard, D. and Tarascon, J.M. (1995) J. Power Sources, 54, 92–98. 47 Rahner, D. (1999) J. Power Sources, 81–82, 358–361. 48 Xu, K., Ding, S.P., and Jow, T.R. (1999) J. Electrochem. Soc., 146, 4172–4178. 49 Johansson, P. (2006) J. Phys. Chem. Lett. A, 110, 12077–12080.

50 Nakajima, T., Mori, M., Gupta, V., Ohzawa, Y., and Iwata, H. (2002) Solid State Sci., 4, 1385–1394. 51 Kroon, M.C., Buijs, W., Peters, C.J., and Witkamp, G.-J. (2006) Green Chem., 8, 241–245. 52 Buxton, G.V., Greenstock, C.L., Helman, W.P., and Ross, A.B. (1988) J. Phys. Chem. Ref. Data, 17, 513–886. 53 Meiss, S.A., Rohnke, M., Kienle, L., Abedin, S.Z.E., Endres, F., and Janek, J. (2007) Chem. Phys. Chem., 8, 50–53. 54 Zein El Abedin, S., P¨olleth, M., Janek, J., and Endres, F., (2007) Green Chemistry, 9, 549–553. 55 P¨olleth, M., unpublished results. 56 Janek, J. and Rosenkranz, C. (1997) J. Phys. Chem. B., 101, 5909–5912. 57 Vennekamp, M. and Janek, J. (2003) J. Electrochem. Soc., 150, C723–C729. 58 Kl¨upfel, K. (1905) Annal. Phys., 16, 574–583. 59 Klemenc, A. and Hohn, H.F. (1931) Z. Phys. Chem. A, 154, 385. 60 Endres, F. (2004) Z. Phys. Chem., 218, 255–284. 61 Fendler, J.H. (2002) Nanoparticles and Nanostructured Films. Preparation, Characterisation and Applications, Wiley-VCH, Verlag GmbH. 62 Schmid, G. (2003) Nanoparticles. From Theory to Application, Wiley-VCH, Verlag GmbH. 63 Trindade, T., O’Brien, P., and Picket, N.L. (2001) Chem. Mater., 13, 3843– 3858. 64 Weller, H. (1996) Philos. Trans. R. Soc. London, Ser. A, 354, 757. 65 Gleiter, H. (1989) Progr. Mater. Sci., 33, 223–315. 66 (1998) Nanoparticles and Nanostructured Films (ed. J.H. Fendler), Wiley-VCH, Verlag GmbH. 67 Puippe, J.-C., Leaman, F. (1986) Theory and Practice of Pulse Plating, American Electroplaters and Surfacefinishers Society, Orlando. 68 Erb, U. (1994) US patent, 5352266. 69 Natter, H., Krajewski, T., and Hempelmann, R. (1996) Ber. Bunsen-Ges. Phys. Chem., 100, 55–64. 70 Natter, H. and Hempelmann, R. (1996) J. Phys. Chem., 100, 19525–19532.

c10 (JWBG008-Endres)

January 11, 2008

9:34

Char Count=

References 71 Natter, H., Schmelzer, M., L¨offler, M.-S., Krill, C.E., Fitch, A., and Hempelmann, R. (2000) J. Phys. Chem., 104, 2467–2476. 72 Przenioslo, R., Wagner, J., Natter, H., Hempelmann, R., and Wagner, W. (2001) J. Alloys Compd., 328, 259–263. 73 Natter, H., Schmelzer, M., and

Hempelmann, R. (1998) J. Mater. Res., 13, 1186–1197. 74 Tench, D. and White, J. (1984) Metall. Trans. A, 15, 2039–2040. 75 Cheung, C., Djuanda, F., Erb, U., and Palumbo, G. (1995) Nanostr. Mater., 5, 513–523.

285

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

287

11 Technical Aspects Debbie S. Silvester, Emma I. Rogers, Richard G. Compton, Katy J. McKenzie, Karl S. Ryder, Frank Endres, Douglas MacFarlane, and Andrew P. Abbott

11.1 Metal Dissolution Processes/Counter Electrode Reactions

While the subject of this chapter may seem counter to the title of the book, metal dissolution is vital in numerous aspects of metal deposition, counter electrode processes, pre-treatment protocols and electropolishing. This chapter outlines the current state of understanding of metal dissolution processes and discusses in some detail an electropolishing process that has now been commercialised using a Type III ionic liquid.

11.1.1 Counter Electrode Reactions

Little or no information is available in the open literature about counter electrode reactions occurring during deposition processes in ionic liquids. No data exist on anodic dissolution efficiencies and hence many practical issues associated with process scale-up are unknown at present. In ionic liquids the issues associated with pH can largely be ignored since the passivating layers either dissolve, e.g. in high chloride media, or trans-passive corrosion occurs at high enough over-potentials. This means that even metals such as Cr and Al have been used as soluble anodes as they can be readily oxidised in ionic liquids. Most work to date has either used soluble anodes or has not considered the anodic reaction. A limited amount of information has been collated on the electrochemical windows of ionic liquids but this tends to be on either platinum or glassy carbon, which is not necessarily suitable for practical plating systems [1]. The anodic limits of most liquids are governed by the stability of the anion, although pyridinium and EMIM salts are sometimes limited by the stability of the cation. The widest electrochemical windows are obtained with aliphatic quaternary ammonium salts with fluorous anions. A selection of potential windows is given in Chapter 3. Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

288 11 Technical Aspects

The optimum process would ideally involve the use of soluble anodes, as the over-potential required to drive the deposition process will be small. This is especially important with ionic liquids because the ohmic loss across the cell can be significant. In aqueous solutions the use of soluble anodes is not often possible due to passivation of the electrode surface at the operating pH. While no systematic studies have been carried out, to date the only metal that we have been unable to electrochemically dissolve in eutectic-based ionic liquids is iridium. Although our research has not studied all metals we have found that even Pt, Au and Ti can be made to dissolve in eutectic-based liquids. Hence, in principle, soluble anodes could be used for the deposition of most pure metals from ionic liquids. This is, however, a considerable over-simplification and a number of factors need to be considered before employing a soluble metal anode. In ionic liquids with discrete anions attention needs to be given to the ligand present that will solvate the dissolving metal. It is highly unlikely that an unsolvated anion could exist in an ionic liquid and no evidence has been obtained to date to suggest otherwise. Metals are known to be soluble in ionic liquids based upon Tf2 N− and BF4 − anions but the nature of the metal complexes is unknown [2]. It could be that dative bonds are formed with oxygen or fluorine moieties or it could be that trace water acts as a ligand. In eutectic-based ionic liquids, the chloride ions act as strong ligands for the oxidized metal ions, forming a range of chlorometallate anions. The free chloride ions are present in very low concentrations as they are complexed with the Lewis acidic metal ions and so the dissolution of metal ions must lead to a complex series of equilibria such as − 2+ 4 ZnCl− ↔ Zn2 Cl− 3 + Zn 5 + Zn3 Cl7

(11.1)

Therefore it can be seen that metal dissolution is easier in Lewis basic melts. The zinc and aluminum deposition processes, which are by far the most frequently studied, are almost totally reversible. Since these metals have no other stable oxidation states the deposition and dissolution processes are very efficient [3–6]. This has the distinct advantage that the composition of the ionic liquid remains constant and the process becomes the removal of metal from one electrode and its deposition on the other electrode. Graphite has been used, but it fragments following electrolysis at high overpotentials leaving a black powdered residue at the base of the cell. Glassy carbon has been used extensively in voltammetric studies, but its stability at high applied current densities has not yet been tested. While anodic dissolution of metals may be advantageous for some metal deposition processes, for others it may prove problematic e.g. for alloy deposition or for electrowinning applications. In some cases, e.g. chromium, it may be impossible to obtain electrodes because the metals are not commercially available in a suitable form from which to make electrodes. Another issue that needs to be considered is metals that exist in different oxidation states, e.g. Cr and Mn. The use of inert anodes could potentially lead to the build up of metals in a higher oxidation state. However, unlike aqueous solutions, ionic

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.1 Metal Dissolution Processes/Counter Electrode Reactions 289

liquids tend to lack strong ligands such as oxygen which can stabilise higher oxidation states and this tends to negate the potential problem. Again no information exists in the open literature but experiments carried out in our laboratory showed that this was not an issue. The type II ionic liquid choline chloride: 2CrCl3 ·6H2 O was studied for the deposition of chromium using a soluble chromium anode [7, 8]. Prolonged electrolysis was carried out over several months using the sample liquid and at the end of this period the liquid was analyzed. It was found that there was no discernible breakdown of the choline cation and no chromium species other than Cr(III) was detected. The chromium content of the liquid was approximately the same as the initial sample and the only discernible change was that the water content of the liquid had decreased, presumably due to both anodic and cathodic decomposition. The chromium rod anodes were also severely etched over the process confirming that they can act as soluble anodes. The anodic processes occurring in the ionic liquids containing discrete anions have not been well characterized. They will be extremely complex as the fluorinated anions tend to be very stable and act as poor ligands. This means that both metal dissolution and solution oxidation will be difficult. If inert anodes, e.g. iridium oxide coated titanium, are used then it is difficult to envisage what the anodic process will be and this is important to determine as the systems will have to operate at relatively high current densities. Electrolysis of the ionic liquid itself must be avoided from the obvious economic viewpoint but also from the practical perspective that most electrolytes will give off toxic fluorinated products. This is analogous to the primary production of aluminum by the Hall–H´eroult process where perfluorocarbons (PFCs) CF4 and C2 F6 are produced at the anode. In the United States, aluminum smelting is the primary source of PFC emissions [9, 10]. Hence it can be concluded that little or nothing is known about the practical issues associated with suitable anode design. With such an array of ionic liquids, metals and deposition conditions available it is impossible to make specific predictions of how all anodic materials will behave. Some general conclusions can, however, be drawn, which should be good starting points from which to design specific processes. Where possible soluble anodes should be used as these improve process efficiency and bath longevity. Decomposition of the ionic liquid should be avoided at all times as it is naturally costly to reprocess the liquid and shortens its use. Processes are in general more current efficient than corresponding aqueous systems. 11.1.2 Pre-treatment Protocol

The surfaces of metal substrates require preparation and cleaning in order to ensure adhesion and effectiveness of the finishing or coating treatment. Cleaning is also employed for the removal of oil, grease or scale from metal surfaces. Abrasive blasting, acid washes, multi-stage chemical cleaning and priming are some of the techniques used for surface preparation and cleaning [11]. Typical surface preparation and cleaning operations such as abrasive blasting are used for removal of paint, rust and scale prior to painting or refinishing. Organic solvents are used

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

290 11 Technical Aspects

for degreasing; aliphatic petroleum, aromatics, oxygenated hydrocarbons and halogenated hydrocarbons are all applied to metal surfaces [11]. Electrocleaning techniques make use of a direct, reverse, or periodic reversed electric current, in combination with an alkaline cleaning bath for the removal of soil and smut and the activation of the metallic surface. The workpiece may be set up as cathode or anode. Electrocleaning baths contain a solution with ingredients similar to those of alkaline cleaning and can be operated either at ambient temperatures or in the range 40–80 ◦ C [12]. To date no processes have demonstrated this in conjunction with an ionic liquid but there is no technical reason why this should not be possible and in cases where the substrate etching is not reversible it may be advantageous (vide infra). In principle. there is no difference between the pretreatment that a metal should undergo before immersion in an ionic liquid or in an aqueous solution. The sole difference is that the workpiece must be dry before immersion in the ionic liquid. The sensitivity of the ionic liquid to water content is dependent upon the ionic liquid. Eutectic-based ionic liquids are less sensitive to water content than liquids with discrete anions. This is thought to be due to the ability of the chloride anions in the former interacting strongly with the water molecules, decreasing their ability to be reduced. Especially with AlCl3 -based ionic liquids water has to be strictly avoided. Most pre-treatment protocols studied so far follow the aqueous protocol quite closely. Good adhesion is obtained by degreasing in a chlorinated solvent, followed by an aqueous pickle in aqua regia, a water rinse and drying. A typical pre-treatment protocol is shown schematically in Figure 11.1. Several groups have then used an anodic etch in the ionic liquid prior to deposition. Anodic etch potentials and times are dependent on the substrate and the

Fig. 11.1 Flow chart for the pretreatment of substrates before electrodeposition in ionic liquids

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.1 Metal Dissolution Processes/Counter Electrode Reactions 291

Fig. 11.2 AFM image of aluminum etched for 20 s at 10 V in a Type III eutectic of 1 choline chloride: 2 ethylene glycol (right side of image masked during experiment).

ionic liquid used but generally less than 1 min is required to achieve a suitably etched substrate. The etch process has the dual purpose of removing any remaining oxide film and roughening the surface to act as a “key” for the coating layer. Metal oxide dissolution is easier in ionic liquids containing metal ions that are good oxygen scavengers, e.g. Type I eutectics, because the oxygen scavengers ‘mop’ up any oxygen moieties which have been generated during the etch process. Figure 11.2 shows an atomic force microscopy (AFM) image of an aluminum electrode etched for 20 s at 10 V in a Type III eutectic of 1 choline chloride: 2 ethylene glycol. The right side of the image was masked with a lacquer during the experiment which was then removed before the sample was imaged. It can be seen that the left side of the sample was significantly etched even during the short duration of the anodic pulse. Dissolution rates of between 50 and 150 µm min−1 are observed under these conditions and result in a pitted surface. The sample in the image is too well etched for practical purposes and hence shorter times or lower over-potentials should be employed. Figure 11.3 shows an analogous experiment for a copper electrode and it can be seen that significantly less metal is removed in the same period. The surface has approximately the same roughness as the original sample but there are more micro-pits on the sample, leading to a better key with the subsequently deposited film. Figure 11.4 shows that the etch rate for aluminum is almost three times that of copper under the same conditions. These figures show that in ionic liquids passivating films on electrode surfaces play a smaller role in controlling metal dissolution kinetics. The metals behave more characteristically, as would be predicted by their standard reduction potentials, i.e. metals with a more negative reduction potential are easier to etch. Many plating protocols advocate the use of a ‘flash’ step where a significantly higher overpotential is applied to ensure that the entire substrate is covered with metal before the potential is reduced to the plating potential. This has been shown to be effective in ionic liquid and significantly improves the corrosion resistance of the coatings [7].

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

292 11 Technical Aspects

Fig. 11.3 AFM image of copper etched for 20 s at 10 V in a Type III eutectic of 1 choline chloride: 2 ethylene glycol (right side of image masked during experiment).

One issue that has to be addressed is the reversibility of the dissolution and deposition of the substrate. If the dissolution of the substrate is reversible, i.e. all the metal dissolved can be redeposited, then etching in situ in the plating liquid is possible. If the substrate cannot be redeposited then the metal will clearly build up its concentration as the ionic liquid is used and this will significantly shorten

Fig. 11.4 Line traces taken through Figures 11.2 and 11.3; aluminum (dotted line) and copper (solid line).

