Electrochemical Storage Materials: From Crystallography to Manufacturing Technology 9783110493986, 9783110491371

Citation preview

Dirk C. Meyer, Tilmann Leisegang, Matthias Zschornak, Hartmut Stöcker (Ed.) Electrochemical Storage Materials

Also of interest Electrochemical Energy Systems Foundations, Energy Storage and Conversion Braun, 2018 ISBN 978-3-11-056182-1, e-ISBN 978-3-11-056183-8

Electrochemistry A Guide for Newcomers Baumgärtel, 2019 ISBN 978-3-11-044340-0, e-ISBN 978-3-11-043739-3

Power-to-Gas Renewable Hydrogen Economy for the Energy Transition Boudellal, 2018 ISBN 978-3-11-055881-4, e-ISBN 978-3-11-055981-1

Hydrogen-Air PEM Fuel Cell Integration, Modeling, and Control Tong, Qian, Huo, 2019 ISBN 978-3-11-060113-8, e-ISBN 978-3-11-060215-9

Physical Sciences Reviews e-ISSN 2365-659X

Electrochemical Storage Materials From Crystallography to Manufacturing Technology

Dirk C. Meyer, Tilmann Leisegang, Matthias Zschornak, Hartmut Stöcker

Editors Prof. Dr. rer. nat Dirk C. Meyer Technische Universität Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg [email protected] Dr. rer. nat. Tilmann Leisegang Technische Universität Bergakademie Freiberg Experimentelle Physik Leipziger Str. 23 09599 Freiberg

Dr. rer. nat. Matthias Zschornak Technische Universität Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg [email protected] Dr. rer. nat Hartmut Stöcker Technische Universität Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg [email protected]

Samara State Technical University, Samara Center for Theoretical Materials Science, Molodogvardeyskaya street 224, 443100 Samara, Russia [email protected]

ISBN 978-3-11-049137-1 e-ISBN (PDF) 978-3-11-049398-6 e-ISBN (EPUB) 978-3-11-049187-6 Library of Congress Control Number: 2018962402 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the Internet at http://dnb.dnb.de. © 2019 Walter de Gruyter GmbH, Berlin/Boston Typesetting: Integra Software Services Pvt. Ltd. Printing and binding: CPI books GmbH, Leck Cover image: SVEN JACHALKE PHOTOGRAPHY www.degruyter.com

Preface The efficient, safe, and sustainable storage of electricity in the form of chemical energy is among the most important challenges of present research and technological development. Beyond lithium-ion batteries, many other electrochemical concepts, which are potentially even better-suited for future applications on a wide scale, are under investigation. Accordingly, further optimization and development of electrochemical cells requires consideration of the system as a whole, taking into account the complex interplay of all individual components. Considering the availability of material resources, the environmental impact and requirements for recycling, the design of new concepts has to base upon the understanding of relevant processes from the atomic level to the macroscopic scale. Following these demands, this book aims to give a comprehensive overview on materials, processes and associated technological challenges for electrochemical energy storage aiming at bridging the gap between academic and industrial disciplines. On this account, international teams of authors from science and industry give insights into various essential aspects. The presentation follows all stages of the value and innovation chain, from raw materials via manufacturing to recycling as well as from crystallography across materials science to technology. The first three chapters give general introductions into the present battery market, the fundamental functional principles of batteries, and the progression of battery concepts, respectively. Chapter 4 focuses on battery materials and contains a comprehensive overview of computational methods to identify new candidates followed by reviews based on the relevant battery components: anodes, cathodes, separators, and electrolytes. The last part of the book presents selected reviews on the structural and electrochemical characterization of battery materials covering, for example, expedient state-of-the-art methods in the fields of spectroscopy, diffraction, and electron microscopy. This collection of chapters is not meant as an all-round introduction to the vast field of battery technology, but as an up-to-date starting point for those new to the field and for those seeking current trends and prospects. We hope that the insights of the present book may lead to original and sustainable developments helping accomplish the current challenges of battery research. The editors are actively involved in the coordination and processing of relevant collaborative research projects. For example, in the framework of the collaborative project “CryPhysConcept: Crystal physics for the future concept of electrochemical energy storage” (funded by the German Federal Ministry for Education and Research, promotional reference 03EK3029A), a systematic search for new material concepts for electrochemical energy storage was carried out. Important criteria were resource availability, environmental compatibility and recyclability. In summary, new systems based on high-valent cations have been identified, exhibiting a high potential to replace the lithium technology also for consumer applications. In the context of the https://doi.org/10.1515/9783110493986-201

VI

Preface

currently implemented follow-up project “R2R Battery: Tailored material systems and technologies for the role-to-role production of electrochemical energy storage on flexible substrates” (funded by the German Federal Ministry for Education and Research, promotional reference 03SF0542A) the development of aluminum-ion thin-film solid-state batteries and their cost-effective production are in the focus of our efforts. Among others, this project enriches the research program of the “Center for efficient high-temperature material conversion (ZeHS)” at our home institution “Technische Universität Bergakademie Freiberg”, for which up to the year 2020 in total 41.5 million € are available for construction and the acquisition of selected equipment (Dirk C. Meyer acts as the head of this project). We would like to thank Ms Kerstin Annassi from Projektträger Jülich for her professional support of the collaborative research projects mentioned above. Thanks to her support, the results included in this publication have been achieved. Experiences, knowledge, and viewpoints from the series of the “International Freiberg Conference on Electrochemical Storage Materials (EStorM)”, founded by two of the editors (Dirk C. Meyer and Tilmann Leisegang), have also been integrated into the book. Results of the first two conferences in 2013 and 2015 were discussed in the context of dedicated “AIP Conference Proceedings”, namely volume 1597 and 1765, respectively. This book is regarded as continuation of the proceedings series. The invitation of the publishing house DeGruyter was a great honor for us. For managing all the organizational parts and the kind assistance we are thankful to all staff members of DeGruyter involved in the preparation of this book, in particular, Kristin Berber-Nerlinger, Vivien Schubert, Jana Habermann, and Holger Kleeßen. We would like to gratefully acknowledge all authors who contributed and shared their expertise to cover the scientific field that we outlined in the beginning of this book project. We furthermore thank all referees for their efforts, helping to meet high scientific standards. For providing the cover photo we thank Sven Jachalke. Freiberg, August 15th 2018 Dirk C. Meyer, Tilmann Leisegang, Hartmut Stöcker, Matthias Zschornak (Eds.) and Theresa Lemser (Assistant)

Contents Preface

V

List of Contributing Authors

XIII

Robert Schmid, Christophe Pillot, Axel Thielmann and Hubertus Bardt 1 Introduction to energy storage: market analysis, raw materials, recycling, new concepts 1 1.1 Analysis of the emerging battery market and outlook 1 1.1.1 Global battery markets 1 1.1.2 LIB demand forecasts 3 1.1.3 LIB production capacity announcements 3 1.1.4 Regional distribution of cell production 4 1.1.5 LIB demand by applications 4 1.1.6 Developments, requirements, and future challenges across the segments EV, ESS, 3C 5 1.2 Assessment of chemical elements for new battery concepts 7 1.2.1 Parameters for assessment 8 1.3 Conclusions 14 References 14 Matthias Zschornak, Falk Meutzner, Jessica Lück, Arnulf Latz, Tilmann Leisegang, Juliane Hanzig, Melanie Nentwich, Jens Zosel and Perla B. Balbuena 2 Fundamental principles of battery design 17 2.1 Nernst equation 20 2.2 The Daniell cell 23 2.3 Reactions at electrified interfaces 24 2.4 Diffusion and migration in crystals 26 2.5 Crystallographic, crystal chemical, and crystal physical peculiarities 28 2.6 Classification of battery applications and types 34 References 37 Melanie Nentwich, Bianca Störr and Juliane Hanzig 3 Battery concepts: The past, the present, and research highlights 3.1 The past 41 3.2 The present 45 3.2.1 Flow accumulator 45

41

VIII

3.2.2 3.2.3 3.2.4 3.2.5 3.2.6 3.2.7 3.2.8 3.2.9 3.2.10 3.2.11 3.2.12 3.2.13 3.3 3.3.1 3.3.2 3.3.3 3.3.4 3.3.5 3.3.6 3.3.7 3.3.8 3.3.9 3.3.10 3.4

Contents

Lead-acid accumulator 47 Lithium ion accumulator 47 Lithium–iron sulphide battery 49 Lithium–sulphur and sodium–sulphur accumulator Metal fluoride accumulator 51 Metal–air battery 52 Mercury battery 52 Molten salt accumulator 53 Nickel–cadmium accumulator 54 Nickel–metal hydride accumulator 55 Silver oxide battery and accumulator 56 Zinc–manganese oxide batteries 57 Research highlights 58 High-valent metal batteries 58 3D printed cell 61 Hybrid accumulator/fuel cell 62 Liquid metal accumulator 63 Metal–air accumulator 64 Paintable accumulator 65 Super-iron accumulator 65 Tin–sulphur–lithium accumulator 66 Virus-enabled electrodes 66 Water accumulator 67 Conclusion and outlook 68 References 68

49

4 Battery Materials 75 Falk Meutzner, Tina Nestler, Matthias Zschornak, Pieremanuele Canepa, Gopalakrishnan S. Gautam, Stefano Leoni, Stefan Adams, Tilmann Leisegang, Vladislav A. Blatov and Dirk C. Meyer 4.1 Computational analysis and identification of battery materials 75 4.1.1 Introduction 75 4.1.2 Voronoi-Dirichlet partitioning 81 4.1.3 Bond valence methods 92 4.1.4 Density functional modelling and the materials project 100 4.1.5 Molecular dynamics 110 4.1.6 Summary 113 References 114

Contents

Falk Meutzner, Matthias Zschornak, Melanie Nentwich, Damien Monti and Tilmann Leisegang 4.2 Electrodes: definitions and systematisation – a crystallographers view 123 4.2.1 Definitions 124 4.2.2 Systematisation 128 4.2.3 Categorisation 134 4.2.4 Summary 136 References 136 Goriparti Subrahmanyam, Ermanno Miele, Remo Proietti Zaccaria and Claudio Capiglia 4.3 Nanostructured anode materials 138 4.3.1 Introduction 139 4.3.2 Intercalation/Deintercalation materials 141 4.3.3 Alloy/de-alloy materials 147 4.3.4 Conversion materials 152 4.3.5 Conclusions 154 References 155 Tina Weigel and Florian Schipper 4.4 Modification of cathode materials for Li batteries 157 4.4.1 Introduction 157 4.4.2 Surface coating 158 4.4.3 Doping 159 4.4.4 Summary 161 References 162 Max Stöber and Charaf Cherkouk 166 4.5 Positive electrodes based on Ion-implanted SrTiO3 4.5.1 Introduction 167 4.5.2 Synthesis and structural characterization of SrTiO3 single crystal 168 O2-electrodes 4.5.3 Oxygen diffusion and defect chemistry in strontium titanate 169 4.5.4 Oxygen solid electrolyte coulometry on ion-implanted SrTiO3 single crystals 170 4.5.5 Summary 172 References 173

IX

X

Contents

Tina Nestler, Elsa Roedern, Nikolai F. Uvarov, Juliane Hanzig, Giuseppe Antonio Elia and Mateo de Vivanco 4.6 Separators and electrolytes for rechargeable batteries: Fundamentals and perspectives 174 4.6.1 Introduction 175 4.6.2 Fundamentals and categorization 176 4.6.3 Solid electrolytes 179 4.6.4 Ionic liquids 202 4.6.5 Conclusion 208 References 209 Giovanni Battista Appetecchi 4.7 Safer electrolyte components for rechargeable batteries 220 4.7.1 Introduction 221 4.7.2 Components for lithium battery electrolytes 224 4.7.3 Electrolyte confinement in polymer hosts 227 4.7.4 Electrolytes based on ionic liquid media 230 4.7.5 Electrolytes based on polymer media 235 4.7.6 Future trends 243 4.7.7 Acronym glossary 246 References 247 5 Characterization methods 261 Thomas Köhler, Juliane Hanzig and Victor Koroteev 5.1 Optical spectroscopy as a tool for battery research 261 5.1.1 Raman spectroscopy 262 5.1.2 Fourier transform infrared spectroscopy 5.1.3 UV/Vis spectroscopy 274 5.1.4 Summary 276 References 277

268

Anastasia Vyalikh, Thomas Köhler, Tatiana Zakharchenko, Daniil M. Itkis, Andraž Krajnc and Gregor Mali 5.2 Magnetic resonance spectroscopy approaches for electrochemical research 282 5.2.1 Nuclear magnetic resonance 282 5.2.2 Electron paramagnetic resonance 298 5.2.3 Conclusion 303 References 306

Contents

Lyubov G. Bulusheva, Alexander V. Okotrub, Lada V. Yashina, Juan J. Velasco-Velez, Dmitry Yu. Usachov and Denis V. Vyalikh 5.3 X-ray photoelectron spectroscopy study of the interaction of lithium with graphene 311 5.3.1 XPS introduction 312 5.3.2 Fast or time-dependent XPS 314 5.3.3 Disclosing structural properties of a graphene/metal interface by XPS 315 5.3.4 Lithium interaction with a metal-supported graphene monolayer 317 5.3.5 Lithium interaction with few-layer graphene and N-graphene 318 5.3.6 A step forward: XPS under operation conditions 320 5.3.7 Summary 321 References 322 Hartmut Stöcker 5.4 X-ray diffraction methods References 330