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.1 Metal Dissolution Processes/Counter Electrode Reactions 293

the life of the bath. In this case a pre-etch in a different liquid should take place before the substrate is transferred to the ionic liquid. This is shown schematically in Figure 11.1. No literature has been published in this area but, as a rule of thumb, metals which dissolve to give complexes that have linear or tetrahedral geometries, e.g. Cu, Ag, Zn, Sn, Pb, can be reversibly deposited and etched. Those with octahedral geometries, e.g. Fe, Ni, Co and Cr, are less reversible. The exceptions to this are the very electronegative metals, most notably Al which is difficult to electrodeposit from some ionic liquids. The reversibility is also dependent upon the type of ionic liquid and the metal being deposited. Endres has shown that the adhesion of aluminum to mild steel is greatly enhanced by an anodic pulse prior to deposition. It has been shown that this alloy was formed between the steel substrate and the aluminum coating [1]. 11.1.3 Electropolishing of Stainless Steels

Electropolishing is the controlled corrosion of a metal surface to bring about a reduction in surface roughness and an increase in corrosion resistance of the components. Electropolished pieces also decrease wear and increase lubricity in engines, thus reducing a major cause of failure, and offer several other functional benefits. The first systematic study of electropolishing was carried out by Jacquet and led to a patent in 1930 [13]. The majority of studies have been carried out on stainless steel although metals such as copper, nickel and titanium have also been studied [14–16]. The current stainless steel electropolishing process is performed worldwide on a commercial scale and is based on concentrated phosphoric acid and sulfuric acid mixtures. The polishing process is thought to involve the formation of a viscous layer at the metal surface and many processes employ viscosity improvers such as glycerol. The practical aspects of electropolishing have been reviewed by Mohan et al. [17], whereas the more fundamental aspects are covered in a review by Landolt [18]. While electropolishing is an extremely successful process there are major issues associated with the technology, most notably that the solution used is highly corrosive and extensive gassing occurs during the process, which results in very poor current efficiency. As explained previously, electrodissolution in ionic liquids is a simple and efficient process, particularly in chloride-based eutectics. Type III eutectics based on hydrogen bond donors are particularly suitable for this purpose. However, it has been noted that the polishing process only occurs in very specific liquids and even structurally related compounds are often not effective. It has been shown that 316 series stainless steels can be electropolished in choline chloride: ethylene glycol eutectics [19] and extensive electrochemical studies have been carried out. The dissolution process in aqueous solutions has been described by two main models; the duplex salt model, which describes a compact and porous layer at the iron surface [20], and an adsorbate–acceptor mechanism, which looks at the role of adsorbed metallic species and the transport of the acceptor which solubilises

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

294 11 Technical Aspects

them [21]. Voltammetry and impedance spectroscopy have been used to confirm that the dissolution mechanism in an ionic liquid is different from that in aqueous acidic solutions. Preliminary results suggest that a diffusion-limited process in the viscous ionic liquid appears to be responsible for electropolishing [22]. Impedance spectroscopy has also shown that one of the main differences between the electropolishing mechanism in the ionic liquid and the aqueous solution is the rate at which the oxide is removed from the electrode surface. The electropolishing mechanism in the ethylene glycol eutectic is described in more detail in two recent publications [21, 22]. Highly polished surfaces were obtained with current densities between ca. 70 and 50 mA cm−2 with an applied voltage of 8 V. Below this current density a milky surface was obtained and above this range some pitting was observed on an otherwise bright surface. It should be noted that the polishing region was narrower than that in aqueous phosphoric/ sulfuric acid mixtures, but the current density requirements were considerably lower using the ionic liquid. In acidic solutions typical current densities are 100 mA cm−2 but much of this results in gas evolution at the anode. With the ionic liquid no gas evolution was observed, suggesting that there are negligible side reactions occurring with the ionic liquid. The current efficiency of the 1ChCl: 2EG electrolyte has been determined using coulometry and gravimetry and was found to be in excess of 90%, which is significantly higher than the aqueous-based electrolytes which are typically ∼30%. Given that the current density used for the 1ChCl: 2EG electrolyte is considerably lower than that used in the aqueous solution the slight difference in the conductivity of the two solutions does not lead to a significant difference in the ohmic loss through the solution. In fact preventing a passivating layer at the electrode surface during polishing decreases the overall ohmic resistance of the non-aqueous system. Hence the current going to metal dissolution is probably similar for the two systems, explaining why the polishing process takes approximately the same time. Analysis of the polished surface and residue left in the polishing tank showed conclusively that no dealloying of the surface took place and AFM analysis of preand post-polished samples showed that the polishing process was effective at preventing corrosion because it removed the micro-cracks from the steel surface [22]. Various pre-treatment protocols have been developed including pickling and anodic/cathodic pulses to remove the oxide films. It was apparent that different types of steel require different pre-treatments, i.e. cast pieces behave differently to rolled pieces. Significant success was achieved in electropolishing cast pieces and the finish obtained with the ionic liquid was superior to that with phosphoric acid, however, the converse was true for rolled pieces because the oxide film is thicker in the latter samples and hence slower to dissolve in the ionic liquid. Similar electropolishing experiments were carried out using different grades of stainless steel (410, 302, 304, 316 or 347) and it was found that the mechanism of metal dissolution and the oxidation potentials for the metals were very similar. The slight exception was the 410 series steel (which has no Ni, unlike the 300 series steels which have 8–14%). The 410 steel required a more positive oxidation potential to break down the oxide in the ionic liquid whereas once the oxide was removed the

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.1 Metal Dissolution Processes/Counter Electrode Reactions 295

Fig. 11.5 AFM image of a 316 stainless steel sample in which one side has been electropolished while the other has been masked with lacquer.

metal was more easily oxidised than the other grades of steel. This shows why the 410 steel was more likely to pit during the polishing process. The pitting could be reduced, however, by chemically pickling the steel with a proprietary phosphoric acid etch before electropolishing [23]. This technology was scaled-up to a 1.3 tonne plant by Anopol Ltd (Birmingham, UK). Results have shown that the technology can be applied in a similar manner to the existing technology. The ionic liquid has been found to be compatible with most of the materials used in current electropolishing equipment, i.e. polypropylene, nylon tank and fittings, stainless steel cathode sheets and a titanium anode jig. Extended electropolishing using the same solution leads to a dark green–brown solution arising from the dissolved iron, chromium and nickel. The solubility of the metals in the ionic liquid is relatively high and a dense sludge forms in the base of the tank when the saturation concentration is exceeded. The metals are present as glycolate and chloride complexes and numerous solvents have been tested to determine their efficacy at precipitating the metal salts. (See SubChapter 11.3). Water is completely miscible with the spent ionic liquid but the resulting mixture leads to a completely transparent liquid and almost all the metal complex is precipitated to the base of the cell. The water can be distilled from the mixture to leave a dry ionic liquid which has lost only ca. 15% ethylene glycol, mostly in the form of the metal complex. The residual concentration of each metal in the ionic liquids was less than 5 ppm. Hence, not only has it been demonstrated that electropolishing can be carried out in this non-corrosive liquid, but also that the liquid can be completely recycled and all of the metal can be recovered. Figure 11.6 shows a variety of stainless steel pieces electropolished using the choline-based ionic liquid.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

296 11 Technical Aspects

Fig. 11.6 A variety of pieces electropolished using a choline-based ionic liquid.

This subchapter has shown that metal dissolution processes are important to numerous aspects of metal plating, however, very few concerted studies have been made in this area. An understanding of dissolution rates and processes, together with information on the stability of oxide films in ionic liquids, is essential for the development of successful metal finishing processes.

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids

Voltammetric, electrodeposition, electrosynthetic and electroanalytical studies are carried out in room-temperature ionic liquids (RTILs) by a significant and increasing number of both industrial and academic laboratories [23–25]. Such studies, when carried out at anything other than a very empirical level, require the use of a ‘reference electrode’. The purpose of this chapter is to address the special problems this poses and their solutions. First, however, we start by considering the essential features of a reference electrode in general. 11.2.1 What is a Reference Electrode?

Consider a generalized electrode process: A ± ne → B, in which A is electrolytically converted into B at a suitable electrode. The rate at which this happens is measured by the current, I, flowing through the electrode, via. Faraday’s Law: I = nF A

(11.2)

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 297

where F is the Faraday constant (96 485 C mol−1 ), A is the electrode area, and  is the flux of species A to the electrode (mol cm−2 s−1 ) averaged over its surface. This rate of reaction is controlled by the magnitude of the electrical potential applied to the electrode of interest (often referred to as a “working” electrode). The application of this potential, self-evidently, requires the presence of at least a second electrode in the solution of interest, so that a defined potential can be applied between the two electrodes. This second electrode is known as a “reference” electrode. In order that a fixed and known driving potential is applied to the working electrode, it is a minimum requirement that the reference electrode maintains a fixed, constant potential difference between itself (M) and the electrolytic solution (S) with which it is in contact, (M – S )ref [26]. The potential difference (M − S )ref is established by means of a suitable electrochemical equilibrium being established at the surface of the reference electrode: Ox +e ↔ Red, so that the reference potential difference of interest is quantified by means of the Nernst equation [27]: (M − S )ref = a constant −

ared RT ln F aox

(11.3)

where R is the universal gas constant, T is the temperature in Kelvin, and ared and aox are the activities of the reduced and oxidized species, respectively. Examples of reference electrode systems which operate successfully in aqueous solutions include the following ‘potential determining equilibria’: 1 e− + H+  H2 2 1 e− + Hg2 Cl2  Hg + Cl− 2 e− + AgCl  Ag + Cl−

(11.4) (11.5) (11.6)

In the first example, a platinized platinum electrode is immersed in a solution of strong acid. The purpose of the platinizing procedure is to ensure the kinetics of the electrochemical processes are rapid enough to sustain the process at equilibrium. In the other two examples, a metal salt, highly insoluble in water (AgCl or Hg2 Cl2 ) is in contact with a solution containing chloride ions and the corresponding metal (Ag or Hg). The corresponding appropriate forms of the Nernst equations are: pH1/2 RT 2 ln F a H+ RT (M − S )ref = A − ln aCl− F RT (M − S )ref = A − ln aCl− F

(M − S )ref = A −

(11.7) (11.8) (11.9)

In experimental practice, the reference electrode will most likely be used in conjunction with a three-electrode potentiostat with a third electrode, a counter (or

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

298 11 Technical Aspects

auxiliary) electrode, completing the circuit so that negligible currents pass through the reference electrode. The latter feature is of crucial importance otherwise electrolysis perturbs the concentrations (‘activities’) of the desired species establishing the potential determining equilibrium and hence the quantity (M – S )ref which, under conditions of sustained current flow, is neither fixed nor constant [28]. That said, in the limit of microelectrodes, the currents passed can be sufficiently small that a two-electrode system becomes viable. In this arrangement, a single electrode can act as a reference electrode and a counter electrode. Classically, in relation to conventional solvent media, three classes of reference electrodes are recognised [29]: 1. Electrodes of the first kind: These are based on a potential determining equilibrium such as Ag+ + e− Ag or 12 Cl2 + e− Cl− where, for “cationic electrodes”, equilibrium is established between atoms or molecules and their corresponding cations in solution or, for “anionic electrodes”, their corresponding anions. 2. Electrodes of the second kind: These consist of three phases. A metal is covered by a layer of its sparingly soluble salt, and immersed in a solution containing the anion of this salt. The Ag/AgCl/Cl− and Hg/Hg2 Cl2 /Cl− electrodes referred to above are of this type. 3. Redox electrodes: In this case, an inert, non-reactive metal such as platinum or gold, is immersed in a solution containing both species contributing to a redox couple. For example in water: 12 BQ + e− + H+  12 H2 Q, where BQ is benzoquinone and H2 Q is hydroquinone or, in acetonitrile, Cp2 Fe+ + e− Cp2 Fe, where Cp2 Fe is ferrocene and Cp2 Fe+ the ferrocenium cation. 11.2.2 Essential Characteristics of a Reference Electrode

In the context of classical solvent media, Butler [30] suggests: “A satisfactory reference electrode must show one or more of the following properties: (1) have a potential stable with time, (2) return to the same potential after polarization,1) (3) obey the Nernst equation with respect to some species in the electrolyte, and (4) if it is an electrode of the second kind, the solid phase must not be appreciably soluble in the electrolyte.”

In the context of RTILs the criterion (3) raises considerable problems since the concept of activity and activity coefficients of ions is largely unexplored in such media. Accordingly, validation of the applicability of the Nernst equation in such media is a non-simple exercise, given that RTILs are likely to exhibit gross non-ideality. Rather, electrochemical measurements based on otherwise validated reference electrodes, may likely in the future provide a methodology for the study of RTIL non-ideality. 1) By ‘polarization’ is meant the application of a voltage perturbing the equilibrium potential of the electrode.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 299

Accordingly, for our present purposes, namely the identification of satisfactory reference electrodes, the pragmatic criteria of (1), (2), and (4) are pertinent, and (1) in particular is paramount since, in essence, (2) and (4) merely indicate means by which (1) might fail. Underpinning the requirement for a stable electrode potential is, of course, the need for relatively fast electrode kinetics to establish the potential determining equilibrium. To quote Ives and Janz [31]: “Exchange current densities for various kinds of metal–solution interfaces cover a range of about 10 −2 to 10 −18 A cm−2 , but the useful range for reference electrodes is normally much more restricted than this; it will be in part dependant upon the sensitivity of the measuring instrument to be used. One of the highest i0 values is for hydrogen ion discharge at platinum, which is one reason why the hydrogen electrode2) is one of the most satisfactory of all.”

11.2.3 Pseudo-reference Electrodes and Internal Redox Reference Couples

Butler [30] says: “If one is not too critical, many metal electrodes show relatively stable potentials in various electrolyte solutions.”

Accordingly, much voltammetry in non-aqueous solvents has been conducted using a ‘pseudo’-reference electrode (alternatively labelled a ‘quasi’-reference electrode) constituting, quite simply, a metal wire, most often silver or platinum. It is then expected (hoped) that the potential of the wire remains constant throughout the voltammetric experiment. This may be a realistic hope if, as Bard and Faulkner [32] point out, the composition of the bulk solution is essentially constant during the period of experimentation, as may be realized during voltammetric studies but certainly not in electrosynthetic work. When a pseudo-reference electrode is used, good practice [32] dictates that its actual potential is calibrated by measuring, voltammetrically or otherwise, the formal potential of an electrochemically reversible couple. IUPAC recommend the use of either the ferrocene/ferrocenium, Cp2 Fe/Cp2 Fe+ couple [33], alternatively, the cobaltocenium/cobaltocene Cp2 Co+ /Cp2 Co (where Cp ≡ C5 H5 ) has been suggested [34, 35]. In experimental practice, this simply involves measuring the voltammogram of either Cp2 Fe or Cp2 Co+ using the selected metal wire as the pseudoreference electrode before (and after) recording that of the species of interest in the same medium. Since the couples Cp2 Fe+ + e− Cp2 Fe and Cp2 Co+ + e− Cp2 Co are electrochemically reversible in most media and at most electrodes3) , comparison of the measurements allows redox data to be reported against either of the two 2) In aqueous solution. 3) Note that couples which show electrochemically reversible behavior at macro-electrodes may display quasi-reversibility or irreversibility at very small electrodes (ultramicro-electrodes, nano-electrodes)

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

300 11 Technical Aspects

couples. The possibility of using solid Cp2 Fe+ + e− Cp2 Fe in the specific context of RTIL voltammetry has been noted by Zhang and Bond [36, 37]. 11.2.4 Liquid Junction Potentials

Measurements of electrode potentials using reference electrodes are of two general types: those that involve liquid junctions and those that do not. An example of a cell which does not have a liquid junction is: Pt | H2 (g) | HCl(aq) | AgCl | Ag where | denotes a phase boundary. In contrast the cell Pt | Cp2 Fe, Cp2 Fe+ (CH3 CN) || AgNO3 (CH3 CN) | Ag has a liquid junction ( ) since two liquid phases of different compositions are brought into contact. The liquid phases may differ in terms of solvents and/or solutes. When liquid junctions exist, liquid junction potentials (LJPs) can arise due to differing ion mobilities across the interface, leading to charge separation and the development of a potential difference across the liquid junction. These can amount to some tens of millivolts and add a corresponding uncertainty in any voltammetric measurement. It follows that systems that avoid LJPs are generally preferable; otherwise some consideration of their likely magnitude is desirable (see below). 11.2.5 Reference electrodes in RTILs: What has been used?

Table 11.1 presents the results of a literature survey to establish which reference and pseudo-reference electrodes have been and are being used in RTILs. The structures of the constituent anions and cations are shown in Figure 11.7. It is clear that the majority of researchers favor the use of pseudo-reference electrodes but that not all take the trouble to calibrate using internal standards such as Cp2 Co+ or Cp2 Fe. In the latter case, the philosophy is nicely and honestly summarized by Welton and colleagues [38]: “The electrochemistry was performed on the neat ionic liquid. In such a set-up, with no recognised background electrolyte or redox standard, the potential vs. the platinum pseudo-reference is difficult to compare with standard potentials, however, in such unusual conditions it is the qualitative nature of the electrochemistry that is important.”

The most popular pseudo-reference electrodes are Pt or Ag wires. Other pseudoreference electrodes have employed coating, for example Pt with polypyrrole [39] or Ag with AgCl [40] (but in the absence of deliberately added solution phase Cl− ,

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 301 Table 11.1 A cross-section of the different types of reference electrodes

that have been used by various researchers in a range of different RTILs RTIL solution

Reference electrode material

Referenced to . . .

[C4 mim][PF6 ] [C2 mim][NTf2 ] [M(MePEG-bpy)2+ ][DNA][b] Various [NTf2 ] [C4 mim]Cl/AlCl3 DIMCARB [C4 mPyrr][NTf2 ] [N4,1,1,1 ][NTf2 ] [N6,2,2,2 ][NTf2 ] [C4 mim][BF4 ] [C4 mim][PF6 ] [C4 dmim][BF4 ] [C4 mim][PF6 ] [C4 mPyrr][NTf2 ] [C4 mim][BF4 ] [C4 mim][PF6 ] [C4 mim][NTf2 ] [C4 mPyrr][NTf2 ] [C4 mim][Co(CO)4 ] [C2 mim][NTf2 ] [C4 mim][NTf2 ] [C4 mim][PF6 ] [N8,8,8,1 ][NTf2 ] [C2 mim][BF4 ]

Ag wire

Cc+ /Cc[a] Fc/Fc+[a] Fc/Fc+[a] NR NR DMFc/DMFc+[a] NR

[C2 mim]Cl/AlCl3

[PP13 ][NTf2 ] [C6 dmim][NTf2 ] [C6 dmim][CTf3 ] [C6 dmim][PF6 ] [C6 dmim][AsF6 ] [C4 mim][PF6 ] [C2 mim][BF4 ] [C3 mim][BF4 ] [C4 mim][BF4 ] [C2 mim][BF4 ] [C4 mim][BF4 ] [C4 dmim][BF4 ] [C4 mPyrr][NTf2 ]

[C2 mPip][F(HF)2 ] [C4 mPip][F(HF)2 ] [C4 mPyrr][F(HF)2 ] [C2 mim][F(HF)2 ]

Ag wire Ag wire Ag wire Ag wire Ag wire

Ref.