324

Mikhail V. Avdeev, Ivan A. Bobrikov and Viktor I. Petrenko 5.5 Neutron methods for tracking lithium in operating electrodes and interfaces 330 5.5.1 Introduction 331 5.5.2 Neutron diffraction 332 5.5.2 Neutron reflectometry 341 5.5.3 Small-angle neutron scattering 346 5.5.4 Summary 348 References 350 Claudia Funke and Venkata Sai Kiran Chakravadhanula 5.6 Characterisation of battery materials by electron and ionmicroscopy techniques: a review 352 5.6.1 Scanning electron microscopy: particle morphology and surface structure 353 5.6.2 Transmission electron microscopy: high-resolution morphology and microstructure 356 5.6.3 Chemical information and analytical electron microscopy 361 5.6.4 Focused ion beam system (FIB) 366 5.6.5 Overall and lithium specific challenges 374 5.6.6 Summary 380 References 380

XI

XII

Contents

Max Stöber, Jens Zosel and Charaf Cherkouk 5.7 Oxygen solid electrolyte coulometry (OSEC) References 385

383

Wolfram Münchgesang and Ulrike Langklotz 5.8 Electrochemical analytical methods 385 5.8.1 Electrochemical cell 387 5.8.2 Cell components and their conductivity 388 5.8.3 Measurement set-up 389 5.8.4 Open circuit potential measurement 390 5.8.5 Potentiostatic cycling 391 5.8.6 Galvanostatic cycling 392 5.8.7 Cyclic voltammetry 393 5.8.8 Electrochemical impedance spectroscopy 394 5.8.9 Summary and outlook 397 References 397 Erik Berendes, Sebastian Socher and Ulrich Potthoff 5.9 Applied battery diagnosis 399 5.9.1 Impedance spectroscopy as a tool for applied battery diagnosis 399 5.9.2 SoC determination 400 5.9.3 Impedance-based temperature detection 402 5.9.4 Battery degradation analysis 404 5.9.5 Summary 405 References 406 Index

409

List of Contributing Authors Robert Schmid TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 Freiberg 09596, Germany [email protected] Chapter 1

Jessica Lück German Aerospace Center (DLR) Institute of Engineering Thermodynamics Pfaffenwaldring 38-40 70569 Stuttgart, Germany [email protected] Chapter 2

Christophe Pillot Avicenne Energy Paris, France [email protected] Chapter 1

Arnulf Latz German Aerospace Center (DLR) Institute of Engineering Thermodynamics Pfaffenwaldring 38-40 70569 Stuttgart, Germany [email protected] Chapter 2

Axel Thielmann Fraunhofer ISI Karlsruhe, Germany [email protected] Chapter 1 Hubertus Bardt Institut der deutschen Wirtschaft Köln e.V. Köln, Germany [email protected] Chapter 1 Matthias Zschornak TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 2, 4.1, 4.2 Falk Meutzner TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 2, 4.1, 4.2

https://doi.org/10.1515/9783110493986-202

Tilmann Leisegang TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 2, 4.1, 4.2 Juliane Hanzig TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 2, 3, 4.6, 5.1 Melanie Nentwich TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 2, 3, 4.2

XIV

List of Contributing Authors

Jens Zosel Kurt-Schwabe-Institut für Mess- und Sensortechnik e.V. Meinsberg (KSI) Kurt-Schwabe-Str. 4 04736 Waldheim, Germany [email protected] Chapter 2, 5.7 Perla B. Balbuena Texas A&M University, Artie McFerrin Department of Chemical Engineering 100 Spence St, TX 77843 College Station United States of America [email protected] Chapter 2 Bianca Störr TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 2, 3 Tina Nestler TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 4.1, 4.6 Pieremanuele Canepa University of Bath Department of Chemistry Bath BA2 7AY, United Kingdom [email protected] Chapter 4.1 Gopalakrishnan S. Gautam Princeton University Department of Mechanical and Aerospace Engineering Princeton, NJ 08544, USA [email protected] Chapter 4.1

Stefano Leoni Cardiff University School of Chemistry Cardiff CF10 3AT, UK [email protected] Chapter 4.1 Stefan Adams National University of Singapore Department of Materials Science & Engineering Engineering Drive 2 117579 Singapore, Singapore [email protected] Chapter 4.1 Dirk C. Meyer TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 4.1 Vladislav A. Blatov Samara National Research University Samara Center for Theoretical Materials Science Moskovskoye Shosse 34 Samara 443086, Russian Federation [email protected] Chapter 4.1 Damien Monti Institut de Ciència de Materials de Barcelona (ICMAB-CSIC) Campus UAB, E-08193 Bellaterra, Catalonia, Spain [email protected] Chapter 4.2 Goriparti Subrahmanyam Istituto Italiano di Tecnologia via Morego 30 Genova, Italy [email protected] Chapter 4.3

List of Contributing Authors

Ermanno Miele Istituto Italiano di Tecnologia via Morego 30 Genova, Italy [email protected] Chapter 4.3 Remo Proietti Zaccaria Istituto Italiano di Tecnologia via Morego 30 Genova, Italy [email protected] Chapter 4.3 Claudio Capiglia Nagoya Institute of Technology Gokiso, Showa, Nagoya, Japan [email protected] Chapter 4.3 Tina Weigel TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 4.4 Florian Schipper Bar-Ilan University Department of Chemistry Ramat-Gan 52900, Israel fl[email protected] Chapter 4.4 Max Stöber Helmholtz-Zentrum Dresden-Rossendorf Institute of Ion Beam Physics and Materials Research Bautzner Landstraße 400 01314 Dresden, Germany [email protected] Chapter 4.5, 5.7

XV

Charaf Cherkouk TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Straße 23 09596 Freiberg, Germany [email protected] Chapter 4.5, 5.7 Elsa Roedern Swiss Federal Laboratories for Materials Science and Technology (Empa) Materials for Energy Conversion Überlandstrasse 129 8600 Dübendorf, Switzerland [email protected] Chapter 4.6 Nikolai F. Uvarov Institute of Solid State Chemistry and Mechanochemistry Siberian Branch of the Russian Academy of Sciences Kutateladze 18, 630128 Novosibirsk, Russia [email protected] Chapter 4.6 Giuseppe Antonio Elia Technische Universität Berlin Research Center of Microperipheric Technologies Gustav-Meyer-Allee 25 13355 Berlin, Germany [email protected] Chapter 4.6 Mateo de Vivanco TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg, Germany [email protected] Chapter 4.6

XVI

List of Contributing Authors

Giovanni Battista Appetecchi ENEA, Agency for New Technologies Energy and Sustainable Economic Development Department for Sustainability (SSPT) Division Sustainable Materials (PROMAS), Materials and Physicochemical Processes Laboratory (MATPRO), Casaccia Research Center, Via Anguillarese 301, Rome, Italy [email protected] Chapter 4.7 Thomas Köhler TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg, Germany [email protected] Chapter 5.1, 5.2 Victor Koroteev Nikolaev Institute of Inorganic Chemistry Siberian Branch of the Russian Academy of Sciences 3 Acad. Lavrentiev Ave. 630090 Novosibirsk, Russia [email protected] Chapter 5.1 Anastasia Vyalikh TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg, Germany [email protected] Chapter 5.2 Tatiana Zakharchenko Moscow State University Department of Chemistry Leninskie gory 119991 Moscow, Russia [email protected] Chapter 5.2

Daniil M. Itkis Moscow State University Department of Chemistry Leninskie gory 119991 Moscow, Russia [email protected] Chapter 5.2 Andraž Krajnc National Institute of Chemistry Hajdrihova 19 1001 Ljubljana, Slovenia [email protected] Chapter 5.2 Gregor Mali National Institute of Chemistry Hajdrihova 19 1001 Ljubljana, Slovenia [email protected] Chapter 5.2 Lyubov G. Bulusheva Nikolaev Institute of Inorganic Chemistry Siberian Branch of the Russian Academy of Sciences 3 Acad. Lavrentiev ave. 630090 Novosibirsk, Russia [email protected] Chapter 5.3 Alexander V. Okotrub Nikolaev Institute of Inorganic Chemistry Siberian Branch of the Russian Academy of Sciences 3 Acad. Lavrentiev ave. 630090 Novosibirsk, Russia [email protected] Chapter 5.3 Lada V. Yashina Lomonosov Moscow State University Leninskie gory 119991 Moscow, Russia [email protected] Chapter 5.3

List of Contributing Authors

Juan J. Velasco-Velez Max Planck Institute for Chemical Energy Conversion Department of Heterogeneous Reactions Stiftstr. 34–36 45470 Mülheim an der Ruhr, Germany [email protected] Chapter 5.3 Dmitry Yu. Usachov Saint Petersburg State University Department of Physics Ulianovskaya Street, 3 St Petersburg, 199034, Russia [email protected] Chapter 5.3 Denis V. Vyalikh Donostia International Physics Center (DIPC) Departamento de Fisica de Materiales and CFM-MPC UPV/EHU 20080 San Sebastian, Spain [email protected] Chapter 5.3 Hartmut Stöcker TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg, Germany [email protected] Chapter 5.4 Mikhail V. Avdeev Joint Institute for Nuclear Research Frank Laboratory of Neutron Physics Joliot-Curie 6 Dubna, Russia [email protected] Chapter 5.5

XVII

Ivan A. Bobrikov Joint Institute for Nuclear Research Frank Laboratory of Neutron Physics Joliot-Curie 6 Dubna, Russia [email protected] Chapter 5.5 Viktor I. Petrenko Joint Institute for Nuclear Research Frank Laboratory of Neutron Physics Joliot-Curie 6 Dubna, Russia [email protected] Chapter 5.5 Claudia Funke TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg, Germany [email protected] Chapter 5.6 Venkata Sai Kiran Chakravadhanula Karlsruhe Institute of Technology (KIT) Electron Microscopy Spectroscopy Laboratory Hermann-von-Helmholtz-Platz 1 Building 640 76344 Eggenstein-Leopoldshafen, Germany [email protected] Chapter 5.6 Wolfram Münchgesang TU Bergakademie Freiberg Institute of Experimental Physics Leipziger Str. 23 09599 Freiberg, Germany [email protected] Chapter 5.8

XVIII

List of Contributing Authors

Ulrike Langklotz Technische Universität Dresden Institut für Werkstoffwissenschaft Helmholtzstr. 7 01067 Dresden, Germany [email protected] Chapter 5.8

Sebstian Socher Fraunhofer Institute for Transportation and Infrastructure Systems Transportation, Energy and Environment Zeunerstr. 38 01069 Dresden, Germany Chapter 5.9

Erik Berendes Fraunhofer Institute for Transportation and Infrastructure Systems Transportation, Energy and Environment Zeunerstr. 38 01069 Dresden, Germany [email protected] Chapter 5.9

Ulrich Potthoff Fraunhofer Institute for Transportation and Infrastructure Systems Transportation, Energy and Environment Zeunerstr. 38 01069 Dresden, Germany ulrich.potthoff@ivi.fraunhofer.de Chapter 5.9

Robert Schmid, Christophe Pillot, Axel Thielmann and Hubertus Bardt

1 Introduction to energy storage: market analysis, raw materials, recycling, new concepts Keywords: market analysis, raw materials, recycling, new concepts, evaluation

1.1 Analysis of the emerging battery market and outlook 1.1.1 Global battery markets The role of batteries as decentralized energy storage technology has dramatically changed in the past few years and is facing a fundamental global transformation process in the next decades. This process is driven by the need for reducing greenhouse gas (GHG) emissions in particular in the transport sector and by the energy transformation guided by grid expansions, modernizations and increasing decentralization on distribution and local grid levels. The trend towards green as well as autarc, autonomous and thus individualized energy solutions can be observed not only in the transport sector with autonomous and electric cars (smart mobility) but also in the energy sector with decentralized energy storage systems connected to a smart grid [2]. It is connected to all sectors, e.g. industry 4.0, smart health, etc. and finally also changing dramatically the consumer or end-user market. Mobile devices such as digital cameras, smartphones, navigation devices and portable computers are normality nowadays. Nevertheless, smart and small medical devices, wearables for leisure activities, power tools, etc. gain interest with the improvement of energy storage devices with higher energy densities. Advanced high-energy lithium-ion batteries (LIBs) provide the highest energy densities among the electrochemical energy storage technologies and are currently diffusing in electric vehicle (EV), stationary storage and portable devices for consumer markets. Although the LIB market is drastically growing with annual rates of up to 26 % on average (in terms of battery capacity demand), the dominating battery technology in terms of capacity, however, is still the lead-acid battery, representing about 90 % of the global demand in volume around 2010 and 80 % in 2017. The total annual battery market is currently on the level of (60–80) billion USD representing approximately (400–500) GWh [3]. LIBs have succeeded because of their high gravimetric and volumetric energy densities and have a market share of about a third in terms of value among the whole battery market. The growth is driven by the decreasing cell costs less than 200 $/kWh for a standard cylindrical cell and

https://doi.org/10.1515/9783110493986-001

2

1 Introduction to energy storage

approaching the 100 $/kWh mark in the next years. This cost degression is due to increasing capacity demand and production and thus due to scale effects. Other mature battery technologies like NiCd (still partly used in power tools) and NiMH (still used in hybrid electric vehicles, HEVs) are slowly declining and expected to disappear from the market in the next decade. The diversification of future battery applications, however, will also broaden the range of battery technologies. Emerging technologies like Na-based, metal-sulphur or redox-flow batteries are developing and may lead to attractive solutions, e.g. with respect to cost and resource availability advantages. However, due to very strict requirements on volumetric energy densities, these technologies are expected to be relevant for stationary or other special markets and less for the most significant EV market. In general, each technology has its own strengths and weaknesses and none of them can satisfy all user demands. So, a broader application-specific technology portfolio is even urgently needed in order to provide alternative technology solutions in the future. This becomes clear when looking at the emerging future market demand for LIB in Figure 1.1. The extremely dynamic development of the global LIB demand will lead to the TWh level latest around 2030 and grow further after 2030. Thus, the demand will by far exceed the demand for lead-based batteries as of today. LIB will very soon transform into the dominating energy storage technology, raising questions concerning not only the technology development and producibility but also the CO2 footprint, energy demand, efficiency, resource availability, recycling strategies and, not the least, second life and chance for technology substitution.