[37] [52] [23, 25] [53] [36] [54]

[40]

Pt wire

Fc/Fc+[a] (E o = 0.3, 0.39 and 0.49 V/ Ag/AgCl) Fc/Fc+

Pt wire

NR

[45]

Pt wire

Fc/Fc+[a]

[56]

Pt wire Pt wire coated in polypyrrole

NR Fc/Fc+[a] (E o = 0.405 V/ SCE)

[38] [39]

Al wire immersed in a 1.5:1.0 acidic chloroaluminate melt (frit) Al wire in an 0.6 M solution of RTIL [C2 mim]Cl/AlCl3 (porous tip) Mg ribbon in Mg(CF3 SO3 )2 Li foil in Li+ salts

n/a

[57]

n/a

[41]

n/a n/a

[47] [48]

Saturated calomel (aq) Ag/AgCl KCl (sat., aq)

n/a n/a

[43] [44]

Ag/AgCl Na (sat., aq)

n/a

[42]

Ag wire in 0.1 M AgNO3 in RTIL [C4 mim][NO3 ] (glass frit) Ag wire in 0.05 M AgBF4 in RTIL [C2 mim][BF4 ]

n/a

[58]

n/a

[50]

Ag wire coated with AgCl

[24, 55]

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

302 11 Technical Aspects Table 11.1 (Continued)

RTIL solution

Reference electrode material

Referenced to . . .

Ref.

[C4 mim][NTf2 ] [C4 mim][BF4 ] [C4 dmim][PF6 ] [C4 mPyrr][NTf2 ]

Ag wire in 0.1 M AgNO3 in RTIL and Ag/AgCl in 0.1 M Bu4 NCl in RTIL. Ag wire in 0.01 and 0.1 M AgTf in RTIL (frit) Pt wire in 0.06 M N(n-C3 H7 )4I, 0.015 M I2 in [C2 mim][NTf2 ]

n/a

[49]

n/a

[51]

n/a

[46]

[N3,1,1,1 ][NTf2 ] [TES][NTf2 ] [TBS][NTf2 ] a

Cc+ =Cp2 Co+ , Cc=Cp2 Co, Fc=Cp2 Fe, Fc+ =Cp2 Fe+ , DMFc=(C5 Me5 )2 Fe, DMFc+ =(C5 Me5 )2 Fe+ . M= Fe, Co and MePEG-bpy = 4,4’-(CH3 (OCH2 -CH)OCO)-2,2’-bipyridine). NR = no calibration vs. internal reference reported. All RTIL structures are given in Figure 11.7.

b

although some may arise locally from dissolution of AgCl). In both of these cases, the Cp2 Fe/Cp2 Fe+ couple was used as an internal reference for the purposes of calibration. Another type of apparently “pseudo”-reference electrode involves the use of Al wires in contact with solutions containing AlCl4 − ions [41]. A further group of researchers simply use conventional aqueous solution-based calomel or silver/silver chloride/aqueous chloride ion reference electrodes [42–44]. These are included in Table 11.1 for illustration and completeness. The use of such electrodes is highly likely to lead to the introduction of water into the RTIL system in contact with the reference electrode, as well as to unknown problems in respect of LJPs. Properties such as voltammetric windows, diffusion coefficients and RTIL viscosity are all likely to be highly sensitive to trace amounts of water [45]. The following systems, in contrast to the above, are based on well-defined potential-determining equilibria established within a RTIL. 3 − − 1. The iodide/tri-iodide system: 12 I− 3 + e  2 I has been used by Matsumoto et al. [46]. The electrode was prepared by dissolving 60 mM N(n-C3 H7 )I and 15 mM I2 in 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)imide [C2 mim][NTf2 ] and placing a platinum wire in the solution. It is highly likely, but not explicitly reported, that such a reference electrode was used to study the voltammetry in various RTILs based on triallylsulfonium cations. If so, there would be an unknown, but probably not too large and reasonably constant, liquid junction potential between the RTIL under study and the reference electrode cell. 2. The couple 12 Mg2+ + e−  12 Mg, with the cation present as the salt Mg(CF3 SO3 )2 (1 M) has been used as a reference electrode in N-propyl-N-methylpiperidinium bis(trifluoromethylsulfonyl)imide [C3 mPip][NTf2 ] [47]. The authors also considered the use of a magnesium ribbon as a pseudo-reference electrode. The RTIL

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 303

Fig. 11.7 Structures of all RTILs listed in Table 11.1.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

304 11 Technical Aspects

in the reference electrode and in the bulk solution used for voltammetry were the same so that liquid junction potentials were relatively minimized. 3. The couple Li+ + e− Li has been used for RTILs based on [1, 2-dimethyl-3propylimidazolium] [X]− (where [X]− = [NTf2 ]− , [CTf3 ]− , [PF6 ]− and [AsF6 ]− ) [48]. The electrode took the form of Li foil in the same ionic liquids to which was added 0.02 M LiAsF6 , LiPF6 , Li[NTf2 ] or Li[CTf3 ] according to the nature of the anion in the RTIL of interest. Again, this arrangement led to a minimization of liquid junction potentials. 4. Josowicz et al. [49] have developed a reference electrode for use in RTILs based on the equilibrium AgCl + e− Cl− + Ag, in which a chlorinated silver wire is placed in a solution of 0.1 M Bu4 N+ Cl− in the RTIL of interest. The latter solution was separated from the sample under study by a double junction arrangement in which a further compartment contained only the RTIL of interest. 5. Several researchers have used the following potential determining equilibrium: Ag+ + e− Ag as the basis for well-defined reference electrodes. Josowicz et al. [49] dissolved 0.1 M AgNO3 in the RTIL of interest and inserted a silver wire. This was used in a similar double junction arrangement as described above. Hagiwara et al. [50] used 0.05 M AgBF4 in the ionic liquid 1-ethyl-3-methylimidazolium tetrafluoroborate [C2 mim][BF4 ]. Finally, Snook and colleagues [51] devised and voltammetrically characterized a Ag/Ag+ reference electrode which incorporated a known concentration (usually 10 mM) of silver trifluoromethanesulfonate (AgTf; Tf = CF3 SO3 − ) in 1-butyl-1-methylpyrroloidinium bis(trifluoromethylsulfonyl)imide [C4 mPyrr][NTf2 ]. A stable and reproducible potential was reported. In a careful and thorough study, the electrode Ag/Ag+ (10 mM AgTf, [C4 mPyrr][NTf2 ]) was found to be stable to within a millivolt over a period of around three weeks, when used in an argon atmosphere at room temperature. This is a highly important and useful observation since the characterization of the RTIL-based reference electrode (see Section 11.2.2) was significantly more rigorous than in any other study of which the present authors are aware. Specifically, for a high concentration of Ag+ , close to Nernstian behavior was seen and measurements showed the electrode to be significantly more stable than a Ag pseudo-reference electrode, even when the latter was separated by a salt bridge. Above all, voltammetric data recorded in a range of ionic liquids against the Ag/Ag+ (10 mM AgTf, [C4 mPyrr][NTf2 ]) reference electrode showed apparent liquid junction potentials of no more than a few tens of millivolts. 11.2.6 Recommendations and Comments

It is evident from the previous section that a range of approaches have been, and can be, adopted by experimentalists wishing to conduct voltammetric or other studies. The aim of this section is to answer some likely questions.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 305

11.2.6.1 When and How Can I Use a Pseudo-reference Electrode in Voltammetry? The use of a Pt or Ag wire as a pseudo-reference electrode is attractive because of its sheer simplicity and the fact that possible contamination of the test solution is avoided. The issue as to whether this provides a stable reference potential is an important consideration. To illustrate this, the electrochemistry of 5 mM Ferrocene (Cp2 Fe) in the RTIL [C4 mim][PF6 ] was studied on a platinum microdisk electrode (d = 10 µm). The pseudo-reference electrode used in this set-up was simply a platinum wire inserted into a glass tube (standard non-aqueous reference electrode kit from BAS) in the same RTIL [C4 mim][PF6 ], separated from the main solution via a Vycor plug. The reference electrode was, prior to recording the voltammetry of Cp2 Fe, pre-oxidized for different times by holding the potential at ca. +1.75 V in blank [C4 mim][PF6 ] vs. a silver wire pseudo-reference. It may be possible that the pre-oxidizing experiment deposits a layer of some species on the Pt wire, leading to significant shifts in potential. Figure 11.8 shows this effect: with no pre-oxidizing (a), the half-wave potential of Cp2 Fe is +0.275 V, which systematically shifts to more negative potentials with increased pre-oxidizing time (+0.255 V for 5 min (b), +0.225 V for 10 min (c), +0.185 V for 20 min (d), and +0.165 V for 40 min (e)). The Pt wire that had been pre-oxidized for 20 min was then left in air for a further 1 h, after which the half-wave potential of Cp2 Fe had shifted back to a potential (+0.245 V (f)) close to that observed with no pre-oxidizing. The same experiments were repeated with a silver wire inside the reference compartment, and the results are shown in Figure 11.9. Here, although there was no systematic shift in peak potentials with pre-oxidizing time: (+0.385 V for 0 min (a), +0.365 V for 5 min (b), +0.415 V for 10 min (c), and +0.385 V for 20 min), there was still a significant difference in the half-wave potential of Cp2 Fe under different conditions. It is clear that Pt or Ag wires can show significant drift (which depends in part on their recent history as well as the solution in which they are immersed) and that if such pseudo-reference electrodes are used, the regular internal calibration using Cp2 Co+ or Cp2 Fe, as advocated by IUPAC [33–35] and Zhang and Bond [37], is essential if anything other than the qualitative data is sought. Other redox couples with Nernstian characteristics may also be suitable. Examples might include: (i) The benzoquinone/benzoquinone radical anion couple (BQ/BQ•− ):

(ii) the N,N,N  ,N  -tetramethylphenylenediamine radical cation / N,N,N  ,N  tetramethylphenylenediamine couple (TMPD•+ /TMPD):

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

306 11 Technical Aspects

Fig. 11.8 Cyclic voltammograms for the oxidation of 5 mM ferrocene in [C4 mim][PF6 ] on a platinum microelectrode (diameter 10 µm) at 100 mV s−1 . Reference electrode was a Pt wire inserted into [C4 mim][PF6 ]

contained in a glass tube, separated by a Vycor frit. Pre-oxidation of the reference electrode took place for (a) 0 min, (b) 5 min, (c) 10 min, (d) 20 min, (e) 40 min and (f) 20 min with 1 h ‘rest’.

Figures 11.10 (a) and (b) show that the voltammetry of these couples in a range of RTILs is nearly electrochemically reversible. Note however that, unlike the ferrocene- and cobaltocenium-based couples, the reduction potentials are likely to vary significantly from one RTIL to another. In experimental practice it is also important to verify that the calibration molecules do not interfere chemically with the voltammetric process under study. For example, we have investigated the oxidation of molecular hydrogen in the presence of TMPD and observed a reaction of the two species, as noted by the disappearance of the reverse-peak of the first redox couple (see Figure 11.11). This implies that the peak potentials of TMPD•+ /TMPD are no longer obvious, and that this redox couple cannot be used as an internal reference in this type of experiment.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 307

Fig. 11.9 Cyclic voltammograms for the oxidation of 5 mM ferrocene in [C4 mim][PF6 ] on a platinum microelectrode (diameter 10 µm) at 100 mV s−1 . Reference electrode was a Ag wire inserted into [C4 mim][PF6 ] contained in a glass tube, separated by a Vycor frit. Preoxidation took place for (a) 0 min, (b) 5 min, (c) 10 min, (d) 20 min.

11.2.6.2 How Do I Conduct an Electrosynthetic Experiment under Potential Control? In this case, since the aim of the experiment is the bulk concentration of the material being electrolyzed, then any attempts to maintain a fixed potential using a pseudo-reference electrode will likely be hopeless. A properly defined and wellcharacterized electrode is essential and the present authors consider that described by Snook et al. [51] to be very probably the best currently available. Note that

Fig. 11.10 Cyclic voltammograms for (a) the reduction of 12.5 mM benzoquinone (BQ) in [C4 mim][NTf2 ] on a platinum microelectrode (diameter 10 µm) at 100 mV s−1 and (b) the oxidation of 20 mM N,N,N ,N -tetramethylphenylenediamine (TMPD) in [C4 dmim][NTf2 ] on a platinum electrode (diameter 10 µm) at 4 V s−1 .

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

308 11 Technical Aspects

Fig. 11.11 Cyclic voltammetry of 20 mM TMPD in [C4 dmim][NTf2 ] on a platinum electrode (diameter 10 µm) at 100 mV s−1 in the presence of 0% and 100% hydrogen.

the electrode can be constructed in the form of a separate probe as shown in Figure 11.12. 11.2.6.3 What Options Are Available for Rigorous, Quantitative Voltammetry? For most voltammetric purposes, the Ag/Ag+ (10 mM AgTf, [C4 mPyrr][NTf2 ]) electrode discussed above can be recommended as a general, stable and well characterized reference electrode although issues of the possible photo-instability of AgTf in daylight may need to be addressed in some applications. If an electrode of this type is introduced into a RTIL other than [C4 mPyrr][NTf2 ], the most likely source of error will occur from liquid junction potentials at the [C4 mPyrr][NTf2 ]/RTIL interface. As Snook and co-workers [51] point out, these may amount to a few tens of millivolts, but probably no more. It follows that for the more rigorous work, it is worth developing reference electrodes which minimize the liquid junction potentials. This is probably best achieved by using the RTIL under study as the solvent in the reference system. Building on published experiments, the latter is probably most securely based on the Ag/Ag+ system. Thus, for example, in an RTIL in which the anion is [BF4 ]− ,

Fig. 11.12 Outline of components of Ag/Ag+ reference electrode, and the reference electrode inserted into a salt bridge compartment.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.2 Reference Electrodes for Use in Room-temperature Ionic Liquids 309

Fig. 11.13 (a) Cyclic voltammograms for the reduction of 84 mM AgTf in [C4 mPyrr][NTf2 ] on a platinum microelectrode (diameter 10 µm) at scan rates of 200, 400, 700 mV s−1 , 1, 2, 4, 7 and 10 V s−1 . The pseudo-reference electrode

used was a silver wire. (b) Cyclic voltammetry for the reduction of 84 mM AgTf in [C4 mPyrr][NTf2 ] on silver wire (diameter 0.5 mm) at 10 mV s−1 with Ag/AgNO3 reference electrode as in Ref. [58].

the Ag+ could most beneficially be introduced as AgBF4 (as in Ref. [50]). Similarly, in [NO3 ]− -based RTILs, AgNO3 might be a recommended source of Ag+ . We note that the following Ag salts (with anions corresponding to common RTIL anions) are commercially available from Aldrich: AgTf (silver trifluoromethanesulfonate), AgNTf2 (silver tri(fluoromethylsulfonyl)-imide), AgBF4 (silver tetrafluoroborate), AgPF6 (silver hexafluorophosphate), AgNO3 (silver nitrate), AgCl (silver chloride), AgMeSO4 (silver methanesulfonate), AgSCN (silver thiocyanate), AgHF2 (silver hydrogenfluoride), AgAc (silver acetate), AgTFA (silver trifluoroacetate). In Cl− -based RTILs the Ag/AgCl/Cl− system can be used to generate a reference electrode relatively free of liquid junction potentials [49]. Finally, in generating new reference electrodes based on the Ag/Ag+ system, it is worthwhile pointing out that the electrode kinetics of this system are certainly unexplored in almost any RTIL medium. Prudence dictates that some brief study of the aspect precedes any application of newly developed reference systems. Usually, a voltammogram (recorded against a pseudo-reference electrode!) will suffice to show that the Ag/Ag+ couple does or does not possess sufficiently fast (“Nernstian”, “reversible”) electrode kinetics. Figure 11.13(a) illustrates the concept in respect of Ag metal deposited on a Pt microelectrode (d = 10 µm) from AgTf in [C4 mPyrr][NTf2 ]. The relative closeness of the peaks suggests quasi-reversible electrode kinetics and hence that a likely satisfactory reference electrode system can be based on Ag/AgTf in the RTIL of interest. In essence, this approach is equivalent to the micro-polarization test advocated by Ives and Janz in their classic text [31]. Figure 11.13(b) shows similar data to Figure 11.13(a), except performed using a Ag wire electrode; the near lack of hysteresis confirms the near electrochemical reversibility of the system and hence the validation of the system as the basis of a reference electrode, as advocated by Snook et al [51].