1000 900 800 700

GWh

600 500 400 300 200 100 0 2010

2011 2012 2013 2014 Demand (pessimistic) 1 BNEF 2017 Avicenne 2015

2015 2016 2017 2018 Demand (trend) 2 BNEF 2017 Avicenne 2017

2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 Demand (optimistic) 3 LIB Capacities LIB Production B3 Corp 2015 B3 Corp 2016 Avicenne 2013 Roskill 2017 Roskill 2017 Lux 2017

Figure 1.1: Global LIB cell production capacities vs. demand 2010 to 2030 (Sources: [3–16]). The graph shows the installed LIB capacities in blue, whereby the produced capacities are shown in red. The demand (green) is supposed to increase in the following years so that approximately between 2025 and 2027, the demand will be higher than the produced capacity.

1.1 Analysis of the emerging battery market and outlook

3

1.1.2 LIB demand forecasts Regarding the forecasts for the LIB cell demand, three market diffusion scenarios have been developed in [14]: a “technology diffusion scenario” assuming the demand for LIB by applications upon the state of technology and costs (pessimistic), an “early diffusion scenario” (trend) assuming a push, e.g. from original equipment manufacturer (OEM) by supporting EV development and market introduction, and a “forced diffusion scenario” (optimistic) assuming, e.g. state aid measures, regulation, legislation or even prohibition of combustion engines as discussed in some countries (e.g. in Germany) in 2017. The scenarios for LIB demand from EVs have been combined with the demand for LIB from stationary (ESS) and portable device (3C) markets. The scenarios are indicated in Figure 1.1 in green lines. A number of market studies and forecasts from 2013 to 2017 have been integrated for 2020 and 2025 in the figure to analyse how strong different forecasts deviate and to understand the uncertainties in the market development. It can be seen that principally all forecasts are within the window between the three market diffusion scenarios. Until 2025, the uncertainty of the market development leads to a factor of four (200 GWh to 800 GWh demand), which means a 12 % to 26 % average annual growth rate until then.

1.1.3 LIB production capacity announcements To face this increasing demand for LIB cells accordingly, production capacities need to be built up in the near future. Based on the currently existing cell production capacities and the global announcements from established and new cell producers until 2025, the LIB cell production capacities (blue line) have been identified. Compared to 15 GWh added production capacity between 2013 and 2016, in the next years around 50 GWh will be added annually, leading to around 700 GWh production capacities until 2025. The announcements include established players such as Panasonic (JP), LG Chem (KR), Samsung (KR), SKI (KR), BYD (CN), Lishen (CN), CATL (CN), CALB (CN), OPTIMUM (CN) and several further Chinese cell producers. Also, new players such as TerraE (DE), Northvolt (SW), Boston Energy (US, AU), Energy Absolute (Thailand) as well as Tesla (USA) are included, while accounting for the different stages of expansion [15, 16]. However, the degree of utilization of the capacity of a factory is never 100 %. Once it goes over 85 % in the long term, manufacturers tend to think about expanding capacity. When it comes to genuinely used production capacity, therefore, it is advisable to expect values only up to 85 %. Also, factories never produce only “good” components. Well-established factories have a yield of over 90 % “good” products. Yield figures in battery production are today still below this in certain cases [16]. Combining these facts, a real production capacity (red dashed line) can be calculated. Furthermore, typically quality and usability of cells for certain applications, regional distribution and availability and capacities that are hold on stock or time delays (e.g.

4

1 Introduction to energy storage

for new player) during the ramp-up phase would have to be accounted for, which would further reduce and shift the real production capacities into the future. Comparing the red dashed line with the green (middle) line of the trend scenario (which is currently regarded as the most realistic and feasible scenario according to the almost 100 GWh or more LIB cell demand in 2017, it can be seen that latest around 2025 very massive additional production capacities would be needed to be built up in order to address the growing demand. 1.1.4 Regional distribution of cell production Before the competition between the Japanese, Chinese and Korean manufacturers began around 2000, the battery industry (NiCd, NiMH, LIB) was concentrated in Japan with almost 85 % but has declined for the benefit of Korea. Samsung SDI and LG drastically increased their market share and dominated the market together with Panasonic in the last years [6, 17]. Since 2015, Chinese cell producer like BYD, Lishen, CATL and many others are strongly building up production capacities and have a current share of around 70 % of the global production capacities. With the global capacities to be built up until 2025, the share of Chinese cell manufacturers is expected to arrive at merely around 50 %. 1.1.5 LIB demand by applications In Figure 1.2(a)–(e), the global LIB demand is broken down to the three main sectors for battery demand, which can be characterized into the following, each having different profiles of requirement:

70 GWh (2015)

50 Cellproducers/Locations/Ramp-up data worldwide (last update 01/12/2017). Michaelis S, Maiser E, Kampker A, Heimes H, Lienemann C, Wessel S, et al. VDMA Batterieproduktion, Roadmap Batterie-Produktionsmittel 2030, Update 2016. Frankfurt: VDMA Verlag GmbH; 2016. Avicenne Energy (Christophe Pillot). Future trends in the rechargeable battery market. Batteries. Paris: Avicenne Energy, 2017 Thielmann A, Sauer A, Wietschel M. Gesamt-Roadmap Lithium-Ionen-Batterien 2030. Karlsruhe: Fraunhofer ISI, 2015. Wietschel M, Thielmann A, Plötz P, Gnann T, Sievers L, Breitschopf B, et al. Perspektiven des Wirtschaftsstandorts Deutschland in Zeiten zunehmender Elektromobilität. Karlsruhe: Fraunhofer ISI, 2017. U.S. Geological Survey. Mineral commodity summaries. Reston, VA: U.S. Geological Survey, 2016. http://dx.doi.org/10.3133/70170140. Thielmann A, Sauer A, Schnell M, Isenmann R, Wietschel M. Technologie-Roadmap Stationäre Energiespeicher 2030. Karlsruhe: Fraunhofer ISI, 2015. Olivetti EA, Ceder G, Gaustad GG, Fu X. Lithium-Ion battery supply chain considerations: analysis of potential bottlenecks in critical metals. Joule. 2017;1:229. Vidal O, Goffé B, Arndt N. Metals for a low-carbon society. Nat Geosci. 2013;6:894. Bardt H. Raw materials in the field of electrochemical energy storage – A risk analysis. AIP Conf Proc. 2016;1765:020002. DOI: 10.1063/1.4961894. Wellmer F-W. Reserves and resources of the geosphere, terms so often misunderstood. Is the life index of reserves of natural resources a guide to the future? Z dt Ges Geowiss. 2008;159/ 4:575–90. Leisegang T, Treffer F. Recycling of electrochemical storage devices. AIP Conf Proc. 2016;1765:020006. DOI: 10.1063/1.4961898. Recycling Rates of Metals – A Status Report. United Nations Environment Programme, 2011. ISBN: 978-92-807-3161-3. vbw/IW Köln/IW Consult. Rohstoffsituation der bayerischen Wirtschaft. Munich/Cologne: Vereinigung der Bayerischen Wirtschaft e. V., 2015 UN conference on Environment and Development (Rio de Janeiro, 1992).

Matthias Zschornak, Falk Meutzner, Jessica Lück, Arnulf Latz, Tilmann Leisegang, Juliane Hanzig, Melanie Nentwich, Jens Zosel and Perla B. Balbuena

2 Fundamental principles of battery design Abstract: With an increasing diversity of electrical energy sources, in particular with respect to the pool of renewable energies, and a growing complexity of electrical energy usage, the need for storage solutions to counterbalance the discrepancy of demand and offer is inevitable. In principle, a battery seems to be a simple device since it just requires three basic components – two electrodes and an electrolyte – in contact with each other. However, only the control of the interplay of these components as well as their dynamics, in particular the chemical reactions, can yield a highperformance system. Moreover, specific aspects such as production costs, weight, material composition and morphology, material criticality, and production conditions, among many others, need to be fulfilled at the same time. They present some of the countless challenges, which make battery design a long-lasting, effortful task. This chapter gives an introduction to the fundamental concepts of batteries. The principles are exemplified for the basic Daniell cell followed by a review of Nernst equation, electrified interface reactions, and ionic transport. The focus is addressed to crystalline materials. A comprehensive discussion of crystal chemical and crystal physical peculiarities reflects favourable and unfavourable local structural aspects from a crystallographic view as well as considerations with respect to electronic structure and bonding. A brief classification of battery types concludes the chapter. Keywords: battery, electrochemistry, ionic transport, diffusion, migration, interface kinetics, crystallography

Electrical energy is generally considered as highest quality form of energy, since it can easily be converted into other forms such as mechanical, thermal, or radiation energy. The drawback of electrical energy is the problem of its storage. Conventional capacitors provide only limited energy density which is why storage in other forms is inevitable for most applications. The conversion to chemical energy is a very efficient way to increase energy density by orders of magnitude but, due to the dynamics of the involved chemical processes, at the cost of power density (see Figure 2.1). This This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Zschornak, M., Meutzner, F., Lück, J., Latz, A., Leisegang, T., Hanzig, J., Nentwich, M., Zosel, J., Balbuena, P. B. Fundamental principles of battery design. Physical Sciences Reviews [Online] DOI: 10.1515/psr-2017-0111 https://doi.org/10.1515/9783110493986-002

18

2 Fundamental principles of battery design

3.6 ms

0.36 s

3.6 s

106

H2

105

Ga so lin ec om bu st io n

co mb us tio n

Ca pa cit or s

36 s

lc ell s fu e H2

Lipr im ar y

/M

103

Li-

H

io n

1h

Ni

Specific power (W kg-1)

104

El Ca ectr pa oc cit he or mi ca s l

102

Pb Pb O2

I

101

10 h

1

10-2

10-1

1

101

102

103

Specific energy (Wh kg-1) Figure 2.1: Ragone plot comparing specific energy and power of electrical and electrochemical energy storage technologies with conventional combustion technologies. Redrawn from Ref. [1] with additional data from Ref. [2].

principle is realised for example in modern super- or ultracapacitors. In the most usual case, the electrostatic capacitance arises from charge separation at the electrodes’ interfaces with ions in the electrolyte shielding the external electric field, the so-called Helmholtz double layers. By adding fast electronic charge-transfer from additionally reacting adsorbed ions at the electrode, the capacitance can be augmented by this supplementary electrochemical pseudocapacitance. Nevertheless, the accompanied increase in energy density is limited predominantly by the saturated amount of adsorbate at the interface. One solution to overcome this constraint of a limited interface region is to realise the electronic charge transfer and accompanied change of oxidation states for a large volume as is provided by batteries.

2 Fundamental principles of battery design

19

Batteries contain electrochemical cells that convert the chemical energy stored in the active materials into electrical energy by means of ongoing exergonic electrochemical reactions. A cell is made up of two half-cells each consisting of an electronically conducting electrode, to deliver electric potential and current via contacts to an external circuit, and a volume of electrolyte interfacing the electrode. Both spatially apart half-cells are connected in series by a separator that prohibits electronic but permits ionic conduction in combination with the electrolyte to balance the disproportion of charge and to close the charge circuit. A schematic setup of a basic cell, the so-called Daniell cell (see section 2.2 for details), is shown in Figure 2.2.

Negative electrode Anode (oxidation)

Positive electrode Cathode (reduction) V

e–

—

e– R

+

A

Zn

Cu salt bridge KCI

ZnSO4 Zn(s)

Zn2+ + 2e–

CuSO4 Cu2+ + 2e–

Cu(s)

Zn(s) | ZnSO4(aq) || CuSO4(aq) | Cu(S) Figure 2.2: Daniell cell with oxidation of zinc at the anode and reduction of copper at the cathode. The cell voltage (V) will provide a current (A) for the outer electric circuit with resistance R.

The oxidation takes place in one half-cell increasing the oxidation state of the involved ionic species and depositing electrons into the respective electrode, the anode. During the discharge process, without externally applied electric potential, this electrode acquires a negative charge and corresponds to the electrode with the negative electric potential, also called negative mass. In the other half-cell, the reduction takes place decreasing the oxidation state of the participating ionic species and depleting the amount of electrons in the respective electrode, the cathode. During the discharge process, this electrode acquires a positive charge and corresponds to the electrode with the positive electric potential, also called positive mass. As the battery is being charged, an externally applied electric

20

2 Fundamental principles of battery design

potential reverses ionic migration in the electrolyte and therefore anode and cathode and respectively the location of oxidation and reduction, but the negative and positive mass remain fixed. A more detailed description about these terms is given in chapter 4.2.

2.1 Nernst equation The fundamental equation that relates the battery’s cell potential to the concentration of reacting chemical species is the Nernst equation [4]. The charged battery is in a thermodynamically non-equilibrium state with an excess of free enthalpy G [J], also known as Gibbs free energy, i. e. chemically stored energy. This excess drives the chemical redox reaction towards the equilibrium state and is responsible for the electronic charge transfer with the electrodes and the accompanied electric potential providing the electric energy for the external circuit. The change in free enthalpy ΔG following the redox reaction induces a change in electrode potential ΔE [V] according to ΔG = – νe eΔE with νe being the electron transfer number for the reaction and e [C] the elementary charge. In the picture of a grand-canonical ensemble, i. e. systems with boundaries open to energy as well as particle transfer, ΔG is directly related to the chemical potential and can be derived from Boltzmann factors with Boltzmann constant k [J K−1] respecting the probability for the ions to be in the oxidised or reduced state. The Nernst equation ΔE = ΔE0 +

kT aOx ln νe e aRed ,

named after the German physicist and chemist Walter Nernst, reflects this balance with the logarithm of the reaction quotient as the ratio of chemical activities aOx and aRed for the oxidised and reduced form of the relevant ionic species, respectively. Since potential differences can only be measured as a voltage between two electrodes and the theoretical normalisation with respect to the vacuum energy is less accurate than relative voltage measurements, the standard hydrogen electrode H+/H2 is taken as a reference and assumed to be at 0 V for all temperatures T [K]. For standard ambient conditions (IUPAC-SATP), i. e. at T = 298.15K, a pressure p = 101.325 kPa, a concentration c = 1 mol/l with activity a = 1, and pH = 0, the respective standard reduction electrode potentials E0 of common redox pairs are given in the electrochemical series. To present an overview, Table 2.1 summarises typical electrode potentials ranging from about −3 to +3 V. The full cell voltage results in 0 0 – EOx . ΔE0 = ERed

3 3 2 2

U3+ Al3+ Ti2+ V2+

Cr U4+ In3+

3+

Mn Cr2+ 2H2O Cd(OH)2 Zn2+

3 1 1

2 2 2 2 2

1 1 1 2 1 2 2 2 1 3 3 2 3

Li+ Rb+ K+ Ra2+ Cs+ Ba2+ Sr2+ Ca2+ Na+ La3+ Cd3+ Mg2+ Sc3+

2+

νe

Oxidised form

Cr U3+ In2+

Mn Cr H2 + 2OH− Cd + 2OH− Zn

U Al Ti V

Li Rb K Ra Cs Ba Sr Ca Na La Ce Mg Sc

Reduced form

References [5] [5] [5] [5] [5] [5] [5] [5, 9] [5, 9] [5] [5] [5, 9] [5] [5] [5] [5] [5] [5] [5] [5] [5] [5, 7, 9] [5, 9] [5] [5]

E0 (V) −3.05 −2.93 −2.93 −2.92 −2.92 −2.91 −2.89 −2.87 −2.71 −2.52 −2.48 −2.36 −2.09 −1.79 −1.66 −1.63 −1.19 −1.18 −0.91 −0.83 −0.81 −0.76 −0.74 −0.61 −0.49

Table 2.1: Electrochemical series of common redox pairs.

Pu Br2

NO3– 4+

+ 4H

+

Hg22 + NO3– – + 2H Ag+ Hg2+ ClO− + H2O 2Hg2+

+

MnO24 – – + 2H2 O Hg2SO4 BrO− + H2O Fe3+

AgBr Sn4+ Cu2+ Bi3+ AgCl Hg2Cl2 Cu2+ ClO4– + H2 O Cu+ I3– I2 MnO4–

Oxidised form

3 1 2

1 1 2 2 2

2 2 1 2

1 2 1 3 1 2 2 1 1 2 2 1 2

νe

Hg22 + NO + 2H2O Pu3+ 2Br−

NO2 + H2O Ag Hg Cl− + 2OH−

2Hg + SO42– Br− + 2OH− Fe2+ 2Hg

Ag + Br− Sn2+ Cu+ Bi Ag + Cl− 2Hg + 2Cl− Cu ClO3– + 2OH – Cu 3I− 2I− MnO4– MnO2 + 4OH−

Reduced form

0.96 0.97 1.09

0.80 0.80 0.86 0.89 0.92

0.62 0.76 0.77 0.79

0.07 0.15 0.16 0.20 0.22 0.27 0.34 0.36 0.52 0.53 0.54 0.56 0.60

E0 (V)

(continued )

[5, 9] [5] [5]

[5] [5, 7, 9] [5] [5] [5]

[5] [5] [5, 9] [5]

[5] [5, 9] [5] [5] [5] [5] [5, 7–9] [5] [5, 9] [5] [5, 9] [5] [5]

References

2.1 Nernst equation

21

1 1 2

3 1 2 2 1 1 2 2 2 3

1 2

In3+ Tl+ Co2+ Ni2+ AgI In+ Sn2+ Pb2+ O2 + H2O Fe3+

Ti4+ H+

2 2 2 1 2

νe

In Ti3+ PbSO4

2+

S Fe2+ In3+ Cr3+ Cd2+

Oxidised form

Table 2.1 (continued )

Ti3+ H2

Pb + SO24 – In Tl Co Ni Ag + I– In Sn Pb HO2– + OH – – Fe

In Ti2+

+

S2− Fe In+ Cr2+ Cd

Reduced form

[5] [5] [5, 9] [5, 8] [5] [5] [5, 9] [5, 9] [5] [5]

−0.34 −0.34 −0.28 −0.23 −0.15 −0.14 −0.14 −0.13 −0.08 −0.04 [5] [5–9]

[5] [5] [5]

−0.40 −0.37 −0.36

0.00 0.00

[5, 9] [5] [5] [5] [5, 9]

References

−0.48 −0.44 −0.44 −0.41 −0.40

E0 (V)

S2 O28 – O3 + 2H+ F2

Mn3+ 2HBrO + 2H+ Ce4+ 2HClO + 2H+ Pb4+ Au+ H2O2 + 2H+ Co3+ Ag2+

Cr2 O27 – + 14H + Cl2 Au3+ MnO4– + 8H +

O2 + 4H+ ClO4– + 2H + MnO2 + 4H+ O3 + H2O

Oxidised form

2 2

1 2 1 2 2 1 2 1 1 2

2 3 5

4 2 2 2 6

νe

2SO24 – O2 + H2O 2F−

2.07 2.87

1.51 1.60 1.61 1.63 1.67 1.69 1.78 1.81 1.98 2.05

1.36 1.40 1.51

2Cl− Au Mn2+ + 4H2O Mn2+ Br2 + 2H2O Ce3+ Cl2 + 2H2O Pb2+ Au 2H2O Co2+ Ag+

1.23 1.23 1.23 1.24 1.33

E0 (V)

2H2O ClO3– + H2 O Mn2+ + 2H2O O2 + 2OH− 2Cr3+ + 7H2O

Reduced form

[5, 8, 9] [5, 6, 8, 9]

[5] [5] [5] [5] [5] [5] [5] [5] [5] [5]

[5, 9] [5] [5, 6, 9]

[5, 9] [5] [5] [5, 8] [5, 6, 9]

References

22 2 Fundamental principles of battery design

2.2 The Daniell cell

23

2.2 The Daniell cell The Daniell cell is considered the first electrochemical cell able to deliver current for a long time. It was invented by John F. Daniell in 1836 and is based on the observations of Luigi Galvani in 1786 on frog legs and Alessandro Volta’s pile of alternating copper and zinc discs in 1800, which is considered the very first “wet battery cell” [10]. A Daniell cell consists of a zinc and a copper electrode immersed in zinc sulfate and copper(II) sulfate electrolyte solutions, respectively. Considering the metal electrodes, zinc is less noble than copper with a work function of 4.34 eV [11] and 4.65 eV [12], respectively. Furthermore, Zn has a much higher solution pressure [13] causing a significantly higher concentration of Zn2 + in the zinc sulfate than Cu2 + in the copper (II) sulfate solution as well as a significantly higher concentration of electrons in the zinc than in the copper electrode. The metal ions in both electrolyte solutions will stay close to the metal electrode interfaces forming the Helmholtz double layer, with an accompanied charge separation that causes the electric potential difference between each electrode/electrolyte interface. The full cell as combination of both half-cells is driven by the oxidation of metallic zinc ZnðsÞ in the solid state (s) and the simultaneous + reduction of copper(II) ions Cu2ðaq Þ in solution (aq) (see Figure 2.2) according to ZnðsÞ

+ 0.76 V

!

+ – Cu2ðaq Þ + 2e

+ – Zn2ðaq Þ + 2e

+ 0.35 V

! CuðsÞ

0 0 with standard electrode potentials EZn 2 + =Zn = – 0.76 V and ECu2 + =Cu = + 0.35 V providing a total electrode potential difference of + 1.11 V for the full redox reaction: + ZnðsÞ + Cu2ðaq Þ

+ 1.11 V

!

+ Zn2ðaq Þ + CuðsÞ

As soon as the outer electronic circuit is closed, the redox potentials can act and supply electrical energy by means of the electrode potential difference and the used current. The ongoing reactions will accumulate positively charged Zn2 + ions in the electrolyte of the anode half-cell (referred to as anolyte) and SO2– 4 in the electrolyte of the cathode half-cell (referred to as catholyte) due to the depletion of Cu2 + ions. This imbalance of ionic charge between both electrolytes and the resulting electrostatic potential as well as the local excess of product (anode) and shortage of reactant (cathode) species would saturate the reactions at some point. To counter this saturation, a salt bridge (separator with electrolyte) connects the electrolytes of both half-cells. It is made up of a chemically inert salt with respect to the electrolytes, e. g. Na2 SO4 , possibly with similar conductivity of anions and cations. It + ions from the salt bridge to neutralise the Zn2 + and SO2– allows SO2– 4 and Na 4 excess, respectively. The concentration imbalance of the salt’s remaining ions and

24

2 Fundamental principles of battery design

the accompanied difference in chemical potential between the electrolyte/salt bridge interfaces will be reduced by ionic migration of predominantly Na + ions from the anode and SO2– 4 ions from the cathode half-cell through the bridge. The whole cell is a combined series of electrochemical gradients which induce mass and charge transport and determine the dynamics of the battery discharge process towards the equilibrium state. Specific electrochemical cells with a solid electrolyte will be discussed in detail in chapter 4.6.

2.3 Reactions at electrified interfaces Regardless of the choice of system, electrochemical storage of energy is largely based on interfacial reaction and transport processes, which determine the overall performance. In principle, all electrochemical systems have at least two chemically different but conducting phases. At the phase boundary, free charge carriers are rearranged due to different properties of the phases. Thus, the interface is electrified and local electric fields are produced. Those formed space charges and potential gradients affect reactions at or across phase boundaries. For electrochemical systems with a solid electrode and a liquid electrolyte, the charged region at the interface is called electrochemical double layer. In the twentieth century, the conception of the electrochemical double layer was developed. With the work of Helmholtz [14] on charge distributions in conducting materials, the term electrical double layer was introduced for the formation of two oppositely charged layers at the surface. Including work done in the field of electrocapillarity of liquids [15, 16], a first double layer model regarding the metal–electrolyte interface was established by Chapman in 1913 [17]. Stern finalised the concept of the double-layer structure describing a fixed and a diffusive part of the charged layer in the liquid phase [18]. Because of this historical development, the electrochemical double layer is pictured as an inner and outer Helmholtz layer, which denotes specifically and nonspecifically adsorbed ions, and a diffuse or Gouy–Chapman layer in the electrolyte, see Figure 2.3. Based on this

electrode

electrolyte

Gouy-Chapman layer

inner Helmholtz plane outer Helmholtz plane

Figure 2.3: Schematic illustration of the electrochemical double layer formed at the electrode–electrolyte interface.

2.3 Reactions at electrified interfaces

25

concept of the structure, the theory of the electrochemical double layer characterizing charges and electric potentials was extended by several aspects like e. g. dipole properties or volumetric effects [19, 20]. A summary of the theoretical models of the electrochemical double layer combined with the field of electrocapillarity can be found in the classical paper of Grahame [21]. In parallel to this development, the reaction kinetics at the electrolyte– electrode interface were investigated. A general expression for the transfer rate of electrons was derived from kinetic theory for the case of hydrogen evolution and metal deposition [22–24]. This relation describing the solid electrode current density ise [A m−2] for an arbitrary charge transfer reaction is known as the Butler–Volmer equation  α e   α e  A C , ise = i0 exp ηs – exp – η kT kT s where αA and αC with αA + αC = 1 are the weighting factors of the anodic and cathodic part of the reaction due to the overpotential ηs [V]. The exchange current density i0 [A m−2] depends on the composition of the reactants and products. The driving force of the reaction is given by the measured overpotential ηs , which describes the difference of electrochemical potential of reduced and oxidized state (further details given in chapter 4.2). In Ref. [25], the Butler–Volmer equation was used for the first time to describe the intercalation reaction in a lithium ion battery. This has now become the standard approach for modelling electrode reaction kinetics in electrochemical systems. A first relation between the structure of the double layer and the reaction rate of the electrode was found by Frumkin [26]. Here, the overvoltage of a metal deposition reaction was deduced by considering the potential distribution in the double layer. Further developments lead to the generalised Frumkin–Butler– Volmer equation [27], which adapts the classical Butler–Volmer relation to capacitive processes of the double layer. The general form of the Butler–Volmer equation is preserved; however, effective transfer coefficients α* including the structure of the double layer are introduced. This corresponds to a charge free layer between electrode and diffuse double layer, namely the Stern layer, and an electron transfer at the outer Helmholtz plane, which depends on the thickness of the Stern layer. As a result, one implicitly neglects ion adsorption on the inner Helmholtz plane and a dynamic behaviour of the double layer. Approaches have been made to strengthen the connection of the Butler–Volmer equation to electrochemical systems. Derivations based on concepts of non-equilibrium thermodynamics were presented [28, 29]. Electron transfer reactions, where no breaking or establishing of chemical bonds occur, can also be expressed in similar terms [30]. Furthermore, a sequential treatment of an electrode reaction is possible by considering ion adsorption and space charges at an electrified interface [31]. Nevertheless, an accurate description of reaction kinetics at electrochemical

26

2 Fundamental principles of battery design

interfaces is still a challenging task for today’s research. It presents an indispensable element of any physics-based model of electrochemical systems, which is a necessary tool for improving battery design and performance.