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

310 11 Technical Aspects

11.3 Process Scale Up 11.3.1 Introduction

Although the deposition of metals from ionic liquids has been possible for over 50 years, to date no processes have been developed to a commercial scale. There are numerous technical and economic reasons for this, many of which will be apparent from the preceding chapters. Notwithstanding, the tantalizing prospect of wide potential windows, high solubility of metal salts, avoidance of water and metal/water chemistry and high conductivity compared to non-aqueous solvents means that, for some metal deposition processes, ionic liquids must be a viable proposition. To assess the issues that need to be addressed before commercialization can be implemented it is probably easier to analyse the current and future markets for electroplating and compare the limitations of the current technology for various metals. The main metals of interest are Cr, Ni, Cu, Au, Ag, Zn and Cd, together with a number of copper and zinc-based alloys. The electroplating industry, which dates back well over 100 years, is based, naturally, on aqueous solutions due to the high solubility of electrolytes and metal salts resulting in highly conducting solutions. Water does, however, suffer from the drawback that it has a relatively narrow potential window and hence the deposition of electronegative metals such as Cr and Zn is hindered by poor current efficiencies and hydrogen embrittlement of the substrate. In addition there are specific difficulties with certain metals. The most obvious case is that of chromium plating. The major disadvantage of the current process of chrome plating is that it requires the use of chromic acid-based electrolytes comprising hexavalent chromium, Cr(VI). The toxicity and carcinogeneity associated with Cr(VI) [59] has resulted in wide-ranging environmental legislation in the USA (OSHA, EPA) and Europe (IPPC) to reduce its use. For example, the EU End-of-Life Vehicles (ELV) Directive aims to ban the use of Cr(VI) in the manufacture of vehicles, although limits of 2 g per car are to be permitted for the foreseeable future. In addition, the Directive on Waste, Electrical and Electronic Equipment (WEEE) aims to ban the use of Cr(VI). In the US, compelling health data and legal suits are forcing OSHA regulators to lower the exposure limit to chromic acid and it is anticipated that future exposure limits could be established at levels between 20- and 200-fold below the current level. Past work carried out in the US and UK has generally examined the viability of reducing emissions of chromic acid (air pollution control techniques and chemical fume suppressants) rather than applying fundamentally novel chemistries for chrome plating [60]. However, environmental and social pressures of operating chromic acid-based processes are imposing demands upon the industry, which cannot be met through effluent reductions alone. In answer to this, at least three types of aqueous trivalent chromium baths have been developed industrially [61–63]. However, finish quality,

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.3 Process Scale Up 311

cost and a perceived difficulty of operation has hindered the general acceptance of these commercially available baths. Ionic liquid-based processes could provide a route to overcome these problems. Other disadvantages of the existing aqueous technology are economic in nature, such as the low current efficiency of the reduction of Cr(VI) in acid media. In addition, the difference in over-potential between chromium and hydrogen reduction results in the evolution of hydrogen gas, which can lead to hydrogen embrittlement in the substrate. A more general issue associated with aqueous solutions is that at some point all water must return to the watercourse and hence the contamination with toxic metal salts e.g. Cd (II) or Ni (II), complexing agents e.g. CN− or brighteners must be minimized. Hence, while the use of ionic liquids to replace aqueous technology may not seem to have an urgent technological or economic driver there are numerous circumstances where the use of ionic liquids has specific advantages, such as the deposition on passivated substrates, e.g. Al, or the efficient deposition of specialist alloys that could not be carried out in aqueous solutions because of the chemistry of water. It is most probable that practical plating liquids for Cr, Ni, Cu, Au, Ag and Zn will use eutectic-based ionic liquids. Numerous Zn-based Type I eutectics have been applied to small scale deposition studies but it is less likely that these will be viable due to the comparatively low conductivities and high viscosities of these liquids. Several companies are currently using Type III-based eutectic ionic liquids, primarily those with urea and ethylene glycol as the hydrogen bond donor, to electrodeposit zinc and zinc-based alloys [64]. This is at the 10–25 l scale using soluble zinc anodes. High current efficiencies can be obtained at low current densities but the morphology and current efficiency deteriorate as the current density increases. The main technological difficulty associated with the further scale up of these plating baths is the development of effective brighteners that function in ionic liquids. Outside of these seemingly niche markets the main driving force for using nonaqueous electrolytes has been the desire to deposit refractory metals such as Ti, Al and W. These metals have numerous applications, especially in the aerospace industry, and at present they are deposited primarily by PVD and CVD techniques. The difficulty with using these metals is the affinity of the metals to form oxides. All of the metal chlorides hydrolyze rapidly with traces of moisture to yield HCl gas and hence any potential process will have to be carried out in strict anhydrous conditions. Therefore the factor most seriously limiting the commercialization of aluminum deposition is the engineering of a practical plating cell. Notwithstanding the perceived difficulties with commercializing such technology, a commercial aluminum electroplating process is already in existence and has operated for over 10 years [65, 66]. It is based on triethylaluminum in organic solvents such as toluene and although precise technical details are not given in the open literature it is apparent that the process is successful. It is also highly probably that a plating bath based upon a chloroaluminate ionic liquid is less water sensitive than the organics solution.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

312 11 Technical Aspects

11.3.2 General Issues

There are several general issues where ionic liquids differ from aqueous solutions. Some of these are discussed in greater detail in the preceding chapters and all are discussed in more detail in a recent review [67]. The key issues are clearly associated with developing non-aqueous processing protocols and accounting for the differences between the physical properties of a non-viscous polar fluid and a viscous ionic liquid. 11.3.2.1 Material Compatibility In general, ionic liquids tend to be non-corrosive towards most metallic and polymeric materials that would normally be encountered in electroplating or electropolishing situations so there is no reason why they could not be simple “drop-in” replacements for aqueous systems. The majority of plating plants are constructed from polymers such as polyethylene, polypropylene, nylon and PVC, all of which are stable in the majority of ionic liquids. The only large scale tests that have been carried out using ionic liquids were in collaboration between Anopol Ltd and the University of Leicester for the electropolishing of stainless steel. Figure 11.14 shows a 1.3 m3 tank that was constructed from polypropylene with polypropylene, nylon and polyethylene fittings and run as a pilot plant. It has a standard 3 kW heater to maintain the liquid at 50 to 60 ◦ C. Tank agitation was achieved by recirculation of electrolyte via eight banks of inductor nozzles [68]. As with aqueous solutions, “dip coatings” can be obtained when more electronegative metals are placed in ionic liquids containing more electropositive metal ions e.g. silver ions will be deposited onto copper metal. Unlike aqueous solutions,

Fig. 11.14 Electropolishing bath (1300 L) operating at Anopol Ltd. (Birmingham, UK) based on an ethylene glycol:choline chloride eutectic.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.3 Process Scale Up 313

however, these dip coatings tend to be more adherent. It should also be noted that the redox potentials of some metals can be significantly shifted from the standard aqueous redox potentials due to the differences in metal ion speciation. The main differences will occur with the design of baths suitable for aluminum and other water-sensitive metal salts. The evolution of HCl will require materials which are more corrosion resistant, and the main difficulty will be in the development of plants which will allow the transfer of pieces in and out of the liquid under strictly anhydrous conditions. 11.3.2.2 Pre-treatment Protocols The aim of any pre-treatment protocol is clearly to remove non-metallic detritus from the surface and will naturally involve a wash with a solvent to remove organic residues and an acidic or alkaline clean to dissolve inorganic residues. Chapter ? discusses the different approaches that can be used but, in principle, these are the same that are currently employed with standard aqueous electroplating baths. The key issue is to introduce a dry substrate into the ionic liquid and this will involve either a drying stage or a rinse in an ionic liquid prior to immersion in the plating liquid. Several methods have been studied but by far the best adhesion is obtained by degreasing in a chlorinated solvent, followed by an aqueous pickle, rinse, dry and then anodic etch in the ionic liquid prior to deposition. Anodic etch potentials and times are dependent on the substrate and the ionic liquid used. Metals such as Al and Mg will require a larger anodic pulse for a longer period than other metals such as Cu or Ni. Metal oxide dissolution is easier in ionic liquids containing a metal that is a good oxygen scavenger. Endres has shown that the adhesion of aluminum to mild steel is greatly enhanced by an anodic pulse prior to deposition. It was shown that an alloy was formed between the substrate and the coating metal, improving adhesion [69]. In situ anodic etching may not always be feasible if the substrate is difficult to re-deposit e.g. steel. In this situation a build-up of contaminating metal ions in the ionic liquid could change the physical properties of the liquid and damage the quality of the coating. Anodic etching should take place in a compatible ionic liquid and the etched substrate is then transferred to the plating tank. 11.3.2.3 Conductivity The conductivity, κ, of an ionic liquid is also strongly dependent upon temperature and in an analogous manner to viscosity it is found to change in an Arrhenius manner

ln κ = ln κ0 −

E RT

(11.10)

where E is the activation energy for conduction and κ 0 is a constant. It has been noted that the empirical Walden rule (η = constant) is applicable to ionic liquids, where  is the molar conductivity and η is the viscosity [70]. Deviations from the Walden rule have previously been used to explain ionic association in proton transfer ionic liquids [71, 72]. The Walden rule is normally only valid for ions at

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

314 11 Technical Aspects

infinite dilution where ion–ion interactions can be ignored but with ionic liquids it is the availability of holes that allow ion migration that limits charge flow. Since the fraction of suitably sized holes in ambient temperature ionic liquids is very low the movement of holes can be defined by a combination of the Stokes–Einstein and Nernst–Einstein equations [73, 74]: λ+ = z2 F e/6πη R+

(11.11)

where z is the charge on the ion, F is the Faraday constant and e is the electronic charge. This explains why so many studies of conductivity in ionic fluids have noted that the empirical Walden rule is valid. Since the Stokes–Einstein equation is valid for both ions then the conductivity of the salt can be determined since  = λ+ + λ−

(11.12)

an expression can be written for the conductivity, κ κ=

z2 F e 6πη



1 1 + R+ R−



ρ Mw

(11.13)

where ρ is the density and Mw is the molar mass of the ionic fluid. Hence all of the theories developed for limiting molar conductivities in molecular solvents are also applicable to ionic liquids where there is an infinite dilution of suitably sized holes [74]. Using this theory it is possible to estimate the limits of viscosity and conductivity that an ionic liquid can achieve. It is difficult to foresee an ionic liquid that has a conductivity significantly in excess of EtNH3 + NO3 − (ca. 150 mS cm−1 at 298 K) and this must be viewed as a probable upper ceiling without modification [75]. 11.3.2.4 Added electrolytes The conductivities of most aqueous electroplating solutions are in the region of 100 500 mS cm−1 because they are mostly high strength aqueous acids [76] and this allows high current densities to be applied with only limited ohmic loss. Significantly lower conductivities are obtained with ionic liquids and one way to increase the conductivity could be to add a small cation such as Li+ that could have better mobility compared to the large organic cation. This has been attempted by a number of groups, particularly those developing lithium ion batteries, but the effect on the conductivity has not been as significant as expected [77, 78]. The viscosity and freezing point of the liquid are, however, affected as the small cation will be strongly associated with the anions and little increase in the conductivity is generally achieved. Other salts such as Na+ and K+ have negligible solubility in most ionic liquids. The addition of electrolytes is clearly an area that requires considerable investigation in the future. The structure of the double layer is also affected by the addition of lithium ions. Few studies have been carried out on the structure of the double layer in an ionic

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.3 Process Scale Up 315

liquid but those that have tend to suggest that the models used for aqueous solutions are inappropriate in ionic liquids [79, 80]. If the metals are reduced at potentials below the potential of zero charge then the electrode must be coated with a 6–7 Å thick layer of cations. Adding small ions such as Li+ to an ionic liquid will decrease the Helmholtz layer thickness considerably and should make metal ion reduction easier. This should simplify nucleation and it has been shown qualitatively to be the case for the deposition of chromium from a eutectic mixture of chromium chloride and choline chloride. The incorporation of up to 10 mol% LiCl led to a change in deposit morphology from microcrystalline to nanocrystalline and a change in visual appearance from metallic to black [81]. It has also been shown that the addition of LiF to 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)imide allows the deposition of dense, thick, corrosion resistant coatings of tantalum [82]. 11.3.2.5 Brighteners Brighteners are essential to most electroplating systems and act to decrease the surface roughness and improve reflectivity. Brighteners are thought to function by either forming metal complexes, which shift the reduction potential and hinder metal nucleation, or by adsorption on the electrode surface blocking nucleation and hindering growth. In aqueous solution most brighteners are complex mixtures of components, many of which are derived by serendipity, but most have the function of viscosity modifiers or amphiphillic molecules that can specifically interact with the metal surface. No systematic studies have been carried out in ionic liquid using the types of brighteners used in aqueous solutions and this is clearly an area that needs to be addressed to see if the brighteners function in the same way as they do in water. The Abbott group has carried out studies using commercial brighteners for zinc plating from Type III eutectics but to date none of these have shown any improved surface finishes. To some extent this is not surprising given: Ĺ The viscosity of the ionic liquids is much higher than aqueous solutions affecting mass transport, Ĺ The double layer structure is totally different in the two liquids and hence the surface potential will differ, meaning that specific adsorption of organics will differ, Ĺ The metal speciation is different and hence the reduction potential will be shifted, Ĺ Electrode processes will be different due to the lack of proton or hydroxide ions in ionic liquids.

It may seem to be an impossible task to find a brightener compatible with an ionic liquid but comparison of practical aqueous plating solutions with current ionic liquids shows a fundamental difference in the metal speciation. In aqueous solutions most plating is carried out with either strong bases e.g. KOH for zinc plating, strong complexing agents e.g. CN− for silver plating or metals in the oxide form e.g. CrO3 for chromium plating. These will tend to shift the reduction potential to more negative values, decreasing the rates of nucleation and growth.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

316 11 Technical Aspects

Applying the same principle to ionic liquids a number of compounds containing nitrile, carboxylate and amine functionalities have been tested. Limited success has been achieved with Ni, Ag, Cu and Zn baths. Brighteners that involve a complexation with a solution-based species will depend upon the comparative strength of the ionic liquid–metal interactions. It would therefore be logical to suppose that ionic liquids with discrete anions would be likely to work directly with brighteners used in aqueous solutions, as the interaction between the metal salt and the anion will be considerably weaker than those between the metal salt and the brightener. In eutectic-based ionic liquids the chloride anions act as strong Lewis bases and could decrease the relative interaction between the metal salt and the brightener. Brighteners which rely on electrostatic or hydrophobic interactions may function in ionic liquids but their efficacy is likely to be surface and cation/anion specific. As with other solutes in ionic liquids, the general rule of like dissolving like is applicable i.e. ionic species will generally be soluble as will species capable of interacting with the anion. Aromatic species tend to exhibit poor solubility in ionic liquids consisting of aliphatic cations and vice versa. We have also studied the use of brighteners in Type III-based ionic liquids, as well as the majority of brighteners that are used in aqueous zinc plating solutions and none of them are active in ionic liquids. Some success has been achieved using complexing agents such as ethylenediamine and acetonitrile but this has not been a significant improvement. Figure 11.15 shows an AFM image of silver deposited from a ChCl:2urea eutectic. It can be seen that in the absence of any brighteners a relatively rough surface is obtained whereas the addition of ethylenediamine acts as a brightener producing a much smoother surface finish. Endres studied the use of nicotinic acid for the deposition of Pd and Al/Mn alloys from an AlCl3 -1-butyl-3methylimidazolium chloride ionic liquid and showed that, in contrast to producing a brighter surface finish, it aided the formation of nanocrystalline deposits.

Fig. 11.15 Silver coating deposited from a urea:ChCl eutectic in the absence (a) and presence (b) of ethylenediamine as a brightener.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.3 Process Scale Up 317

11.3.2.6 Counter Electrode Reactions As outlined in previous chapters the counter electrode reactions occurring in ionic liquids will be significantly different from those in aqueous solutions. Given the increased ohmic resistance that will be encountered compared to aqueous solutions it will be preferable to use soluble anodes which will decrease the over-potential that needs to be applied between the electrodes. Soluble anodes will also minimize the breakdown of the ionic liquid itself and retain the bath composition in its original state. Anodic dissolution of most metals will occur in ionic liquids due to the absence of passivating films on the electrode surface. Hence metals such as Al and Cr could potentially be used as anodic materials. While this is potentially useful it should also be noted that caution should be exercised when choosing a suitable material for jigs or connectors that will be immersed in the ionic liquid. No systematic study of inert electrode materials has taken place to date and nothing is known about the anodic processes taking place in ionic liquids. It is probable that noble metal oxide coatings should be suitable but processes such as chlorine evolution will clearly have to be avoided for eutectic-based ionic liquids. The breakdown products of most cations are unknown but it is conceivable that some of them could be potentially hazardous.

11.3.2.7 Post-treatment Protocols and Waste Treatment Treatment of the sample following electrodeposition has primarily been carried out using a simple aqueous washing procedure. While this is an extremely effective method it may not ultimately be applicable to large scale production due to toxicological issues with some of the anions or cations. Some ionic liquids have been developed with biodegradable cations and anions but the liquids will still contain large metal ion concentrations and some complexing agents which would be better to keep separate from aqueous systems. The amount of “drag-out” and the extent of the issue will depend upon the viscosity of the liquid. To circumvent the need to process large volumes of rinse water it may be more practicable to rinse the piece with a liquid that is immiscible with the ionic liquid, which will allow the separation of the ionic liquid in a settling tank. The most appropriate washing liquid will depend upon the nature of the ionic liquid and the phase behavior of most ionic liquids is well documented.

11.3.2.8 Supply The majority of aqueous plating solutions are supplied as finished products by major distribution suppliers. No such analogue exists for ionic liquids as no plating processes have been developed with a sufficiently good surface finish to replace the aqueous competitor. A number of companies make or distribute ionic liquids on the > 100 kg scale. These include BASF, Merck, Scionix and Solvent Innovation although laboratory scale amounts can now be obtained from a wide range of chemical supply houses.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

318 11 Technical Aspects

Fig. 11.16 Recycling of ionic liquid (1 ChCl : 2 ethylene glycol) used to electropolish stainless steel; (a) used liquid containing Fe Cr and Ni salts, (b) as (a) with 1 equiv. v/v added water, (c) as (b), after gravity filtration and subsequent removal of residual water by distillation.

11.3.2.9 Recycling Given the cost and environmental compatibility of most ionic liquids, recycling protocols will be essential. Many of the issues will be associated with separating the metals from the ionic liquids. The same issues also exist in aqueous solutions and they are usually addressed by the addition of concentrated base which precipitates the metals as an oxide or hydroxide. The solutions then need to be filtered and neutralized before disposal. Similar ideas will need to be developed for ionic liquids i.e. ligands that can be added to precipitate the metals. An alternative approach is to add sufficient diluent to the ionic liquid, thus changing the solvent properties such that the specific metal becomes insoluble. This idea has been applied to the recycling of the commercial electropolishing solution. The electrodissolution of stainless steel produces an ionic liquid that contains high iron, chromium and nickel concentrations. The metals are present as glycolate complexes and the addition of water renders the complexes insoluble, Figure 11.16. This has the advantage that it decreases the viscosity of the mixture and permits easier filtration. The water can be distilled from the mixture with minimal loss of the ionic component. While this process will only be applicable to a limited number of ionic liquids analogous processes should be possible using other solvents. 11.3.3 Conclusions

Although no plating processes have been developed to date using ionic liquids, it is clear that the advantages afforded by this new technology will certainly have commercial applications. There are some issues associated with process scale up but these are only analogous to the aqueous solutions and are not insurmountable. The potential high current efficiency and longevity of the ionic liquids should make the economics of the processes beneficial and with the current groundswell of interest in the area it is highly likely that the plating industry will see at least some processes entering the market within the next ten years.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 319

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating Daniel Watercamp, and Jorg Th¨oming

In electroplating, impurities can be assumed to interfere with the intended deposition, deteriorating surface qualities or narrowing drastically the potential window available. Due to their relatively high price and the anticipated cost for discharge of spent liquors a breakthrough of ionic liquids in electroplating applications can be expected to be linked to successful regeneration options. Because ionic liquids are non-volatile and typically one to three orders of magnitude more viscous than water, their regeneration by separation from mixtures and purification is a challenging task. However, it can be assumed for most electrochemical applications of ionic liquids, especially for electroplating, that suitable regeneration procedures can be found. This is first, because transfer of several regeneration options that have been established for aqueous solutions should be possible, allowing regeneration and reuse of ionic liquid based electrolytes. Secondly, for purification of fresh ionic liquids on the laboratory scale a number of methods, such as distillation, recrystallization, extraction, membrane filtration, batch adsorption and semi-continuous adsorption in a chromatography column, have already been tested. The recovery of ionic liquids from rinse or washing water, e.g. by nanofiltration, can also be an important issue. 11.4.1 Introduction

For electroplating purposes ionic liquids show several attractive properties, such as large electrochemical windows, specific solvent characteristics and extremely low vapor pressures compared to ordinary solvents. When used as base electrolytes in electroplating, ionic liquids can allow new processes that are impossible in conventional electroplating where the main solvent used is water. Despite the great electrochemical potential offered, these new compounds have to compete with water in terms of “greenness”, given that water is itself the most environmentally benign solvent. Nonetheless, an advantage of ionic liquids is that they do not evaporate, even at elevated bath temperatures, avoiding heat and mass losses during processing. However, “greenness” should not be attributed to a compound due to a single characteristic, the whole system has to be considered and every single chemical entity has to be assessed with respect to its entire life cycle. Based on the findings of a case specific analysis, a relative degree of greenness can be attributed to comparative process and compound alternatives. In principle, the degree of greenness can be determined through the following four main aspects: Ĺ Greenness of the manufacturing process of the ionic liquid Ĺ Risk potential of the technical application of the ionic liquid (leakages, toxic and eco-toxic effects, fate of the compounds in the environment) Ĺ Possibility of regeneration, recycling and reuse of the ionic liquid Ĺ Waste treatment options.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

320 11 Technical Aspects

In general, recyclability is crucial for the design of sustainable chemical processes [83]. The aspect that should be elaborated here is the possibility of regeneration and reuse of the ionic liquid, depending on the type of impurity and the sensitivity of the specific application towards contamination. Despite the huge number of publications dealing with the application of ionic liquids, there are only a couple that include reuse aspects. To the best of our knowledge, there is none that deals with regeneration of spent ionic liquid based electrolytes. The intention of this contribution is to bridge this gap and suggest potential concepts for ionic liquid regeneration. In this chapter, an introduction to the principles of regeneration as they have been developed in the field of water-based electroplating is given. With this background, a discussion of the purification options for ionic liquids is presented, followed by a first case study. 11.4.2 Recovery, Regeneration and Reuse of Electrolytes in Electroplating 11.4.2.1 The Concept A general approach towards both more economical and more environmentally benign applications of electrolytes in electroplating is the minimization of losses and purge stream optimization. Losses are caused by drag-out, i.e. electrolyte that clings to workpieces when they are removed from the plating bath. This makes subsequent rinsing of the workpieces necessary, through which the losses are diluted and discharged into the wastewater. Purge streams could be necessary as a measure for product quality assurance. This implies that, by replacing all these losses with fresh electrolyte, the so called make-up, relevant contamination can be kept below critical levels. To reduce the consumption of fresh electrolyte, diverse general approaches are possible, such as the recovery and reuse of the losses from product and wastewater streams, the recycling of spent liquid back to the manufacturing of the electrolyte and the reuse of purge streams within the plating process. Fundamentally, all these approaches require regeneration of the electrolyte prior to reuse or recycling. Ideally, the regeneration makes use of the selective separation of the minor compound, the impurity, from the electrolyte. Without such a regeneration step, impurities would accumulate and eventually interfere with the intended functions of the electrolyte, thereby reducing product quality. There are several possible sources of impurities in the electrolytes and reasons for their potential accumulation during use. Key amongst the sources, are the unavoidable side-reactions. Others include the widespread practice in electroplating processes of using the more convenient open systems that allow easier handling of workpieces. Consequently the absorption of atmospheric gases and particles might introduce impurities. The overall concept for recovery, regeneration and reuse in electroplating is shown in Figure 11.17. It includes the recovery stage, in which the workpieces are rinsed for further cleaning and the diluted electrolyte received. The diluted solution

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 321

Fig. 11.17 Concept of sustainable use of process bath liquors in electroplating: recovery (rinsing and concentration), regeneration (concentration and purification) and reuse.

is then concentrated using membrane systems [84] producing wastewater. However, there exists another strategy for avoidance of the production of wastewater and the reuse of the diluted stream in rinsing, thereby achieving zero-water discharge systems [85]. In this case, both, the concentration and the purification units can be part of the regeneration, as shown in Figure 11.17. More significantly, the design of the whole system should be subject to an optimization process with respect to sustainability aspects such as cost and wastes [86]. As shown in industrial applications, there are surface finishing systems for which the recovery of electrolytes is feasible and economically attractive [87]. 11.4.2.2 Regeneration Options for Water-based Process Liquors Chemical and electrochemical surface treatment processes such as electroplating, pickling, and etching often have a high consumption of chemicals and produce a lot of wastewater and heavy metal wastes. Consequently, cost saving and environmental compatibility lead to the necessity of applying purification and concentration units. Purification units can be divided into two groups. The first group treat spent plating solutions, while the latter treat rinsing discharges. Regenerators for Spent Process Liquors. Most effort in developing regeneration methods for water-based process liquors in metal finishing has been spent on chromium plating baths. These solutions contain a significant amount of chromium and a lesser amount of other heavy metals, which make them a significant environmental concern and obvious targets for regeneration and reuse. Typically a two-chamber electrolytic cell is applied and different electrode materials have been tested [88]. The cell allows oxidation of Cr(III) to regenerate Cr(VI) in the anode compartment. The removal of dissolved metal impurities such as Fe(II), Fe(III), Cu(II), and Ni(II) from contaminated chromic acid solutions can be performed through electrodialysis in the same two-chamber cell as the chromic acid recovery, where the impurities that electromigrated into the cathode compartment are deposited or precipitated. To achieve chemically robust low-cost separators, ceramic membranes have been suggested by Sanchez et al. [89]. A Nafion 117 membrane and a ceramic diaphragm

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

322 11 Technical Aspects

separator were compared by Huang et al. [90]. Their results indicated that a system using the Nafion separator and a small catholyte/anolyte volume ratio was best suited for removing impurities from concentrated plating solutions. Similarly, using a ceramic membrane for chamber separation, Jegadeesan et al. found up to 69% impurity removal [91]. To reduce the energy demand in such a system Huang et al. [92] modified the set-up successfully and found that the average removal rate of each impurity was approximately proportional to the product of its initial concentration and the separator area/anolyte volume ratio. More detailed investigations have been reported in Huang et al. [93]. In the case of dissolved metal as major additive compounds, a combination of precipitation and redissolution can be applied for recovery from spent solutions. Gyliene et al. [94] found, for recovery of the main additive in nickel electroless plating, that the Ni(II)-citrate complex could be precipitated with alkali followed by redissolution in citric acid for reuse in electroless nickel plating after separation of the precipitate. Additionally, for decontamination of spent electroless nickel plating solutions Fe(III) can be used to precipitate the pollutant. To simultaneously recover the metal and sulfuric acid from spent process liquors of nickel electrolysis, Xu and Yang [95] tested diffusion dialysis successfully. The membrane used was surface-cross-linked with aqueous ammonium to decrease waste volume expansion caused by the water osmosis. They could control nickel leakage within 4% and recover about 70% of the acid. Alternatively to diffusion dialysis, Pierard et al. [96] suggested electrodialysis as a regeneration process. In the case study involving acid pickling before electroplating, they demonstrated the selection of ion-exchange membrane couples as well as the development of tools to promote the use of electrodialysis in industrial applications. For removing organic compounds, adsorption could be a good choice. This applies to many decomposition products that might occur during electroplating as well as for additives such as polyethylene glycol, a major organic additive in copper electroplating solution, used as a brightening and stabilization agent in the low ppm concentration range. Chang et al. [97] reported a successful application of activated carbon, Calgon Filtrasorb 400, to remove polyethylene glycol from used electroplating solution in order to reuse it. Other unit operations that can be used in this field of process liquor treatment are evaporation and crystallization. Both were tested by Ozdemir et al. [98] for regenerating waste pickling liquors from hydrochloric acid pickling baths and are reported to be suitable for small to mid-scale plants, currently neutralizing and discarding waste pickling liquors. Even the relatively expensive crystallization process, which can be used for removal of ferrous chloride to enable the recycling of unused acid, was found to bring some improvement. Purification Units for Rinsing Solutions. The second group of purification units comprises those for treating rinsing discharges. For example, dissolved metals can be separated by applying a combination of electrodeposition and electrodialysis, as reported by Bolger and Szlag [99]. They recovered nickel from the rinse water

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 323

cathodically in an electrolytic cell separated by an anion exchange membrane. Depending on the anions used in the electrolyte such a process generates anodically a sulfuric/hydrochloric acid mixture. However, additives like boric acid that are characterized by high acid dissociation constants cannot be recovered by anion exchange. A widespread technology for purifying diluted aqueous solutions and even electroplating waste solutions is ion exchange [100]. This is also true for rinsing solutions [101, 102]. In technical systems a set of ion exchange columns is applied [85]. Usually liquid for regenerating the columns is discharged afterwards, but in some cases recovery of valuables is also possible [103]. For purification of aqueous solutions the use of adsorption processes for cationic impurities is also common. As economical adsorbents, montmorillonite, tobermorite, magnetite and silica gel were found sufficient for the removal of Cd(II), Cr(VI) and Cu(II) in rinsing wastewater from a plating factory [104]. From this investigation, it was found that the removal efficiency tended to increase with increasing pH and decrease with increasing metal concentration. This method allows the realization of a rapid, simple and cheap rinse water treatment system for the removal of heavy metals. A complete process scheme for regeneration and reuse of spent final rinse water from an electroless plating operation has been developed by Wong et al. [105]. It includes (i) pre-treatment by microfiltration, UV irradiation, carbon adsorption; (ii) heavy metal removal by nanofiltration and (iii) polishing using an ion exchange mixed bed. The results of a pilot study showed that high quality product water with an overall water recovery of 90% could be produced with an estimated payback period of less than 18 months. Concentration Units. Typical concentrators for rinsing solutions are membrane filtration units, which split the feed into diluate and concentrate streams, meaning purification and recovery, respectively [106]. Both nanofiltration and reverse osmosis might be applied, depending on the physico-chemical properties of the solutes. To produce highly concentrated solutions suitable for re-use in plating baths, high pressure reverse osmosis might be necessary [84]. A combination of electrodialysis with a concentrator media, ion-exchange resins or activated carbon in the catholyte chamber has been suggested by Chaudhary et al. [107]. Besides anodic chromium regeneration about 90% of dissolved copper could be recovered. Another approach to achieve purification of rinses and recovery in one step, electrodialysis has been suggested for chromic acid recovery and removal of metallic impurities [108]. As the authors point out there are two main process limitations: first, the poor stability of most anion-exchange membranes against the oxidative chromic acid solution and secondly the increase in membrane resistance due to the formation of polychromates in the membrane. Recovery of Minor Compounds. Extraction and separation of nickel(II) and its recovery from spent electroplating bath residue is reported by Singh et al. [109].