2.4 Diffusion and migration in crystals As already indicated above, electrochemical energy storage is based on the movement and valence state change of ionic species in matter. Therefore, the most important atomic processes involved in these reactions comprise the diffusion or, in terms of an electric field driven process, the migration of charged ions through a liquid or solid matrix. Both processes strongly depend on the type of electrolyte and electrochemical redox reaction. The electrolyte is especially important for the kinetics providing ionic and preventing electronic migration, as it may and in case of solid electrolytes often does present the limiting transport rates and thus the electric power capabilities of the device. The redox reaction describes the participating redox couples and available electrons per reaction and determines, apart from overpotential losses, the potential differences or respective voltage of the cell and as a consequence the highest achievable theoretical energy density. Both factors are of tremendous importance for electrochemical energy storage since they eventually define the field of application and the impact of a certain chemistry within the respective field. Diffusion is a very general concept that is observed on very different scales and even different fields of science. Generally speaking, diffusion describes the movement of a “particle” due to a concentration gradient. In chemistry and physics, this phenomenon can be observed for example when a drop of coloured solute is added to a colourless solvent (see Figure 2.4(a)). If the mixture is not stirred, the solute will slowly start to colour the solvent, starting from the highly concentrated point where the drop was added. This process is driven by the different concentrations of solute in the solvent: solute particles move towards the regions of lowest solute concentration in order to equalise the concentration in the whole solvent. Similar concentration-driven redistribution of particles takes place in solids, as is exemplified in Figure 2.4(b) for atomic interdiffusion at metal/metal interfaces. Mathematically, based on continuity equation for a conserved number of particles, the particle diffusion flux jD [m−2 s−1] causes the change of particle density n [m−3] with progressing time t [s] and is described by Fick’s first and second law [32] and the general local relations ∇jD = ∇ð – D  ∇nÞ = – ∂n=∂t = – DΔn, where D [m2 s−1] is the so-called diffusion coefficient that describes the magnitude of the process – the proportionality constant with respect to the density gradient. The

2.4 Diffusion and migration in crystals

27

(a)

time initial

(b)

100%

Cu

intermediate

Ni

100%

Cu

Ni 0%

50%

50%

50%

0%

0%

Concentration Profile

100%

Concentration Profile

Figure 2.4: Diffusion of dye in a liquid (upper part). After an initially convection-dominated redistribution, the dye moves from highest to lowest concentration until it is evenly distributed throughout the liquid. [Photo by S. Jachalke]. Diffusion of atoms in a solid (lower part) illustrated at the interface of two phases with similar atomic structure but made up of different types of atoms Cu and Ni with a certain degree of solubility for the other type in each phase. With increasing time, the concentration step will smear out as atoms switch positions and move into the other phase.

higher the density gradient between two reservoirs, the faster will be the transport process.1 This Laplacian relation reflects the local particle redistribution towards an equilibration distribution, i. e. a positive local curvature of the particle density will promote a local inward flux, whereas a negative curvature will cause an outward flux. In the picture of charge redistribution, it is convenient to include the elementary charge e [C] into the basic physical quantities, i. e. multiplying the above equation with particle charge q [C]. This way jD becomes the diffusion current density JD representing the flux of charge diffusion in [C m−2 s−1] or [A m−2] with ρ = qn in [C m−3] denoting the charge density. Apart from a density gradient, charged atoms or ions can also move through arrangements of ions due to the application of an electric field. Similar to electrons, ions move along the electric field or respective electric potential ±∇’. When the

1 The model of heat flow uses the same kind of equation showing the generality of the concept.

28

2 Fundamental principles of battery design

gradient is inverted, the ions’ movement will be inverted as well. This process is called ionic “conduction” or “migration” and thus also allows the directed movement of ions in absence of a density gradient and even against it, if the electromotive force2 is high enough. Similarly to Fick’s equation, conduction is described as JC = jρjμE = σE = – σ  ∇’, where JC [A m−2] is the current density induced by the electric field E [V m−1], μ [m2 V−1 s−1] is the mobility of the charge carrier, σ [S m−1 = Ω−1 m−1] the conductivity, and φ [V] the electric potential. Since μ is positive by definition, the absolute value of charge density ρ has to be used to ensure a unidirectionality of current density JC and electric field E. The ionic transport process is a superposition of diffusion and migration and depends on two potentials: the chemical and the electric potential, in sum referred to as the electrochemical potential. The stronger will dominate the movement of ions resulting in a total current density J = – D∇ρ – jρjμ∇’. Conductivity is closely related to the diffusion coefficient and describes the magnitude of achievable movement of ions due to a varying local electric potential ’ ðrÞ. In a classical Boltzmann picture, the equilibrium distribution according to the particle’s electrostatic energy q’ follows for a given temperature T [K], the exponential relation n ⁓ e – q’ =kT . The gradient of the particle density then yields ∇n = – qn=kT  ∇’. From this relation originates, at zero net flux JD + JC = 0 and absence of external electric fields, the Nernst-Einstein equation D = μkT=jqj which connects material parameters of diffusion and migration. The above-given basics are valid for liquid and solid systems. The following discussion focusses on solid electrolytes and all-solid-state type batteries, in particular based on crystalline materials, since these systems have a prospering future (cf. chapter 4.6).

2.5 Crystallographic, crystal chemical, and crystal physical peculiarities This article presents the electrochemistry of batteries from a solid-state physics point of view. The following section covers ideas and considerations that originate from the

2 Electromotive force is the voltage supplied to an electrical circuit by a source of electrical energy, such as a capacitor, a battery, or an electric generator.

2.5 Crystallographic, crystal chemical, and crystal physical peculiarities

29

ordered structure of crystals specifically reflecting the interdisciplinary field of crystallography3 as a tool to give a non-standard but appropriate approach for describing this field of research and application. Ions in crystals are coordinated by other oppositely charged ions. In order to move from one site to another, this coordination needs to be weakened, which requires energy. Since diffusion or migration is described as an atomic jump from one equilibrium structural site via an intermediate region, possibly an intermediate site, to an equivalent equilibrium structural site (see Figure 2.5), crystallography is a powerful tool for the analysis of conduction processes. These capabilities are predominantly based on the high potential for the elucidation of equilibrium structures and electron distributions, which coupled with theoretical techniques provide useful predictions of ionic conductivity. The derived parameters cover primarily the potential barrier EB [J or eV] and, taking into account the particle density as well, the diffusion coefficient. On an atomic level, the moving ion needs to overcome this energetic barrier as it passes through the atomic environment of the minimum-energy pathway for conduction. This is a statistical, thermally activated process based on the dynamic movement of the migrating ions with a certain attempt frequency fi [s−1] to overcome the migration barrier. This barrier is biased by the gradient of the electric

site 1

site 2

site 1

Pair Potential Energy W (energy units)

3 Lennard-Jones Potential Morse Potential 2

1 EB

0

-1 0,0

1,5 1,0 Pair Distance r (spacial units)

2,0 periodicity a

Figure 2.5: (a) Typical pair-interaction potentials of two atoms according to the Lennard–Jones and Morse potential type with a stable minimum position (here normalised to r = 1), a well depth (here normalised to W = –1) giving the dissociation energy and a certain well width defining strength and range of the bond. (b) Thermally activated hopping process illustrated for an atom on a stable site 1 via an intermediate site 2 to an equivalent stable site 1 in form of a double well potential with lattice periodicity a and migration barrier EB.

3 Crystallography describes the atomic arrangement of crystalline matter and offers general tools for the description of non-periodic structures as well.

30

2 Fundamental principles of battery design

potential reducing or increasing the barrier by the electrostatic energy ± qΔ’i =2 for forward and backward current, with an electric potential difference Δ’i per jump along the path i and a rate of success given by a classical Boltzmann-factor (for details see e. g. Ref. [33]). The accompanied drift velocity vD = μE [ms−1] in field direction for a specific migration path i with effectively moved distance si [m]      EB – qΔ’i =2 EB + qΔ’i =2 – exp – vD, i = fi  si  Ni exp – kT kT depends further on the degeneracy4 of the migration path Ni within the crystal structure of the ion conducting material. In crystalline materials, the drift velocity will depend on the crystallographic direction, with respect to path degeneracy, moved distance per successful jump, as well as barrier bias. Transport will preferentially take place in the vacancy or void network in between the discrete atomic positions, which are fixed by the crystal structure. Usually, cations present the most mobile species in solid-state batteries, although anions may be used as well [34]. The interatomic energy of an atom in the vicinity to another atom can be described by pair potentials, e. g. the Lennard–Jones potential, balancing out repulsive (due to the Pauli principle in case of overlapping orbitals) and attractive forces (e.g. van der Waals). Typically, three parameters are characteristic: the dissociation energy, an equilibrium bond distance, and the potential’s well depth describing the range of the interaction (see Figure 2.5). The so-called Morse potential extends this description, taking into account anharmonicity of vibrational modes, the temperature dependence of the equilibrium distance, as well as the description of unbound states and bond breaking (see e. g. Ref. [35]). With respect to the equilibrium distance with another atom, pair potentials describe the increase in energy if the atom is displaced. Since the terms of attraction and repulsion between the coordinating atoms in summary depend on the geometrical arrangement and the direction into which the atom is displaced, there is always an energetically preferred path where migration proceeds. The stronger the bonds between a moving cation and the surrounding anions, the more energy is required for the movement. In particular, ionic compounds may show the disadvantage of stronger, usually more isotropic interactions. Less electronegative elements – such as S instead of O, Cl instead of F – show better flexibility to electron density delocalisation and polarisability. The liquid-like density smooths the energy landscape of the migration path, in particular the discrete singularities of the ionic charge. These elements may form more directed, covalent bonds with other structureconstituting matrix-cations, which offer additional relaxational degrees of freedom

4 In the context presented here, the term degeneracy describes the number of symmetry-equivalent paths within the unit cell of the structure. The symmetry may be reduced by the specific direction of the electric field.

2.5 Crystallographic, crystal chemical, and crystal physical peculiarities

31

during the hopping process. Preferentially, they are also less electron-affine so that they may compensate the electronegativity of the anions without altering the charge state of the moving cation as it approaches the bottleneck position. In general, decreasing changes in the interaction of the moving cation with the structural environment will flatten the energy landscape and decrease the energy barrier. These interactions are even stronger for higher valent ions, like Mg2 + , Ca2 + , Zn2 + , or Al3 +. Therefore, the local chemistry is extremely important for conductors of these cations. A chemical concept regarding the possible bonding and bond strength of atoms and ions is called “HSAB” – theory of “hard” (strong) and “soft” acids and bases [36] – stating that strong acids form strong bonds with strong bases and weak bonds with soft bases. Analogously, soft acids and soft bases form strong bonds. The conduction channel in general provides an anionic environment, which is held together by matrix-cations. In order to have more independent conduction ions, the environment should mainly consist of anions with low affinity for a chemical bond to the moving ion and high affinity for bonding to the matrix-cations. Most of the (technologically) interesting cations for ionic conduction are “hard” acids. Therefore, matrices built from soft acids and bases are of particular interest [37]. In the case of Li, the S-containing ionic conductor materials identified to this point already show higher promise than their O-containing counterparts. Ionic compounds mostly show isotropic, non-directional bonds, as each ion tries to surround itself by oppositely charged ions to passivate or shield central charge and minimise the energy of electric fields. This is best achieved in highly symmetric atomic environments and can be observed in many oxidic compounds, where metal-cations are in many cases tetrahedrally or octahedrally coordinated. These highly symmetric building blocks can then be constructed into crystal structures that may show very differing magnitudes of symmetry: from highly symmetric cubic structures like spinels to completely amorphous silicates. Both of these extremes may show cations tetrahedrally coordinated by oxide ions [38]. “Classic crystallography” deals with 3D periodic structures that are described by a unit cell containing symmetry elements that leave the structure invariant. The more of these elements can be identified, the higher is the intrinsic symmetry of the system. In addition to the translation due to 3D periodicity, a high-symmetry structure presents multitudes of symmetry-equivalent regions and respective structure motives in a unit cell. For a structure of certain symmetry, from a crystallographic point of view, the lower the symmetry of a site,5 the more symmetrically equivalent atoms are

5 The International Union for Crystallography defines a symmetry operation as “a transformation under which two objects, or two configurations of an object, are brought to coincide” [62]. The crystallographic description of a structure is based on the placing of different symmetry elements – geometric elements, mostly points, lines, and planes, that apply a symmetry operation – in a 3D space. Each specific set defines a unique space group. If an atom is placed on a specific position in this

32

2 Fundamental principles of battery design

generated and the higher is this site’s multiplicity. The set of symmetry-produced atoms of a specific atomic position is called crystallographic orbit, for which all atoms have an identical atomic environment and thus the same energy. As described in Ref [39], the most important chemical design principle for superionic conductors6 is the presence of a plurality of partially occupied, energetically identical, or at least similar sites in the crystal structure. This principle is illustrated schematically in Figure 2.6. Multitudes of energetically identical sites can be identified in crystal structures by low-symmetry, high-multiplicity sites. In the best case, only one site of very high multiplicity is involved in the conduction process. The occupancy of this site is also an important factor in crystallography that is determined during the experimental crystal structure analysis. Of course, these sites have to be interconnected in order to allow the movement of an ion through the structure. The connectivity is defined by a percolation of conduction and intermediate sites. This means that at least one “infinite” path can be identified in the crystal structure, starting on a point of one face of a unit cell and arriving to this point, but on the exact opposite face of the unit cell or supercell. If this is not the case, a 0D path is formed (a loop). Infinite paths can be of 1D (tubes), 2D (planes), and 3D (networks) dimensionality (see Figure 2.7). Crystal structures can be assigned to a crystal class because of their inherent symmetry. Because of Neumann’s principle, which states that “if a crystal is invariant with respect to certain symmetry elements, any of its physical properties must also be invariant with respect to the same symmetry elements” [40], certain dimensionalities can be expected: In cubic, only 3D paths are possible. If a conduction path is parallel to the designated axis (c axis) in hexagonal, trigonal, and tetragonal structure, it is at least 1D, if the path is perpendicular to the designated axis, it must be 2D. For all less symmetric crystal classes, no preferences can be deduced from the available symmetry operations. Ideal candidates for good conduction are therefore highly symmetric structures. In the case of cationic transport, more accurately, a high-symmetry anion-sublattice should be aimed at [41]. High packing density anion networks exhibit multitudes of

3D space filled with symmetry elements, all symmetry operations are applied to this position. If the position does not coincide with any symmetry element, it is called “general position”, if it coincides with one or more symmetry elements it is called a “special position.” The more symmetry elements leave an atom invariant, the less “symmetrically equivalent” atoms are generated and the higher the position’s local symmetry. The sets of symmetry elements of all positions reflect the possible “site symmetries” in a space group and define the “crystallographic sites.” The amount of symmetrically equivalent atoms generated on a certain site is called “multiplicity.” In case there are free positional parameters on a crystallographic site, as is always true for the coordinates of the general position, a specific coordinate triplet defines a “crystallographic orbit.” 6 Ionic conductors with very high conductivity are commonly referred to as superionic conductors.

2.5 Crystallographic, crystal chemical, and crystal physical peculiarities

33

Figure 2.6: Graphic representation of the three main “crystallographic” features of a superionic conductor; in blue: matrix anions (matrix cations are omitted to avoid confusion), in red, dotted: void sites, in red, solid: conducted cations. A void network with void sites is needed for fast conduction. For a flat energy surface, energetically equivalent sites (left) are preferential, since larger or more displaced anions decrease the space and thus change the energy of the neighbouring voids (middle). For low energy sites within the conduction network – the dynamical traps, population of other void sites can be neglected in time-average – an occupancy of about one half is best, since each major hopping process can only be successful with an adjacent void. If the concentration is too low, the current is limited due to the migrating species. If it is too high, the current is limited due to the limited amount of necessary voids.

(a)

(b)

(c)

(d)

Figure 2.7: Basic matrix dimensionalities of an ion-conducting matrix (grey polyhedra) with highlighted conduction network (blue network): (a) 0D within a 3D conduction network, (b) 1D within a 1D conduction network, (c) 2D within a 2D conduction network, and (d) 3D within a 3D conduction network.

energetically identical or similar sites that are interconnected by likewise identical or similar coordination-polyhedra faces through which the movement between sites is mediated. In these cases, the channel between these sites consists of a triangle of anions that is penetrated by the moving cation. In a recent study, sulphur-containing Li+-conductors were systematised according to their anion-package by comparing the

34

2 Fundamental principles of battery design

anion-network to ideal packings of bcc, fcc, and hcp arrangements [37]. The highest promise was deduced from bcc arrangements of anions: the conduction sites are situated in tetrahedral sites which are directly connected to each other via a single jump from tetrahedron to tetrahedron. In fcc and hcp structures tetrahedral sites have no direct connection, they do not percolate, but an intermediate octahedral site needs to be passed. Additionally, there are no direct octahedron–octahedron connections. Therefore, two possibly energetically different sites increase the “complexity” of the migration network. In fact, the highest Li+-conductivity sulphur-containing compounds show anion-arrangements that can be mapped to bcc. In summary, the conduction network should show as little energetic deviation in the conduction path as possible. This is achieved by having percolating sites with more or less the same energy. Recently, it was also shown that this implication stays true even if only energetically unfavourable sites are occupied as long as this path is not connected to lower energy sites [42]. Crystallographic science along with X-ray diffraction, powder X-ray diffraction (PXRD), resonant X-ray diffraction (RXD), and neutron diffraction measurements provide excellent characterisations of structural features that are useful to infer ionic transport (details in chapter 5). PXRD in particular [43] is of relevance to identify porosity regions that could eventually become ionic conduction channels in solid state electrolytes or in solid conductors. In addition, RXD techniques provide spectroscopic sensitivities for the investigation of specific chemical species in the crystal structure, their valence states, crystallographic sites, as well as their respective structural phase [44, 45]. Once the crystal structure is characterised, first principles computational methods (discussed in chapter 4.1). are extremely useful to identify minimum potential energy pathways for ions traveling the solid structures [46, 47] and diffusion mechanisms that can compete in a given solid or through solid/solid interfaces [48].

2.6 Classification of battery applications and types Once adequate materials have been identified, the electrochemical cell, which is the smallest unit of a battery, may be designed. The general sequence of material layers is set by the respective electrochemistry, but the final form of the housing holding the cell may differ according to voltage, capacity, or space requirements of the application [49]. Small batteries, e. g. for consumer applications like flashlights, remote controls, etc., just consist of one electrochemical cell. Other applications like Laptop batteries require the connection of several cell units in parallel, in series, or both. Standards in cell packaging, e. g. of typical Li-ion batteries, comprise button, cylindrical, prismatic, and pouch cell geometries [50]. For the cylindrical battery, the material stack is cut to a specified width and rolled to the desired radius, whereas for the prismatic geometry the

2.6 Classification of battery applications and types

+ve/-ve Terminals and Safety Vent

35

+ve/-ve Terminals

Metal Case

Pressure Relief Vent

Anode

Separator

Cathode Metal Case Anode

Separator Cathode

+Ve/-ve Terminals

Metallised Foil Pouch Anode Separator

Cathode

Figure 2.8: Standard geometries of conventional Li-ion batteries: cylindrical (upper left), prismatic (upper right), and pouch cell geometries (lower). Copyright Johnson Matthey Plc © 2012 [51] – reprinted with permissions.

wrapping is flat, and for the pouch cell setup the flat material stacks are packed on top of each other to the desired height or are designed as prismatic flat wrap (see Figure 2.8). Prismatic cells can be packaged more efficiently than cylindrical cells because of their form factor and therefore the packing density is higher [51]. They are available in sizes up to 100 Ah, whereas sizes of up to 200 Ah are available for cylindrical cells. While cylindrical and prismatic cells are predominantly produced in hard cases, so-called pouch cells are mostly soft packs. They make the most efficient use of available space and achieve a packaging efficiency of 90–95 %. Because of the absence of a metal can, the pouch pack has a lower weight and therefore the battery pack will have a higher

36

2 Fundamental principles of battery design

energy density. However, the cells have low mechanical stability and therefore a more robust packaging is required for many applications. Typically, several of these cells are connected to modules, which then contain also individual cell monitoring and temperature control. The assembly of individual modules eventually forms the final battery pack. All mentioned geometries are typically sealed with a metallic casing due to the impermeability for liquids, moisture, and air as well as for mechanical stability. In the case of commercially available Li-ion battery packs for automotive applications, about 500 single parts have to be assembled in one battery system, together with about 300 screw connections and 250 welding points [49]. The most widespread Li-ion battery on the market today is the 18,650 round cell. It has to be noted that for the emerging battery technologies, like the high-temperature Na–S battery, the cell design is different owing to the fact that the electrodes are liquid and therefore separated by a solid-state material. Hence, tube-shaped ceramics are used as containers here [52]. In today’s life, batteries power numerous applications with diverse needs and demands for electrical energy in terms of parameters (voltage, energy density, current, power density, etc.) and handling (mobility, chargeability, cycle life, etc.). With respect to the powered device, one basic classification of a battery is therefore made according to its portability. Mobile applications comprise a variety of everyday consumer tools like mobile phones and smartphones, laptops, or watches up to electrically powered cars or busses, to name a few. Stationary applications may be used to balance generated and consumed electricity on all scales, from single houses to interregional or even national and international electrical grids, or to ensure power supply in case of grid failure, e. g. by means of uninterruptible power supplies. Accessibility as well as power demand often determine if a battery may be changed or recharged regularly or not, as is the case e. g. for medical applications like implantable neurostimulators or cardiac pacemakers, or batteries for military and space applications. Batteries are further classified regarding their chargeability. Primary batteries are based on chemical reactions that cannot be easily reversed, which usually allows for higher energy densities. These batteries are disposed after use. Secondary batteries or accumulators can be recharged. They convert externally provided electrical energy back to chemically stored energy in the battery, reversing the internal chemical reactions. Thus, the battery can be cycled many times before degradation processes start to limit the capacity. Tertiary cells are sometimes referred to and describe cells with quasi-unlimited energy and power. They basically describe two electrodes that are independent of the active masses consumed during the electrochemical reactions. The more active mass is available, the more energy can be generated. Power can be increased by increasing the surface of the electrodes and thus the amount of charge generated per time. A fuel cell is the prime example for a tertiary cell, depending only on the amount of hydrogen and oxygen to be converted to water.

References

37

Battery types comprise various concepts to realise conversion of chemically stored energy to electrical energy. Among the most common, prevalent types are based on cation migration. Representatives cover, e. g., Li batteries [53], Ni–Cd batteries [54], polymer batteries [55], redox flow batteries [56–58], metal–air batteries [59], Pb accumulators [60], and concentration cells [61]. Several of these types and concepts are presented in detail in chapter 3.

References [1] [2] [3]

[4] [5] [6] [7] [8] [9] [10]

[11] [12] [13] [14]

[15] [16] [17] [18]

Simon P, Gogotsi Y. Materials for electrochemical capacitors. Nat Mater. 2008;7:845–54. Linpo Y, Chen GZ. Redox electrode materials for supercapatteries. J Power Sources. 2016;326:604–12. Meyer DC, Leisegang T, Stöcker H, Zschornak M. Electrochemical storage materials: from crystallography to manufacturing technology. Berlin, Germany: De Gruyter Oldenbourg Publishing House, 2018. Nernst W. Die elektrolytische Zersetzung wässriger Lösungen. Eur J Inorg Chem. 1897;30: 1547–63. Atkins PW, de Paula J. Physikalische Chemie, 4th ed. Weinheim, Germany: Wiley-VCH Verlag, 2006. ISBN: 978-3-527-31807-0. Mortimer CE, Müller U. Chemie: Das Basiswissen der Chemie, 10th ed. Stuttgart, Germany: Georg Thieme Verlag, 2007. ISBN: 978-3134843095. Hollemann AF, Wiberg N. Lehrbuch der Anorganischen Chemie, 102nd ed. Berlin, Germany: Walter de Gruyter Verlag, 2007. ISBN: 978-3110177701. Schmidt VM. Elektrochemische Verfahrenstechnik – Grundlagen, Reaktionskinetik, Prozessoptimierung. Weinheim, Germany: Wiley-VCH Verlag, 2003. ISBN: 978-3-527-29958-4. Riedel E, Janiak C. Anorganische Chemie, 7th ed. Berlin, Germany: Walter de Gruyter Verlag, 2007. ISBN: 978-3110189032. Gupta SV. Units of measurement: past, present and future. international system of units. In: Hull R, et al., editor(s). Springer series in materials science, Vol. 122. Heidelberg, Germany: Springer Science & Business Media, 2009. ISBN: 9783642007378. Kuchling H. Taschenbuch der Physik, 11th ed. Thun and Frankfurt/Main, Germany: Verlag Harri Deutsch, 1988:635. ISBN: 3-8171-1020-0. Bergmann L, Schaefer C, Kassing R. Lehrbuch der Experimentalphysik. Band 6: Festkörper, 2nd ed. Berlin, Germany: Walter de Gruyter, 2005:361. ISBN: 3-11-017485-5. Nernst W. Theoretical Chemistry from the Standpoint of Avogardro’s Rule & Thermodynamics, 4th ed. London, UK, New York, USA: The MacMillan Company, 1904. von Helmholtz H. Über einige Gesetze der Vertheilung elektrischer Ströme in körperlichen Leitern, mit Anwendung auf die thierisch-elektrischen Versuche. Ann Phys Chem. 1853;89: 353–77. Lippmann G. Beziehungen zwischen den Capillaren und elektrischen Erscheinungen. Ann Phys. 1873;225:546–61. Gouy M. Sur la constitution de la charge électrique à la surface d'un électrolyte. J Phys Theor Appl. 1910;9:457–68. Chapman DL. LI. A contribution to the theory of electrocapillarity. Lond Edinb Dubl Phil Mag J Sci. 1913;25:475–81. Stern O. The theory of the electrolytic double-layer. Z Elektrochem. 1924;30:1014–20.