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

324 11 Technical Aspects

Along with Cr(III), Fe(III), Mn(II), Co(II), Cu(II) and Zn(II), Ni(II) was removed from sulfuric acid media, employing a Cyanex 301–toluene system. The success depended on various parameters such as the concentration of the acid, metal ion and extractant and the nature of the diluent. A more selective recovery of nickel from plating wastewater was described by Eom et al. [103].They used a column packed with strongly acidic cation resin through which over 99% nickel ion was removed. In this process, sulfuric acid was employed with a reagent in order to regenerate nickel ions from the resin adsorbed. Moreover, the nickel ions recovered by sulfuric acid were obtainable up to 120 g-Ni L−1 allowing reuse in the plating bath. Investigations by Malinowska et al. [110] have shown that absorption can be used to recover 90% of ammonia that is vaporized during chemical bath deposition of cadmium sulfide thin layers from which concentrated solutions with more than 10 mol L−1 of pure ammonia can be obtained. Additionally, a cake with mixed cadmium sulfide–cadmium cyanamide is produced, from which cadmium can be recovered hydrochemically as cadmium sulfate [111]. The global process recovers up to 99.999% of cadmium and generates only solid sulfur and a liquid effluent containing traces of cadmium. Finally, impurities that accumulate during usage of electrolyte can also be recovered. For example, Ni–Cu–Zn ferrite powder can be prepared from steel pickled liquor and electroplating waste solutions by a hydrothermal process [112]. Transfer from Water-based to Ionic Liquid Based Liquors. In the case of water-based electrolytes, there are two economic incentives for the above mentioned approaches: the recovery of valuables and the avoidance of wastes and wastewaters. Despite the environmental attractiveness of such measures economic constraints may become an obstacle in industrial application. For ionic liquid based process liquors, the contrary can be assumed. Due to their relatively high prices and anticipated costs for discharge of spent liquors a breakthrough of ionic liquids in plating applications can be expected to be linked to successful regeneration options. Even though regeneration units have not yet been reported for ionic liquid based electrolytes, it is most likely that some of those mentioned above could be transferred to this new field. For example, the application of electrodialysis could presumably allow removal of ionic impurities from ionic liquids. As in water-based electrolytes, it should be possible to separate small and relatively highly charged metal cations across cation exchange membranes and then to precipitate them out in an alkaline catholyte. But for such a method the complementary anodic process has to be designed carefully. For example, there should be another species to be oxidised such as Cr(III) in spent chromic plating baths or the separated cations could be replaced, for example, by anodic dissolution of the metal that is to be plated. However, electroneutrality has to be guaranteed as a crucial constraint in electric field driven separation processes. Other unit operations that have been established for aqueous solutions could be considered, to allow regeneration and reuse of ionic liquid based electrolytes.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 325

Actually, as can be seen in the following section, several of the separation methods mentioned above have already been tested in the purification of at least fresh ionic liquids. However, there is still some development necessary to come up with sustainable regeneration units. 11.4.2.3 Regeneration Options for Ionic Liquids in Electroplating Despite the huge number of publications dealing with the application of ionic liquids, to the best of our knowledge, there is only one paper [113] that mentions general problems related to purification of ionic liquids for electrochemical applications and it appears that there is none so far that deal with regeneration of spent ionic liquid based electrolytes. This is amazing, considering that the influence of impurities often narrows drastically the potential window available, as illustrated by Zhang and Bond [113]. However, a number of purification procedures have already been tested on the laboratory scale for fresh ionic liquids with respect to their downstream processing but little is known about efficiency on a technical scale. The reason for this lack of experience in large scale purification is quite simple: downstream processing is avoided so as to minimize the production cost of ionic liquids. On a commercial scale separation processes needed for purification can be assumed to be more costly than improvements in the synthesis stage [114]. Regeneration Options for Ionic Liquids in Other Fields of Application. In fields of application other than electroplating several examples of ionic liquid regeneration and reuse are described in the literature. For example, in the field of new reaction media [115, 116] or in the field of catalysis [117–121]. Even though they do not deal with electrolytes, they are a useful guide to learning about possible concepts and challenges. For example, Song et al. [122] described the reuse of an aminofunctionalized ionic liquid applied as a nucleophilic scavenger in solution phase combinatorial synthesis. Here regeneration was necessary to remove extracted electrophiles, such as benzoyl chloride and phenyl isocyanate, by a combination of extraction and phase separation steps, such as decanting and filtration. Neither FTIR nor 1 H NMR spectra showed any significant differences between the freshly prepared and the regenerated ionic liquid. Here, the reusability of the regenerated ionic liquid was demonstrated by reusing it three times as scavenger with comparable activity in terms of product yield and purity. Thermal Unit Operations. The easiest case for regenerating ionic liquid electrolytes is when the impurity is volatile. This is due to the negligible vapor pressure of the ionic liquid and the resulting extreme vapor pressure difference. For such a task, simple distillation in a single step is sufficient. If more than one volatile solute is present in the solution from which one is to be removed selectively, the task becomes more demanding. In this case, the other solutes would be lost through the simple distillation process. Alternatively, the volatile components could be separated from each other by repeated vaporization–condensation cycles within a packed fractionating column. If the other solutes show lower boiling points

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

326 11 Technical Aspects

another method should be considered. Finally, the method is chosen on economic grounds. The same is true for the technical application of vacuum distillation that can be performed by means of a rotary evaporator. For relatively high boiling temperature compounds such as water, that is a common impurity in ionic liquids from many applications, this technique is in general very useful, as it is for removing compounds with boiling points near or beyond the decomposition temperature of the ionic liquid at atmospheric pressure. For high purity purposes, the target concentrations of contaminants are extremely low and vacuum distillation might also be an option. For example, vacuum distillation at 120 ◦ C resulted in ionic liquids with moisture content below 10 ppm, as reported by Appetecchi et al. [123]. Scott et al. [115] found that successful reuse of an ionic liquid in a new synthetic route required regeneration by removing the methanol, which was used as a precipitating agent, under vacuum,. As an alternative to simple distillation, pervaporation could be used [124]. This technique makes use of non-porous membranes with a selective layer consisting of hydrophilic or hydrophobic polymer. Those compounds, which are volatile and soluble in the membrane, are evaporated into the vacuum on the permeate side. By this means, selective separation, for example of volatile impurities from volatile auxiliary agents in the ionic liquid, should be possible. To the best of our knowledge this possibility has not yet been shown to work. The two major challenges are the relatively high Reynolds numbers necessary inside the membrane module and the need to find selective membranes suitable for ionic liquids. Ceramic membranes show great potential for this application but so far there are only a few choices available on the market. Another thermal separation unit often used for the laboratory scale purification of ionic liquids is recrystallization [125]. It is an attractive option for those ionic liquids that can form solids with a high degree of crystallinity. Crystals of ionic liquids are expected to be pure because each molecule or ion must fit perfectly into the lattice as it leaves the solution. Impurities preferentially remain in solution as they do not fit as well in the lattice. The level of purity of the crystal product finally depends on the extent to which the impurities are incorporated into the lattice or how much solvent is entrapped within the crystal formed. In single-solvent recrystallization, the impure ionic liquid is dissolved in the minimum amount of a single solvent necessary to give a saturated solution; the solution is then allowed to cool. As cooling progresses, the solubility of the compounds in solution drops, resulting in the desired recrystallization. To enhance the process, a seed crystal of the pure ionic liquid is preferably added to the saturated solution resulting in these crystals forming first and thus leaving a greater ratio of impurity in solution. In the case of unsatisfactory separation factors, multi-solvent recrystallization can be tested. Here a second solvent, in which the impurities are soluble and the ionic liquid is not, is added carefully to the solution. As mentioned above, a possible drawback of recrystallization is the potential presence of solvent traces in the ionic liquid. This might result in the formation of yellowish compounds, as was reported by Appetecchi et al. [123].

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 327

Extraction Processes. The extraction procedure usually applied for hydrophobic ionic liquids containing hydrophilic impurities is “washing” with water. However, this method is a problem for certain types of ionic liquids that undergo hydrolytic decomposition, such as those containing hexaflurophosphate. At first glance, washing appears to be a simple and cheap method, but in large scale applications problems related to the wastewater issues may arise. Even though hydrophobic ionic liquids have a low solubility in water, the concentrations are relatively high, typically ranging from 0.1 to 10 g L−1 in the discharge. The fact that most cations and hydrophobic anions do not show significant biodegradability, coupled with the loss of costly materials in the discharge, explains the problem in large scale applications. A potential solution of this problem lies in the application of nanofiltration. The success of solvent extraction to remove polar or non-polar compounds from ionic liquids appears to depend strongly on the system for which it is used. While in some cases only “mixed success” is reported [126], in other applications solvent extraction has been shown to lead to excellent results, for example extraction with hexane [127]. It has also been shown that in some cases consecutive removal of the extractant is necessary if it partly dissolves in the ionic liquid, as Zulfiqar and Kitazume [116] reported for the application of diethyl ether. They purified the ionic liquids after extraction by distillation at 80 ◦ C. Therefore, before planning for a process scale up, there are some questions that need to be answered such as: (i) How often could the solvent be reused directly? (ii) By what means could the impurity be removed from the solvent? (iii) To what extent does the ionic liquid accumulate in the solvent? and (iv) How does this accumulation influence the performance of the intended separation? The separation of the auxiliary agent can be easily handled on a technical scale if it forms a pure phase. Otherwise more sophisticated separation methods are needed. In the case of ionic liquids a process termed organic solvent nanofiltration has been tested successfully [120, 128]. Adsorption Processes. Since adsorption processes often show high distribution coefficients, several adsorbents are favorite candidates for removing low concentrations of impurities. An important group are chromophores. In the synthesis of ionic liquids the formation of color, generally ranging from yellowish to orange, has been attributed to side reactions, e.g. from excessive heating during synthesis [129]. It can be assumed that color occurs at elevated temperatures for instance due to formation of a dimer of the amine and the ionic liquid or ionic liquid precursor in which the amine is dissolved. As an alternative to the avoidance of such side reactions during synthesis, several approaches for decolorization have been developed, ranging from recrystallization or adsorption to extraction. The prevalent method in the literature is definitely adsorption. A number of attractive adsorbents have been tested already, among which are activated carbons and aluminas, synthetic zeolites, and silica gels. Both batch contactors as well as chromatographic columns have been suggested for decolorization [130]. Nevertheless, to remove color with a single adsorbent was

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

328 11 Technical Aspects

not always sufficient, neither with powdered carbon in a batch contactor [131] nor with alumina or silica in a semi-continuous column [125]. However, with subsequently applied powdered carbon and alumina [123] or in a combined chromatographic column with granular carbon and silica gel, as described by Earle et al. [132], the adsorption process provided even better results. However, losses of ionic liquid were significant in such adsorption procedures [123]. Redox Processes. Among the most serious impurity problems for electrochemical applications is the contamination of electrolytes with halides. Since they easily react anodically they can be expected to reduce the size of the electrochemical window drastically but the readiness of their anodic decomposition can be used for a decontamination procedure. This was recently described by Li et al. [133] for chloride impurities. They found that, in combination with a subsequent removal of the gaseous product Cl2 by absorption, electrochemically pure ionic liquids can be obtained. Ethylene was bubbled through the solution to absorb the chlorine gas. Without such an absorption step, the soluble complex Cl3 − was formed which could not be removed by vacuum distillation. Both formation and subsequent removal of the complex Cl3 − can be easily followed spectrometrically due to a strong band of this species at 302 nm. The crucial parameter for the anodic decomposition of halides is the anodic potential. This is simply due to the dilemma that a minimum potential for decomposition is needed but degradation of the ionic liquid cation is enhanced with increasing potential. It was found for the [BMIM] cation that at a voltage about 20% above the decomposition value the appearance turned gradually from colorless to light yellow. Another option could be photochemical decomposition of impurities. Yang and Dionysiou [134] described a combined approach for treating solids or liquids which contain environmentally important organic contaminants. They suggest using room-temperature ionic liquids as solvent media and a subsequent photolytic degradation of the contaminants. The second step, the photolytic degradation, could, in principle, also be used for regeneration. It can be assumed that photolytic degradation is capable of degrading components in ionic electrolytes to below the required limit concentrations. The constraint here is that metabolites may be produced, which accumulate instead of the primary compound, exceeding their own required limits. Mechanical Processes. For removing particulate matter from low viscous liquids filtration generally is the technique of choice. Gan et al. [135] studied microfiltration characteristics of room temperature ionic liquids. They found that due to the relatively high viscosity it was impossible to get the tested liquids permeated through the microfiltration membranes with ease. They suggested mixing the ionic liquid with 20 % volumetric proportion of diluting polar agents, preferably methanol or ethanol, to drastically reduce viscosity. Alternatively, it can be assumed that at elevated temperatures it should be possible to receive comparable results at high temperatures without addition of another solvent.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 329

The separation of non-volatile products from ionic liquid solutions using nanofiltration was suggested by Kr¨ockel and Kragl [136]. It was shown for both bromophenol blue and lactose, each in ionic liquid, that the product was rejected while the ionic liquid permeated. It should be noted that in such cases the products are not isolated. Instead, concentrated ionic liquid solutions are produced. However, depending on the solubility, phase separation might occur. Another already mentioned application of membrane filtration is for the recovery of ionic liquids from wastewaters. Here the challenge is to find appropriate membranes, since rejection values that have been reported to date [136] are too low for industrial application. However, for similar ionic liquids we found a membrane that shows rejection rates above 99% throughout at considerably high permeate flow rates above 50 L m−2 h−1 in cross-flow filtration. Such numbers make washing in combination with nanofiltration an interesting option.

11.4.3 Case Study

Every single regeneration problem has to be analysed individually, however, the following case study demonstrates how a selection of separation techniques, extraction and phase separation, can successfully be applied to regenerate a spent ionic liquid based electrolyte satisfactorily. As a case study the electrolyte 1-butyl1-methylpyrrolidinium bis(trifluoromethanesulfonyl)amide ([BMP]Tf2 N) was chosen, which is used for electrodeposition of aluminum as described in the literature [137, 138]. The spent electrolyte was prepared by IoLiTec GmbH as follows: Dry AlCl3 (26.2 wt%) was dissolved in [BMP]Tf2 N, resulting in a solid at room temperature. At the process temperature (100 ◦ C) the mixture formed two liquid phases. The deposition took place in the upper phase at a voltage of –2 V (anode: Al plate, cathode: gold plate). After 90 min of deposition (charge: 145 A s) a mixed black and silver colored coating was received cathodically, and the electrolyte was collected for regeneration (Figure 11.18(a)). Starting the regeneration procedure a 10 mL sample of the spent ionic liquid/ AlCl3 mixture was heated to 75 ◦ C and stirred under nitrogen flow (Figure 11.18(b)). Deionized water was added stepwise (in amounts of 1 mL) with a syringe to the stirred two-phase liquid. During addition of water gas evolution could be observed in the vials. This could either be due to the strongly exothermic hydration (Hsolvation = –330 kJ mol−1 ) of the AlCl3 (Eq. (11.14)) leading to generation of water vapor or to the thermal decomposition of the hexaaquaaluminum trichloride resulting in the liberation of HCl gas (Eq. (11.15)).

AlCl3 + 6 H2 O → AlCl3 ·6 H2 O

(11.14)

AlCl3 ·6 H2 O → Al(OH)3 + 3 HCl + 3 H2 O

(11.15)

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

330 11 Technical Aspects

Fig. 11.18 Samples of spent [BMP]Tf2 N electrolyte (a) directly after electrodeposition of Al and (b) stirred at 75 ◦ C under nitrogen atmosphere.

After total addition of 9 mL water, resulting in the weight fractions wIL = 0.45, wAlCl3 = 0.16 and wH2 O = 0.39, the mixture was stirred for 15 min. Subsequently the sample was shaken for an additional 10 min while cooling to ambient conditions. At room temperature the mixture divided into two liquid phases. These phases were separated (Figure 11.19) by centrifugation (20 min at 2460 g). The lower clear and more viscous phase was presumed to be the IL phase and the upper liquid the water phase. At the interface fine particles were collected. Using a syringe the phases were carefully separated and transferred to different vials. The ionic liquid phase was then submitted to an evaporation procedure (rotary evaporator, 50 ◦ C, 10 mbar, 6 h). Shortly after connecting to vacuum bubble generation could be observed.

Fig. 11.19 Samples after mixing with water, showing phase separation.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 331

Fig. 11.20 Recovered ionic liquid phase ([BMP]Tf2 N) after regeneration.

After evaporation the ionic liquid phase contained a dispersed precipitate. These solid particles were concentrated at the bottom of the flask by centrifugation (20 min at 2460 g) resulting in a very clear, slightly yellowish ionic liquid phase (Figure 11.20). It should be stated here that on a technical scale washing requires a concept for water reuse and recovery of ionic liquid from the wastewater. As already discussed, nanofiltration is likely be a successful approach for the recovery task. The regenerated ionic liquid phase was investigated electrochemically to determine its quality. Cyclic voltammetry was performed using a rotating platinum disk electrode (500 rpm), a platinum counter electrode and a platinum wire as (quasi-) reference electrode placed closed to the rotating disk. In Figure 11.21 two ionic liquids are compared, a freshly synthesized [BMP]Tf2 N received from Iolitec GmbH and the regenerated ionic liquid. It can be clearly seen that current densities after regeneration are lower than for the fresh electrolyte throughout the entire potential range. No additional signal can be recognized for the regenerated ionic liquid. This indicates that none of the electrochemically active additional ingredients, water and Al(III), remain. The regenerated ionic liquid appears to be at least as pure as the originally synthesized ionic liquid. The regeneration was successful. In the fresh electrolyte a first anodic step starts at 1500 mV. This could be a hint for chloride impurity. Since this signal almost vanished for the regenerated ionic liquid, it can be assumed that the procedure presented is suitable also for purifying fresh ionic liquids.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

332 11 Technical Aspects

Fig. 11.21 Cyclic voltammograms of original and regenerated ionic liquid [BMP]Tf2 N)). The potential was determined vs. Pt as quasireference electrode. Scan rate 5 mV s−1 .