38

2 Fundamental principles of battery design

[19] Bikerman J. Structure and capacity of electrical double layer. Lond Edinb Dubl Phil Mag J Sci. 1942;33:384–97. [20] Freise V. Zur Theorie der diffusen Doppelschicht. Z Elektrochem Ber Bunsenges physik Chem. 1952;56:822–7. [21] Grahame DC. The electrical double layer and the theory of electrocapillarity. Chem Rev. 1947;41:441–501. [22] Erdey-Gruz T, Volmer M. Zur Theorie der Wasserstoff Überspannung. Z Phys Chem. 1930;150:203–13. [23] Erdey-Gruz T, Volmer M. Zur Frage der elektrolytischen Metallüberspannung. Z Phys Chem. 1931;157:165–81. [24] Butler JAV. The mechanism of overvoltage and its relation to the combination of hydrogen atoms at metal electrodes. Trans Faraday Soc. 1932;28:379–82. [25] Doyle M, Fuller TF, Newman J. Modeling of galvanostatic charge and discharge of the lithium/ polymer/insertion cell. J Electrochem Soc. 1993;140:1526–33. [26] Frumkin A. Wasserstoffüberspannung und Struktur der Doppelschicht. Z Phys Chem. 1933;164:121–33. [27] Van Soestbergen M. Frumkin-Butler-Volmer theory and mass transfer in electrochemical cells. Russ J Electrochem. 2012;48:570–9. [28] Latz A, Zausch J. Thermodynamic derivation of a Butler–Volmer model for intercalation in Li-ion batteries. Electrochim Acta. 2013;110:358–62. [29] Rubi J, Kjelstrup S. Mesoscopic Nonequilibrium Thermodynamics Gives the Same Thermodynamic Basis to Butler−Volmer and Nernst Equations. J Phys Chem B. 2003;107:13471–7. [30] Zeng Y, Smith RB, Bai P, Bazant MZ. Simple formula for Marcus–Hush–Chidsey kinetics. J Electroanalytical Chem. 2014;735:77–83. [31] Lück J, Latz A. Theory of reactions at electrified interfaces. Phys Chem Chem Phys. 2016;18:17799–804. [32] Fick A. Über Diffusion. Annalen der Physik 170.1 (1855): 59-86. Fick A. On liquid diffusion. Phil Mag Series. 1855;4:30–9. [33] Hanzig J, Zschornak M, Mehner E, Hanzig F, Münchgesang W, Leisegang T, et al. The anisotropy of oxygen vacancy migration in SrTiO3. J Phy Cond Matter. 2016;28:225001. [34] Hanzig J, Zschornak M, Nentwich M, Hanzig F, Gemming S, Leisegang T, et al. Strontium titanate: an all-in-one rechargeable energy storage material. J Power Sources. 2014;267:700–5. [35] Adams S, Prasado Rao R. High power lithium ion battery materials by computational design. Phys Status Solidi. 2011;208:1746–53. [36] Pearson RG. Hard and soft acids and bases, HSAB, part 1: fundamental principles. J Chem Educ. 1968;45:581–7. [37] Wang Y, Richards WD, Ong SP, Miara LJ, Kim JC, Mo Y, et al. Design principles for solid-state lithium superionic conductors. Nat Mater. 2015;14:1026–31. [38] Meutzner F, Münchgesang W, Kabanova NA, Zschornak M, Leisegang T, Blatov VA, et al. On the way to new possible na-ion conductors: the voronoi–dirichlet approach, data mining and symmetry considerations in ternary Na oxides. Chem Eur J. 2015;21:16601–8. [39] Huggins RA. Chapter 9: Very Rapid Transport in Solids. In: Nowick AS, editor(s). Diffusion in solids: recent developments. New York, USA: Academic Press, 1975:445–86. ISBN: 0-12522660-8. [40] Neumann F. In: Meyer OE, editor(s). Vorlesungen über die Theorie der Elastizität der festen Körper und des Lichtäthers. Leipzig, Germany: B. G. Teubner-Verlag, 1885. [41] van der Veen A, Bhattacharya J, Belak AA. Understanding Li Diffusion in Li-Intercalation Compounds. Acc Chem Res. 2013;46:1216–25.

References

39

[42] Rong Z, Malik R, Canepa P, Sai Gautam G, Liu M, Jain A, et al. Materials design rules for multivalent ion mobility in intercalation structures. Chem Mater. 2015;27:6016–21. [43] Yakovenko AA, Wei ZW, Wriedt M, Li JR, Halder GJ, Zhou HC. Study of guest molecules in metal organic frameworks by powder X-ray diffraction: analysis of difference envelope density. Cryst Growth Des. 2014;14:5397–407. [44] Zschornak M, Richter C, Nentwich M, Stöcker H, Gemming S, Meyer DC. Probing a crystal’s short‐range structure and local orbitals by Resonant X‐ray Diffraction methods. Crystal Res Technol. 2014;49:43–54 [45] Richter C, Zschornak M, Novikov D, Mehner E, Nentwich M, Hanzig J, Gorfman S, Meyer DC. Picometer polar atomic displacements in strontium titanate determined by resonant X-ray diffraction. Nat Comms. 2018;9:178. [46] Wengert S, Nesper R, Andreoni W, Parrinello M. Ionic diffusion in a ternary superconductor: an ab initio molecular dynamics study. Phys Rev Lett. 1996;77:5083–5. [47] Shi SQ, Lu P, Liu ZY, Qi Y, Hector LG, Li H, et al. Direct calculation of li-ion transport in the solid electrolyte interphase. J Am Chem Soc. 2012;134:15476–87. [48] Soto FA, Yan P, Engelhard MH, Marzouk A, Wang C, Xu G, et al. Tuning the solid electrolyte interphase for selective li- and na-ion storage in hard carbon. Adv Mater. 2017;29:1606860. [49] Maiser E. Battery packaging – technology review. AIP Conf Proc. 2014;1597:204–8. [50] Korthauer R. Handbuch Lithium-Ionen-Batterien. Berlin/Heidelberg, Germany: Springer-Verlag, 2013. ISBN: 978-3-642-30652-5. [51] Johnson Matthey Battery Systems (former Axeon © 2012). Our Guide to Batteries. Rooksley, Milton Keynes, UK: Johnson Matthey, Precedent House. 2nd edition, 2018. accessed on Jan 26th. [52] Dunn B, Kamath H, Tarascon J-M. Electrical energy storage for the grid: a battery of choices. Science. 2011;334:928–35. [53] Kraytsberg A, Ein-Eli Y. Higher, stronger, better? a review of 5 volt cathode materials for advanced lithium-ion batteries. Adv Ene Mat. 2012;2:922–39. [54] Reddy TD, Linden D. Chapter 19 – 21: Nickel-Cadmium Batteries. Linden’s handbook of batteries, 4th ed. New York City, USA: McGrawHill Verlag, 2011. ISBN: 978-0071624213. [55] Page KA, Soles CL, Runt J. Polymers for energy storage and delivery: polyelectrolytes for batteries and fuel cells, Vol. 1096. Washington D.C., USA: American Chemical Society, 2012. ISBN: 9780841226319. [56] Skyllas-Kazacos M, Chakrabarti MH, Hajimolana SA, Mjalli FS, Saleem M. Progress in flow battery research and development. J Electrochem Soc. 2011;158:R55–R79. [57] Wang W, Luo Q, Li B, Wei X, Li L, Yang Z. Recent progress in redox flow battery research and development. Adv Funct Mater. 2013;23:970–86. [58] Ponce De León C, Frías-Ferrer A, González-García J, Szánto DA, Walsh FC. Redox flow cells for energy conversion. J Power Sources. 2006;160:716–32. [59] Cheng F, Chen J. Metal–air batteries: from oxygen reduction electrochemistry to cathode catalysts. Chem Soc Rev. 2012;41:2172–92. [60] Reddy TD, Linden D. Chapter 16 & 17: Lead-Acid Batteries & Valve Regulated Lead-Acid Batteries. Linden’s Handbook of Batteries, 4th ed. New York City, Vereinigte Staaten: McGrawHill Verlag, 2011. ISBN: 978-0071624213. [61] Beattie GW. Nernst’s theory of the concentration cell. Charleston SC, USA: BiblioBazaar, 2015. ISBN: 9781343047952. [62] Hahn T (Hrsg.). International tables for crystallography. Bd. A: Space-group symmetry. 5., rev. ed., repr. with corr. Dordrecht: Kluwer Academic Publishers, 2002. ISBN: 0-7923-6591-7.

Melanie Nentwich, Bianca Störr and Juliane Hanzig

3 Battery concepts: The past, the present, and research highlights Abstract: The concept of a battery is not a modern invention, as first proofs go back to 200 BC. The development of electrochemical cells similar to those that we use today started at the end of the eighteenth century with the experiments of Luigi Galvani. The following paragraphs will give an overview of the progress in electrochemistry from the very early reports to the state of the art. Additionally, some future perspectives from the recent years will be highlighted. Keywords: history, perspectives, overview

3.1 The past The earliest proof for electrochemical devices is the Battery of Bagdad, which was found at Khujut Rabu, close to Bagdad. It was most likely created between 227 and 126 BC during the Parther Empire. The device consists of a copper cylinder with an iron rod in its center. Both are connected via isolating asphalt and shielded to the outside by a clay jar [1]. The discoverer of this device assumed that several such cells had been used in series connection. Three applications are possible: (i) the generation of an electric current, (ii) gold plating of metal objects, and (iii) the treatment of diseases with electric shocks [2]. The voltage gained from this device is certainly less than expected from theory due to the use of juice as electrolyte [1]. With nowadays knowledge, this ancient setup can be described easily by chemical reaction equations. We will use the term battery exclusively for primary cells and accumulator exclusively for secondary cells. Additionally, we will indicate the reverinstead of in chemical formulas. The discharging process is sibility by using shown in reading direction. Furthermore, the order of the reaction is always negative electrode (anode during discharge), positive electrode (cathode during discharge), (complex forming), and overall reaction.

2+

Cu

Fe + 2e–

CuO + Fe

Fe2+ + 2e– Cu

0.44 V 0.34 V

Cu + FeO

0.78 V

This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Nentwich, M., Störr, B., Hanzig, J. Battery concepts: The past, the present, and research highlights. Physical Sciences Reviews [Online], DOI: 10.1515/psr-2018-0038 https://doi.org/10.1515/9783110493986-003

42

3 Battery concepts: The past, the present, and research highlights

The term electrical battery was first used by Benjamin Franklin (1706–1790) in 1749 to describe an arrangement of several capacitors [3]. This term is adapted from the military to describe a set of similar objects operating together. Later, the term was adapted for electrochemical cells, as the principle of gaining electrical energy from this group setup is similar. In 1780, the medical scientist Luigi Galvani (1737–1798) observed muscle contractions of frog legs and assumed animalistic electricity to be the cause. He described the frog legs as some kind of Leyden jar, a capacitor consisting of a glass jar covered with metal foils from the inside and outside [4]. However, he unconsciously formed a circuit between the legs, a piece of copper, and a piece of iron, thus gaining electricity due to the potential difference of the two metals [5]. In his honor, the Galvanic cells carry his name. Twelve years later, Alessandro Volta (1745–1827), who had already invented some electrical devices, read Galvani’s reports. He discovered that the frog’s muscle contractions are caused by an external voltage due to contact electricity [6]. The different conclusions from Galvani and Volta caused a massive argumentation among all European scientists. Volta started a long-lasting research that leads to a first electropotential series and the construction of the first battery of the modern age. The Voltaic pile is a stack of silver and zinc plates separated by cloth or cardboard soaked in acid, thus forming a serial connection of Ag-Zn cells [7]. Alternatively, copper and tin were used. The underlying nonreversible reactions are:

Zn 2Ag + 2e–

Zn2+ + 2e– 2Ag

0.76 V 0.80 V

2Ag+ + Zn

2Ag + Zn2+

1.56 V

+

Sn + 2e–

Sn2+ + 2e– Cu

0.14 V 0.34 V

Cu2+ + Sn

Cu + Sn2+

0.48 V

2+

Cu

In the same year, William Nicholson (1753–1815) and Anthony Carlisle (1768–1840) performed the first electrolysis using this Voltaic pile. Using the electrical energy to start a chemical reaction, they created a gas on each electrode, later determined as hydrogen and oxygen [8]. Two years later, in 1802, Johann W. Ritter (1776–1810) constructed the first accumulator, i. e. the first rechargeable electrochemical cell. This cell consists of a pile of copper plates separated by carton moistened with an acid [9]:

3.1 The past

2+

2Cu

2O2– + 4e–

O2 + 4e– 2Cu

2CuO

O2 + 2Cu

43

These piles can create voltages in the range of kV, but with very low currents. They were used to power Zamboni pendulums, which work for 187 years without additional supply of energy [10]. In 1831, Michael Faraday (1791–1867) discovered electromagnetic induction [11]. In his book, he coined the terms of anode and cathode as we use them today. In the late 1830s, the sending of telegrams became more frequent. Thus to transmit the messages over longer distances, the demand for batteries with better performance increased. The combination of copper and zinc plates in the Daniell cell by John F. Daniell (1790–1845) met these requirements [12]: Zn Cu2+ + 2e–