The pros and cons of this approach are summarized as follows: Ĺ Ease of process. Ĺ Small amount of losses of ionic liquid, but aluminum salts are completely lost after hydrolysis for the plating process. Ĺ Pure ionic liquid as product. Ĺ Only water as solvent necessary, which can be re-used in the regeneration process to a certain extent. Ĺ Before discharge of the wastewater dissolved amounts of ionic liquid need to be recovered for environmental and economical reasons, e.g. by nanofiltration. Ĺ In contrast to dead-end microfiltration, which could also be used to remove solids from spent electrolytes producing (after addition of a solvent and at elevated temperatures) an ionic liquid as residue, the residue in the extractive regeneration is wet sludge only.

Despite the conspicuous advantages of the presented water-based regeneration approach, it is still to be shown whether it can be transferred to other tasks and whether the reuse of the regenerates in plating processes leads to surface qualities similar to those received from fresh electrolytes. 11.4.4 Conclusions

A general approach towards both more economical and more environmentally benign applications of ionic liquids is maximization of their lifetime. The measures to be applied in electroplating are recovery and regeneration, both to allow

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.4 Towards Regeneration and Reuse of Ionic Liquids in Electroplating 333

reuse. This study focuses mainly on regeneration, but also recovery of drag-out is considered, which should be possible in conventional counter-current rinsing systems, albeit in combination with regeneration units such as membrane concentrators. This study focuses firstly on the transfer of regeneration principles as they have been developed in the field of water-based electroplating and of purification options for ionic liquids as they are experienced in other fields of ionic liquid application. A number of purification procedures for fresh ionic liquids have already been tested on the laboratory scale with respect to their finishing in downstream processing. These include distillation, recrystallization, extraction, membrane filtration, batch adsorption and semi-continuous chromatography. But little is known yet about efficiency on the technical scale. Another important aspect discussed is the recovery of ionic liquids from rinse or washing water. However, the financial and environmental cost might be too high for a certain approach. Hence the optimality of a solution is always subject to technical constraints and the technical bottleneck of each option has to be identified. For any optimization approach it has to be considered that the demand for regeneration is finally related to large scale applications. The mass flow that has to be treated during regeneration will range typically from grams to kilograms per minute. Since little is yet known about efficiency on the technical scale, future investigation should focus on (i) efficiency with respect to separation yield, energy demand and amount of mass separation agents required, (ii) long-term re-use options of auxiliary agents such as extractants or adsorbents and (iii) ease of scaling up. Furthermore, a crucial point for further development of regeneration will be to identify the pollutants that disturb the main process as well as their critical concentration levels in the electrodeposition process. In a case study, the extraction of a spent, turbid electrolyte with water at elevated temperature and subsequent phase separation is shown as an example. It could be demonstrated that purification of ionic liquids for re-use is not necessarily as difficult as suspected in the literature [113]. The case study is going to be continued to demonstrate whether the application of the regenerates will lead to comparable surface qualities. Accordingly, it is of future interest to see whether the roughness of cathodic deposits using regenerated electrolytes shows similar dependences on current density and temperature as does the roughness of deposits from fresh electrolytes. Additionally, future investigations should consider gaseous impurities, which could be dragged in easily due to the high solubility and capacity of many ionic liquids for trace gases, especially sulfur compounds. Acknowledgments

The authors wish to thank IoLiTec GmbH for providing ionic liquid and spent electrolyte, and several partners of the BMBF project NEMESIS for fruitful discussions. Financial support by VDI/VDE-IT (project No. 16SV1970) is gratefully acknowledged.

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

334 11 Technical Aspects

11.5 Impurities

Impurities are a concern in ionic liquids electrochemistry. Whereas even considerable amounts of impurities, like different metal ions, water or organic impurities, might not disturb a technical process (e.g. extractive distillation, organic synthesis) the wide electrochemical windows of an ionic liquid (∼± 3 V vs. NHE) allow the electrodeposition of even reactive metals like lithium and potassium, as well as the oxidation of halides to the respective gases. In the best case this codeposition only leads to a low level of impurities, in the worst case fundamental physicochemical studies are made impossible as the impurities are adsorbed onto the electrode surface and subsequently reduced. Furthermore, passivation or activation effects at the counter electrode have to be expected. In the last few years the different suppliers of ionic liquids have developed several purity grades. Merck has introduced “synthesis”, “high purity” and “ultrapurity” (see Chapter 1.2) and other suppliers also follow this purity scheme. In the past much attention was focussed on impurity effects. However, as the different suppliers are about to establish large scale production lines where the costs for the educts have to be quite low one has to be prepared that the problem of impurities may return. As many groups (in part without any experience at all) have entered the field of ionic liquids in recent years we would like to draw attention to the subject of impurities. Impurities can be a concern but do not necessarily have to be a concern. We could also imagine that for some processes impurities are beneficial but, as a minimum, one should know their role in the respective process. 11.5.1 Origin of Impurities 11.5.1.1 Synthetic Impurities The synthesis process represents a very significant source of impurities in ionic liquids. Because of their typically low volatility, which makes distillation impractical, and the lack of any straightforward crystallization method of purification, ionic liquids are often delivered in a semi-impure state. Significant impurities include starting materials, such as halides, and metal cations, such as lithium, sodium or silver, and any impurities carried through from the synthesis of the organic cation, in particular amines. Where halides such as bromide and iodide are present, some oxidised species such as I3 − are often also present, generating color in the otherwise colorless ionic liquid. Most of these can be quite difficult to remove; however, at the 1% or below level, in many cases they can be tolerated as long as the impurity level is consistent from batch to batch. Seddon [139] has discussed the impact of low levels of impurities on physical properties including viscosity. Chloride ions are particularly notable for their effect in lowering viscosity. If analytical information is not available from the supplier of the ionic liquid it is advisable to carry out analysis, using traditional AAS or ICP-MS methods for the metals and halide-selective electrode analysis for the halides. Residual amine

c11 (JWBG008-Endres)

January 11, 2008

9:55

Char Count=

11.5 Impurities 335

can be easily detected using a Cu(II) complex formation and UV/vis absorbance measurements. In some ionic liquids, acid (proton) impurities are significant. This is common in the phosphonium cation family of ionic liquids and can also be the case with nitrogen-based cations if the synthetic method involves a neutralization reaction. It is relatively easy to deal with this situation. Acidity should be determined by a standard titration method and then the acidity neutralized by addition of an appropriate base. Carbonates are particularly useful in this regard since they produce CO2 as the product. 11.5.1.2 Water Water presents a rather different problem in that its presence can originate from the synthesis, or from handling and storage prior to (or even during) the electrodeposition. Notably, even ionic liquids such as [EMIM]Ntf2 ] that we think of as being hydrophobic are nonetheless reasonably hygroscopic up to their saturation point, so that storage and handling needs to involve an inert atmosphere. The presence of this water is particularly significant in the potential region below –0.5 V (vs Ag/Ag+ ) where it produces reduction products directly and also may cause degradation of the ionic liquid and/or a surface film to form on the deposited metal. Howlett et al. [140] have suggested that [Ntf2 ] ionic liquids produce breakdown products of the anion on metals such as lithium and magnesium in a reaction that is catalyzed by reduction products of water such as the hydroxyl radical. These reduction products may produce useful protective films in some cases, such as lithium, such that further reduction of the metal ion can take place via transport through the film, but this is unlikely to be the situation in the case of more highly charged metal ions such as Ti(II). Water analysis can be routinely carried out by a Karl–Fischer analysis in which the ionic liquid is diluted in methanol before analysis. A spiking approach can be used to produce a calibration curve that allows for background effects. At very low levels of water (99%) and 1-butyl-3-methyl imidazolium chloride ([BMIM]Cl) (Aldrich >99%) were weighed at a 2:1 mole ratio into two separate Schlenk tubes and dried on a vacuum line for 2 h prior to use. The two components were mixed by adding AlCl3 to the [BMIM]Cl tube with stirring at room temperature, leading to a homogeneous, straw-brown liquid of “technical quality”. Finally, 3 mole equivalents of toluene (corresponding to 39 wt.%) were then transferred into the 2:1 neat ionic liquid using a stainless steel cannula. A homogeneous mixture (dark green in color) was obtained by stirring the liquid for 15 min. The liquid was maintained under a dry nitrogen atmosphere at all times.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

12.2 Electrodeposition of Al from 1-Butyl-3-methylimidazolium chloride–AlCl3 – Toluene 357

12.2.3 Pretreatments

To achieve a good adhesive coating and maintain the electrolyte stability, both cathode and anode need to be treated properly before they are mantled for electroplating.

12.2.3.1 Cathode (Mild Steel Rods) Ĺ Polished using P400 sand paper and cleaned using tissues. Ĺ Degreased in acetone under ultrasonic conditions for 15 min. Ĺ Activated chemically in 5 wt.% HCl for 2 min to remove possible oxide layer then rinsed thoroughly using deionized water. Ĺ Degreased in dichloromethane for 10 min to remove any organic impurities and form a chloride layer which is resistant to oxide formation.

12.2.3.2 Anode (Al) The anode was polished using P400 sand paper, and then activated by dipping in (1% HNO3 , 65% H3 PO4 , 5% acetic acid and water) for 5 min, followed by rinsing thoroughly with deionized water and degreasing in acetone for 5 min.

12.2.4 Electroplating and Morphology Analysis

Electroplating experiments were performed using a two-electrode set-up under N2 atmosphere in a Schlenk tube. The cathodes were mild steel rods with diameter 0.6 cm, and the anode was a cylindrical bucket of Al sheet of diameter 1.6 cm placed around the cathode. Anodic etching of mild steel rods was performed by applying +1 V for 30 s to remove any possible oxide layer prior to electroplating. All samples were prepared by applying constant potentials for 60 min. Samples were rinsed using toluene followed by isopropanol and then deionized water after removal from the Schlenk tube. Surface analysis was carried out using scanning electron microscopy (Philips XL30 ESEM) and energy dispersive analysis by X-rays (EDX).

12.2.5 Results

The surface morphologies of the deposits are highly dependent on the potential applied between the anode and cathode. For lower voltages (0.5 V), a growth of nanocrystals dominates, leading to smooth bright and shining samples. Figure 12.5 shows photos of the samples and Figure 12.6 SEM images of the same samples.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

358 12 Plating Protocols

Fig. 12.5 Photos showing (a) the dull finish at 0.5 V and (b) the bright finish at 1.0 V.

12.3 Electrodeposition of Al from 1-Ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl)amide/AlCl3

In this protocol we describe the electroplating of mild steel with thick layers of aluminum in the water and air stable ionic liquid 1-ethyl-3-methylimidazolium bis(trifluoromethylsulfonyl) amide [EMIM]TFSA containing AlCl3 . We aim to electroplate mild steel with dense, adherent aluminum layers in the employed ionic liquids. 12.3.1 Experimental Set-up

A quartz round flask was used as an electrochemical cell with three electrodes. Al-wires (Alfa, 99.999%) were used as reference and counter electrodes. Mild steel sheets were employed as working electrodes. The working electrodes were mechanically polished with emery paper, cleaned with acetone in an ultrasonic bath, treated with dilute hydrochloric acid and rinsed with distilled water. Prior to the electrodeposition process the electrodes were anodically polarized in the employed ionic liquid to remove as far as possible the native oxide layer. Removal of the

Fig. 12.6 SEM images of (a) the dull finish at 0.5 V and (b) the bright finish at 1.0 V.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

12.3 Electrodeposition of Al from 1-Ethyl-3-methylimidazolium 359

surface air-formed oxide layer is a prerequisite for achieving adherent coatings. The cell was thoroughly cleaned in a mixture of 50/50 vol% H2 SO4 /H2 O2 followed by refluxing in bi-distilled water. The deposition experiments were performed in an argon-filled glove box using a Parstat 2263 Potentiostat/Galvanostat (Princeton Applied Research). 12.3.2 Chemicals and Preparation

The ionic liquid [EMIM]TFSA was purchased from Merck KGaA(EMD) in the highest available quality and was dried under vacuum for 12 h at a temperature of 100 ◦ C then stored in an argon-filled glove box with water and oxygen below 1 ppm (OMNI-LAB from Vacuum-Atmospheres). Anhydrous AlCl3 (Fluka, 99%) was used without further purification as a source of aluminum. It is important that AlCl3 grains are employed, as powders (even in 99.999% quality) only contain low amounts of active AlCl3 , according to our experience. The ionic liquid [EMIM]TFSA shows, at room temperature, biphasic behavior on addition of AlCl3 . AlCl3 dissolves well in [EMIM] TFSA up to a concentration of about 2.5 mol L−1 , then a biphasic mixture is obtained on further addition of AlCl3 . The upper phase of the mixture AlCl3 /[EMIM] TFSA is clear and colorless while the lower one is pale and more viscous. On further addition of AlCl3 the viscosity of the lower phase increases. It is worth noting that Al can only be electrodeposited from the upper phase, the clear one, at AlCl3 concentrations ≥ 5 mol L−1 . Furthermore, after a few days a precipitate which contains Al(TFSA)3 forms as a third phase. 12.3.3 Results

The SEM micrograph of Figure 12.7(a) shows the surface morphology of a deposited aluminum layer obtained galvanostatically at a current density of −5 mA cm−2 for 2 h in the upper phase of the biphasic mixture [EMIM] TFSA/6 M AlCl3 at room temperature. Prior to Al electrodeposition, the electrode was anodically polarized at a potential of 1 V (vs. Al) for 2 min. As seen, the deposited Al layer is dense and contains crystallites in the micrometer regime. Figure 12.7(b) shows SEM micrographs of the cross-section of the deposited aluminum layer on a mild steel substrate. As shown in the SEM micrograph the deposited Al layer adheres well to the mild steel substrate and the layer is homogeneous with a thickness of about 10 µm. Also, with higher magnification, the Al layer exhibits a good adhesion without any splits between it and the substrate, inset of Figure 12.7(b). Figure 12.8 shows the photo of a deposited aluminum layer obtained potentiostatically on a mild steel substrate at −0.3 V (vs. Al) for 4 h in the upper phase of the mixture [EMIM] TFSA/6 M AlCl3 . The substrate was electrochomically etched at 1 V (vs. Al) for 2 min prior to electrodeposition. The aluminium layer adheres so well that it can be mechanically polished to a mirror appearance.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

360 12 Plating Protocols

Fig. 12.7 (a) SEM micrograph of an about 10 :m aluminum layer electrodeposited galvanostatically on a mild steel substrate at −5 mA cm−2 . Inset: SEM micrograph of higher magnification showing the excellence of the coating adhesion. (b) SEM micrograph of the polished cross-section of the deposited aluminium layer [2].

12.4 Electrodeposition of Al from 1-Butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide/AlCl3

In this protocol we describe the electrodeposition of nanocrystalline aluminum without additives in the water- and air-stable ionic liquid 1-butyl-1methylpyrrolidinium bis(trifluoromethylsulfonyl) amide [Py1,4 ]TFSA containing AlCl3 . 12.4.1 Experimental Set-up

The experimental set-up used was as described in Section 12.3.1. Gold substrates from Arrandee (gold films of 200–300 nm thickness deposited on chromiumcovered borosilicate glass) and glassy carbon (Alfa) and mild steel sheets were used as working electrodes, respectively. Directly before use, the gold substrates were heated in a hydrogen flame to slightly red glow for several minutes. The glassy carbon substrate was degreased with acetone in an ultrasonic bath for 10 min. The mild steel substrates were mechanically polished with emery paper, cleaned with

Fig. 12.8 An optical photo of a deposited Al layer made potentiostatically at −0.3 V (vs. Al) in the upper phase of the mixture [EMIM] TFSA/6 M AlCl3 at room temperature.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

12.4 Electrodeposition of Al from 1-Butyl-1-methylpyrrolidinium 361

acetone in an ultrasonic bath, treated with dilute hydrochloric acid and rinsed with distilled water. 12.4.2 Chemicals and Preparation

The ionic liquid 1-butyl-1-methylpyrrolidinium bis (trifluoromethylsulfonyl)amide was purchased from Merck KGaA(EMD) in the highest available quality and was dried under vacuum for 12 h at a temperature of 100 ◦ C then stored in an argonfilled glove box with water and oxygen below 1 ppm (OMNI-LAB from VacuumAtmospheres). Anhydrous AlCl3 (Fluka, 99%) was used without further purification as a source of aluminum. Similar to the AlCl3 /[EMIM]TFSA mixture, the mixture of AlCl3 /[Py1,4 ] TFSA shows biphasic behavior with increase in the concentration of AlCl3 up to 1.6 M. In contrast to the AlCl3 /[EMIM]TFSA mixture, the lower phase is colorless while the upper one is pale and more viscous. By adding more AlCl3 the volume of the lower phase decreases till a concentration of 2.7 mol L−1 is reached, then only one solid phase can be formed at room temperature. The biphasic mixture of AlCl3 /[Py1,4 ]TFSA becomes monophasic by heating to a temperature of about 80 ◦ C. The electrodeposition of aluminum occurs only from the upper phase at AlCl3 concentrations ≥ 1.6 mol L−1 . 12.4.3 Results

Nanocrystalline aluminum can be made in the employed ionic liquid without additives, see Chapter 8. The SEM micrograph of Figure 12.9 shows the surface morphology of a deposited aluminum layer obtained potentiostatically on mild steel at −0.75 V (vs. Al) for 2 h in the upper phase of the biphasic mixture [Py1,4 ] Tf2 N/2 M AlCl3 at 100 ◦ C. Prior to Al electrodeposition, the electrode was anodically polarized at a potential of 1 V (vs. Al) for 2 min. The deposited layer is dense, shining and adherent to the substrate with crystallites in the nanosize regime. Figure 12.10 shows a high resolution SEM micrograph of an about 5 µm thick layer of Al on gold substrate electrodeposited potentiostatically at 100 ◦ C at −0.45 V (vs. Al) for 2 h in the upper phase of the mixture [Py1,4 ]TFSA/1.6 M AlCl3 . Generally, the electrodeposited layer contains very fine crystallites in the nanometer regime. Figure 12.11 shows the XRD patterns of a nanocrystalline Al film obtained at a constant potential of −1.7 V for 2 h at 100 ◦ C in the ionic liquid [Py1,4 ] TFSA containing 1.6 M AlCl3 on a glassy carbon substrate. The XRD patterns show the characteristic diffraction patterns of crystalline Al, furthermore the peaks are rather broad, indicating the small crystallite size of the electrodeposited Al. The grain size of Al was determined using Scherrer’s equation to be 34 nm. For more information on the electrodeposition of nanocrystalline aluminum in the employed ionic liquid we refer to Refs. [3, 4].

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

362 12 Plating Protocols

Fig. 12.9 SEM micrograph of electrodeposited Al on mild steel made potentiostatically at −0.75 V (vs. Al) for 2 h in the upper phase of the mixture [Py1,4 ]TFSA/2 M AlCl3 at 100 ◦ C.

12.5 Electrodeposition of Li from 1-Butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide/Lithium bis(trifluoromethylsulfonyl)amide

A solution of lithium bis(trifluoromethylsulfonyl)amide is made up at ∼0.5 mol kg−1 in the ionic liquid 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide. Both the salt and the ionic liquid are dried prior to use at 100 ◦ C or above, under vacuum for 12 h or more, which gives water values at least below 10 ppm. These materials must be handled only in an argon-filled dry

Fig. 12.10 SEM micrograph of electrodeposited Al on gold formed potentiostatically at −0.45 V (vs. Al) for 2 h in the upper phase of the mixture [Py1,4 ]TFSA/1.6 M AlCl3 ) at 100 ◦ C.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

12.5 Electrodeposition of Li from 1-Butyl-1-methylpyrrolidinium 363

Fig. 12.11 XRD patterns of an electrodeposited Al layer obtained potentiostatically at −1.7 V for 2 h in the upper phase of the mixture [Py1,4 ]TFSA/1.6 M AlCl3 at 100 ◦ C on a glassy carbon substrate.

box. Since lithium reacts rapidly with both oxygen and nitrogen this experiment must be carried out under an argon, or other inert gas, atmosphere. To ensure that the ionic liquid does not contain traces of nitrogen or oxygen the lithium salt solution in the ionic liquid should be degassed by bubbling pure argon through the solution overnight; the solution is held at an elevated temperature (>50 ◦ C) during this process to reduce the viscosity. If available, the water content should be determined at this stage by Karl–Fischer titration and should be < 30 ppm. Plating of lithium can be performed galvanostatically with a current density around 1–2 mA cm−2 on a range of substrates including nickel, copper, platinum and glassy carbon. Nickel and glassy carbon tend to require a greater overpotential to achieve Li deposition. Platinum requires less over-potential, but tends to form alloys easily with lithium and therefore it is not preferred for a simple plating experiment. In all cases a passive film will form on the lithium deposit as it plates. This surface film that forms in 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide is approximately a few hundred nanometers thick and is known to be highly conductive to lithium [5]. Lithium plating occurs through the film but at a rate which, to some extent, is limited by the film, particularly at lower temperatures. Despite the presence of this film, the deposit will remain bright and shiny.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

364 12 Plating Protocols

The counter electrode is preferably lithium metal in order to provide a constant lithium concentration in the electrolyte. Lithium is also a very useful reference electrode in this ionic liquid in the form of a strip of foil. For preliminary experiments platinum is a suitable counter electrode. The electrodeposition can be carried out at room temperature, but is more facile at 50 ◦ C or higher due to the resistance of the passive film. Typically about 50–100 mV of overpotential vs. Li/Li+ is sufficient to obtain a deposit. It is important to limit this overpotential to 99%), ethylene glycol (Aldrich >99%), ZnCl2 (Aldrich >99%) and ethylene diamine (Aldrich >99%) were used as obtained. The eutectic mixture was formed by stirring a 1:2 molar ratio mixture of choline chloride and

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

366 12 Plating Protocols

ethylene glycol at 70 ◦ C until a homogeneous colorless liquid was formed. ZnCl2 was then dissolved in the liquid. As a further experiment two molar equivalents of ethylene diamine were added as a complexing agent. 12.7.2 Pretreatment

To attain an adherent Zn Coating the pre-treatment protocol below was followed: The cathode (mild steel) must be: Ĺ Ĺ Ĺ Ĺ

Polished using P400 sandpaper, rinsed in deionized water and dried. Degreased in acetone for 5 min. Chemically etched in 30% H2 SO4 for 30 s, and rinsed with deionized water. Degreased in dichloromethane for 5 min to remove organic impurities, rinsed in deionized water and dried with N2. Anode (IrOx -coated Ti mesh) must be:

Ĺ Degreased in acetone for 5 min, rinsed thoroughly in deionized water and dried with N2.

Homogeneous coatings were obtained by driving a constant current density of 5 mA cm−2 at 50 ◦ C, without stirring for 60 min. 12.7.3 Results

The deposition obtained from ZnCl2 dissolved in choline chloride: ethylene glycol (Figure 12.13(a)) contains small crystals of a homogeneous size. This deposit has a matte dark gray appearance. The addition of ethylene diamine (Figure 12.13(b)) leads to the growth of larger crystallites and a disperse silvery metallic finish.

Fig. 12.13 Scanning electron micrographs showing the deposits gained from (a) choline chloride: ethylene glycol (1:2) + 0.3 M ZnCl2 and (b) choline chloride: ethylene glycol (1:2) + 0.3 M ZnCl2 + 1 molar

equivalent ethylene diamine. Both experiments were carried out by applying a constant current density of 5 mA cm−2 at 50 ◦ C, without stirring for 60 min.

c12 (JWBG008-Endres)

December 25, 2007

10:24

Char Count=

References

References 1 Liu, Q.X., Zein El Abedin, S., and Endres, F. (2006) Surf. Coat. Technol., 201 (3–4), 1352. 2 Zein El Abedin, S. (2006) Z. Phys. Chem., 220, 1293. 3 Zein El Abedin, S., Moustafa, E.M., Hempelmann, R., Natter, H., and Endres, F. (2006) Chem. Phys. Chem., 7, 1535. 4 Zein El Abedin, S., Moustafa, E.M., Hempelmann, R., Natter, H., and Endres, F. (2005) Electrochem. Commun., 7, 1116.

5 Howlett, P.C., Brack, N., Hollenkamp, A.F., Forsyth, M., and MacFarlane, D.R. (2006) J. Electrochem. Soc., 153, A595. 6 Tiyapiboonchaiya, C., Pringle, J. M., Sun, J.Z., Byrne, N., Howlett, P.C., Macfarlane, D.R., and Forsyth, M. (2004) Nature Mater., 3, 29. 7 Zein El Abedin, S., Farag, H.K., Moustafa, E.M., Welz-Bierman, U., and Endres, F. (2005) Phys. Chem. Chem. Phys., 7, 2333. 8 Zein El Abedin, S., Welz-Bierman, U., and Endres, F. (2005) Electrochem. Commun., 7, 941.

367

c13 (JWBG008-Endres)

December 25, 2007

10:26

Char Count=

369

13 Future Directions and Challenges Frank Endres, Andrew P. Abbott, and Douglas MacFarlane

In this book the current state-of-the art of electrodeposition in ionic liquids has been summarized. In Chapter 2 the key aspects of three types of ionic liquids, i.e. (i) first generation ionic liquids based on AlCl3 , (ii) air- and water -stable ionic liquids and (iii) systems based on choline chloride, were introduced. After an introduction to the physical properties there are four chapters describing the more or less classical electrodeposition of metals, alloys, semiconductors and conducting polymers. The subsequent chapter describes rather novel aspects such as the electrodeposition of nanocrystalline metals and alloys which seems to be quite easy in some ionic liquids. The in situ scanning tunneling microscope gives direct insight into dynamic nanoscale processes during electrodeposition and plasma electrochemistry allows the preparation of suspensions of nanocrystalline metal particles quite simply by discharging a plasma over the ionic liquid. In the following we discuss possible future directions and some challenges.

13.1 Impurities

As briefly discussed in Chapters 4.4 and 11.5 ionic liquids can contain variable amounts of organic and inorganic impurities. The organic impurities, which often give a yellowish color to some liquids, arise either from impurities in the starting material or formed during the synthesis by partial decomposition/oligomerization of the cation and/or the anion. In our experience low levels of such organic impurities are not critical and even with the in situ STM, which is highly sensitive towards impurities that adsorb at an electrode surface, there is quite good picture quality, even in yellowish ionic liquids. Thus a low level of organic impurities might be tolerable for an electrochemical application. Common inorganic impurities in ionic liquids are water, metal ions and halide(s). Water is introduced during the synthesis of most ionic liquids as they are typically made either by the acid–base route from e.g. bis(trifluoromethylsulfonyl)amide-acid and diluted solutions of 1ethyl-3-methylimidazolium hydroxide in water, or via a metathesis reaction from e.g. lithium bis(trifluoromethyl-sulfonyl)amide and 1-ethyl-3-methylimidazolium Electrodeposition from Ionic Liquids. Edited by F. Endres, D. MacFarlane, A. Abbott C 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim Copyright  ISBN: 978-3-527-31565-9

c13 (JWBG008-Endres)

December 25, 2007

10:26

Char Count=

370 13 Future Directions and Challenges

chloride in aqueous solution. Water can be easily removed from such liquids simply by stirring them at elevated temperature (about 100 ◦ C) under vacuum. Water levels of 3 ppm and below are easily achieved. Metal ion and halide impurities are an issue in ionic liquids with discrete anions. As we have demonstrated in Chapter 11.5 Li+ (and K+ ) are common cationic impurities, especially in the bis(trifluoromethylsulfonyl)amides which typically contain 100 ppm of these ions from the metathesis reaction. Although Li and K are only electrodeposited in the bulk phase at electrode potentials close to the decomposition potential of the pyrrolidinium ions, there is evidence for the underpotential deposition of Li and K on gold and on other rather noble metals. For a technical process to deposit nickel or cobalt from ionic liquids the codeposition of Li and/or K, even in the underpotential deposition regime, has to be expected. Halide impurities can alter the complex chemistry in ionic liquids and can lead to unexpected oxidation reactions at the counter electrode. Furthermore even low amounts of e.g. chlorine can be formed, leading to some side reactions. When ILs were first commercially available, the quality of most samples was questionable as they contained numerous organic and inorganic impurities. More recently different quality levels have been introduced (for synthesis, high purity, ultrapurity). Ultrapure ionic liquids usually contain water, halide and metal ion impurities below 10 ppm and they are currently the best choice for fundamental physicochemical studies. There might be two distinct approaches to the purity issue in future: on the one hand ultrapure ionic liquids (i.e. impurity levels below 10 ppm) should be used for fundamental electrochemical studies to understand the electrochemical reactions alone, which can be quite complicated in ionic liquids. On the other hand a deep(er) understanding of the mentioned impurities might allow the use of lower quality ionic liquids for technical electrochemistry or electroplating. Water impurities might be less critical (if not beneficial) if an element like nickel or cobalt is deposited. Halide impurities might not be critical for semiconductor electrodeposition which can be achieved easily from halides. An understanding of the influence of impurities on the electrochemical processes and information on the levels that can be tolerated for a reaction would help in the design of technical processes. Impurities are a lot less problematic for eutectic-based ionic liquids. The strong acid–base nature of these systems leads to predominantly halometallate species which tend to be unaffected by simple salts or other impurities such as water. The strong Lewis acids and bases coordinate well to water and even in the chloroaluminate systems low amounts of water do not significantly affect voltammetric behavior or have a deleterious effect on deposit morphology. 13.2 Counter Electrodes/Compartments

In Chapter 11.1 some aspects of counter electrode reactions and metal dissolution were discussed. An interesting aspect in ionic liquid electrochemistry is that some reactive metals are quite noble. Aluminum, for example, is easily

c13 (JWBG008-Endres)

December 25, 2007

10:26

Char Count=

13.3 Ionic Liquids for Reactive (Nano-)materials 371

oxidized electrochemically in first generation ionic liquids based on AlCl3 , giving a reversible counter electrode. In air- and water-stable ionic liquids with the bis(trifluoromethylsulfonyl)amide anion, aluminum behaves as a passive electrode. On the other hand gold is oxidized in both types of ionic liquids. This is not too surprising: ionic liquids can have wide anodic decomposition potentials (up to 3 V vs. NHE), wide enough to allow the oxidation of almost all elements. In contrast to aluminum, gold can be present as “naked” Au+ ions which seems to facilitate an electrochemical oxidation in the mentioned liquid. In some ionic liquids platinum (especially in the presence of halide) can be oxidized and deposited on the working electrode if cathodic and anodic compartments are not separated. Counter electrode reactions have so far been more or less neglected in ionic liquid electrochemistry. As there can be unusual reactions, more effort should be invested in studying these processes. There are suggestions that the counter electrode can also influence the morphology of deposits at the cathode. In haloaluminate liquids, for example, although aluminum dissolves, the rate is limited by the diffusion of AlCl4 − to the electrode surface. The competition of generated Al3+ with AlCl3 for the halide anion is controlled by the relative Lewis acidity of the ionic liquid components or, more accurately, of the components in the double layer close to the electrode surface. Hence, in Lewis acidic ionic liquids the rate of aluminum dissolution is slower than the rate of deposition and under constant potential the rate is limited by anodic dissolution. Preliminary results have shown that the increased rate of deposition and improved quality of the deposit brought about by the addition of toluene is due primarily to the increase in the rate of the anodic process. The limited reversibility of some electrode reactions might require consideration of consumable (cheap) ionic liquids in the anode compartment for technical applications and commercial electroplating. For example, the electrochemical oxidation of oxalate delivers carbon dioxide, hydride could be oxidized to hydrogen, halides to the halogen or trihalide salt in the case of iodide ionic liquids and so on . Since ionic liquids can readily form biphasic systems an alternative may be to have the anodic reaction in an immiscible solvent. In that case a common ion would be needed that can be transferred from one phase to the other.

13.3 Ionic Liquids for Reactive (Nano-)materials

The electrodeposition of reactive elements like Al, Si, Ge, Ta and a few others is possible. As discussed in Chapter 4.4 the successful electrodeposition of Ti, Mg, Mo and many others in relevant layer thicknesses has not yet been described, though attempts have been made in some cases. Apart from the availability of suitable precursors there is at least one other issue to consider: ionic liquids can be reactive. It was found that magnesium and its alloys can form passivating films in ionic liquids with the bis(trifluoromethylsulfonyl)amide (Tf2 N) anion, especially in the presence of water. It was found by two of our groups (Endres, MacFarlane) that, under certain circumstances, the Tf2 N ion is subject to cathodic

c13 (JWBG008-Endres)

December 25, 2007

10:26

Char Count=

372 13 Future Directions and Challenges

breakdown. It is likely that water in the liquid plays an important role in the breakdown reaction. Attempts to deposit magnesium from Mg(Tf2 N)2 have not yet been successful since, in the presence of water, the Tf2 N is subject to reduction, producing a variety of decomposition products. The IL designers and synthesizers should cooperate more intensively with fundamental electrochemists and theoreticians to develop new ILs which do not exhibit such undesired electrochemical side reactions. A further important aspect is how to handle reactive elements? It was found in the Clausthal group that nanocrystalline aluminum and nanoscale silicon made in 1-butyl-1-methylpyrrolidinium bis(trifluoromethylsulfonyl)amide react even with the comparably low level of oxygen (