Zn2+ + 2e– Cu

0.76 V 0.34 V

Cu2+ + Zn

Cu + Zn2+

1.10 V

A second successful invention was the Grove cell invented by William R. Grove (1811–1896) in 1839. On the cathode side, the cell exhibits a platinum current collector, which also operates as a catalyst. The active material comes from the electrolytes, e. g. nitric acid on the cathode side and sulphuric acid on the anode side. The electrolytes are kept in separate tanks, connected by a diaphragm [13]. The anode consists of zinc. The Grove cell emits noxious nitrogen dioxide: Zn H2SO4 + 2HNO3 + 2e–

Zn2+ + 2e– SO42- + 2H2O + 2NO2

Zn + H2SO4 + 2HNO3

ZnSO4 + 2H2O + 2NO2

In 1866, Georges Leclanché (1839–1882) patented his Leclanché cell [14], which is the predecessor of the dry-cell battery and which was used for telegraphic systems and in railroads. The Leclanché cell is one of the first batteries that use a cathode of manganese dioxide with a current collector made of graphite. The reaction includes the formation of a complex, which is why we added a third reaction equation: Zn 2MnO2 + 2H+ + 2e– Zn2+ + 2NH4+ + 2Cl– Zn + 2MnO2 + 2NH4Cl

Zn2+ + 2e– 2MnO(OH) [Zn(NH3)2]Cl2 + 2H+ [Zn(NH3)2]Cl2 + 2MnO(OH)

44

3 Battery concepts: The past, the present, and research highlights

In 1841, Robert Wilhelm Bunsen (1811–1899) improved the Grove cell by replacing the expensive platinum by carbon (pressed coal) [15]. Exactly like the Grove cell, the Bunsen cell emitted nitrogen dioxide. In 1854, Wilhelm Josef Sinsteden (1803–1891) conducted experiments using two lead plates placed in sulphuric acid. After several cycles of charging and discharging, one of the plates got oxidized and he could measure a capacity. Five years later, Gaston Planté (1834–1889) enhanced this first lead-acid accumulator by using a spiral arrangement of the lead plates [16], which is still used today: Pb + SO42– PbO2 + SO42– + 4H3O+ + 2e– Pb + PbO2 + 2H2SO4

PbSO4 + 2e– PbSO4 + 6H2O

0.36 V 1.68 V

2PbSO4 + 2H2O

2.04 V

In 1873, Josiah Latimer Clark (1822–1898) presented a cell producing a very stable direct voltage of 1.4 V. The Clark cell was used as a reference voltage source [17]. It consists of a zinc amalgam and a liquid mercury electrode in sulphuric acid with the reactions: Zn Hg2+ + 2e–

Zn2+ + 2e– Hg

0.76 V 0.86 V

Zn + Hg2+

Zn2+ + Hg

1.62 V

In 1897, Waldemar Jungner (1869–1924) invented the nickel–iron accumulator [18]. Two years later, he replaced iron by cadmium [19], which was cheaper but also less efficient and which produced more hydrogen. At the same time, Thomas A. Edison (1847–1931) independently invented the nickel–iron accumulator and patented it in the USA in 1901 [20]. He promoted this accumulator with its higher energy density compared to the popular lead-acid accumulator for the application in electric vehicles; however, the nickel–iron accumulator did not prevail: Fe + 2OH– 2NiO(OH) + 2H2O + 2e– Fe + 2NiO(OH) + 2H2O

Fe(OH)2 + 2e– 2Ni(OH)2 + 2OH– Fe(OH)2 + 2Ni(OH)2

In 1965, a first patent mentioning lithium as anode material was filed [21]. As the reduction of lithium has one of the most negative standard electrode

3.2 The present

45

potentials, it is a desired material for electrochemical applications. In 1966, D. A. Swinkels published a paper about the first accumulator containing lithium [22]. On the cathode side, the current is collected by graphite, but the active material is provided by the LiCl electrolyte. The anode consists of metallic Li. To enable ion diffusion, the electrolyte and the anode are in the liquid state (melting points are Tm ðLiÞ = 180  C, Tm ðLiClÞ = 614  C), so that the operating temperature is 650  C:

1 2Cl2

Li + e–

Li + 21 Cl2

Li+ + e– Cl–

3.05 V 0.68 V

LiCl

3.73 V

It took another 14 years until John B. Goodenough proposed lithium intercalation as a suitable technique for electrochemical cells using lithium cobalt oxide as the cathode [23]: LiC6 +

–

CoO2 + Li + e

LiC6 + CoO2

C6 + Li+ + e– LiCoO2 C6 + LiCoO2

In 1976, the sodium–sulphur accumulator was patented [24]. The cell operates at temperatures above 270  C to enable the reaction of molten sodium and molten sulphur. Both electrodes are separated by a solid electrolyte, e. g. β′′-alumina (β′′ Al-2O3):

n 8S8

2Na + 2e–

2Na + 8n S8

2Na+ + 2e– Sn2– Na2Sn

3.2 The present The following sections briefly introduce the battery types most commonly used in the past 50 years, in alphabetical order. Only the major parts of the cells will be described as well as some additives of great importance. For more detailed information, please refer to References [25–27] and those given in the following. 3.2.1 Flow accumulator The flow accumulator consists of two tanks containing the solvents with the active materials. These electrolytes are called anolyte and catholyte based on the terms

46

3 Battery concepts: The past, the present, and research highlights

anode and cathode. Pumps transfer both liquids to the electrochemical cell, where the reactions take place. A separator membrane prevents the exchange of electrons. The advantage of the flow accumulator is the simple construction compared to other cells. If the electrolyte has to be renewed, only the tank needs to be replaced. Additionally, the capacity is determined by the container size and the concentration of the active material, whereas the power depends on the active surface of the current collector. In 1954, Walther Kangro developed the first concepts of the flow accumulator [28, 29]. Later, the NASA went on developing this approach using iron and chromium electrolytes [30]. Here, the mixing of the electrolytes is a problem as it implies a loss of capacity and a lower cycle stability [31, 32]: Cr2+ Fe3+ + e– Cr2+ + Fe3+

Cr3+ + e– Fe2+

0.41 V 0.77 V

Cr3+ + Fe2+

1.18 V

Finally, in 1986, Maria Skyllas-Kazacos established the vanadium system [33]. The V4+: cathode reaction of this vanadium system can be simplified to V5+ + e–

VO2+

V2+ + 2H + e– +

V2+ + VO2+ + 2H+

V3+ + e– VO2+ + H2O

0.26 V 1.00 V

V3+ + VO2+ + H2O

1.26 V

Other research groups focused on other electrolyte combinations, like the bromine– sulphur cell. The polysulphide anode can partially made of waste from NaS accumulators and is thus highly abundant. Additionally, this cell is highly efficient and has a high cycle stability and low corrosion due to its low pH value. On the other hand, the mixing of the electrolytes can cause explosions, the membrane decomposes easily, and gas generation can lead to an expansion of the cells [32, 34]:

Br3–

2S22– + 2e–

2S22– + Br3–

S42– + 2e– 3Br–

0.27 V 1.09 V

S42– + 3Br–

1.36 V

Another example is the zinc–bromine cell, where zinc deposits on the current collector during the charging process. The combination of zinc and bromine in an electrochemical cell was first patented in 1880 [35]. However, the realization was hindered by the dendrite formation of zinc and the dissolving of bromine in aqueous zinc bromine [25]. These two problems were solved by Gould in the 1970s [36]:

3.2 The present

Zn Br2 + 2e– Zn + Br2

Zn2+ + 2e– 2Br–

0.76 V 1.09 V

ZnBr2

1.85 V

47

Table 3.1 gives an overview of the cell parameters as well as advantages and disadvantages of the various concepts of flow accumulators. Table 3.1: Technical data of different flow accumulators [25, 26, 37–40]. Electrodes Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number (in thousands)

V2+|V5+ 1.26 20 ... 33 15 ... 25 0 5 ... 14

Advantages

Simple thermal management, flexible design, rapid charge, no harm by deep discharge Low specific energy Large-scale energy storage devices, electric vehicles

Disadvantages Applications

Cr2+|Fe3+ 1.18 40 15 ... 20 0 5 ... 14

S22+|Br3– 1.52 80 0 5 ... 14

Zn|Br2 1.85 80 50 ... 100 0 5 ... 14

3.2.2 Lead-acid accumulator After the work of Sinsteden and Planté (see Section 1), the Luxembourgian inventor Henri Tudor used the lead-acid accumulator in 1882 to store energy from a waterfall to supply his houses with electricity [41]. In 1885, he founded a factory for the production of lead-acid accumulators. Almost at the same time, in 1887, Adolf Müller started a similar project in Germany resulting in the factory VARTA in 1904 [42]. Table 3.2 lists the cell parameters as well as advantages and disadvantages of the lead-acid accumulator. 3.2.3 Lithium ion accumulator Lithium ion accumulators are a group of electrochemical cells that are based on the exchange of lithium ions from one electrode to the other. Mostly, lithium is intercalated in or removed from certain host structures. Often, the anode host is graphite and the cathode host is a metal oxide [25], with metal M. Lithium ion accumulators are very versatile in their handling due to their high voltage, high cycle life, high energy density, and high specific energy, see Table 3.3. Thus, a single review cannot cover

48

3 Battery concepts: The past, the present, and research highlights

all aspects of these cells. Dedicated reviews focus on applications as microelectronics [45], in electric vehicles [46], and for the storage of renewable energies [47] as well as the individual parts of the cell, e. g. the cathode [48, 49], the anode [50, 51], and the electrolyte [52, 53], or the recycling [54]: Cn + x Li+ + x e– LiM O2

LixCn Li1-xM O2 + x Li+ + x e– LixCn + Li1-xM O2

LiM O2 + Cn

Table 3.2: Technical data of the lead-acid accumulator [25, 26, 43, 44]. Anode Cathode Electrolyte

Pb PbO2 H2SO4

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

2.00 50 ... 110 30 ... 42 3 ... 20 < 350

Advantages

Reliability, abundance, low cost, scalability, high cell voltage, simple production, good recyclability, high discharge current, no overcharging Relatively low cycle number, high self-discharge, deposition of impurities at the anode, H2 production during charging, limited specific energy Starter batteries, electric vehicles, standby units, stationary storage

Disadvantages

Applications

Table 3.3: Technical data of the lithium ion accumulator [25, 26, 44]. Anode Cathode Electrolyte

C LiCoO2, LiNiO2, or LiMn2O4 Water-free solvents with Li-salt, e. g. LiPF6 or LiBF4

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

3.3 ... 3.8 200 ... 500 90 ... 240 3 > 90000

Advantages Disadvantages

High specific energy, high cell voltage Less robust (reactions of Li with H2O), low currents, control of voltage, current, and temperature Electrical vehicles, microelectronics, ...

Applications

3.2 The present

49

3.2.4 Lithium–iron sulphide battery Lithium–iron sulphide batteries were invented in 1977 as a cheaper alternative to zinc–silver oxide batteries [55]. Due to a decrease in the silver prize, the lithium–iron batteries became less attractive and the production stopped until 1992:

4Li FeS2 + 4e–

4Li+ + 4e– Fe + 2S2–

4Li + FeS2

Fe + 2Li2S

Table 3.4 lists the cell parameters as well as advantages and disadvantages of the lithium–iron sulphide battery. Table 3.4: Technical data of the lithium–iron sulphide battery [25, 56]. Anode Cathode Electrolyte

Li FeS2 or FeS LiI in organic solvents

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

1.50 112 350 15 0

Advantages

Similar voltage as alkaline cells, higher durability, and capacity compared to alkaline cells (Not rechargeable) Small cells

Disadvantages Applications

3.2.5 Lithium–sulphur and sodium–sulphur accumulator The lithium-sulphur (LiS) and the NaS accumulator are based on the same principle. During discharge, the metal M (lithium or sodium) of the anode is ionized and the sulphur S8 rings of the cathode split, so that polysulphides emerge:

2M x S + 2e–

2M+ + 2e– Sx2–

2M + x S

M2Sx

50

3 Battery concepts: The past, the present, and research highlights

The cell types are differentiated according to their operating temperature in high-, intermediate-, and low-temperature cells. For NaS accumulators, the high-temperature cells are well established and widely used as stationary energy storages with worldwide 180 installations providing 334 MW [57]. The cell parameters are listed in Table 3.5. The electrodes of these cells are separated by a solid electrolyte, in most cases β′′-alumina. Both electrodes are liquid, which requires an operating temperature above their melting points (Tm ðNaÞ = 98  C, Tm ðSÞ = 115  C), generally between 270  C and 350  C [58]. The sodium ions diffuse through the β′′-alumina separator and react step by step with the liquid sulphur, generating ever shorter polysulphides. The formation of Na2S2 and Na2S is unwanted as they are solid and thus can cause local cell overcharge, increase the internal resistance, and restrain the reversibility [25]. However, avoiding these last two polysulphides reduces the capacity of the cell.

Table 3.5: Technical data of the LiS and NaS accumulator, values in parentheses are practical data [26, 44, 62–64]. Anode/cathode

Li/S

Liquid Na/liquid S

Electrolyte

LiPF6 or LiBF4

β′′ -Al2O3

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

2.5 1,700 (320) 600 (300)

2.08 150 ... 250 150 ... 250 3 ... 20

< 100

> 4,500

Advantages

Cheap, flexible operation

Disadvantages

Applications

Cheap, flexible operation, high cycle life, high energy density, good power density, high energy efficiency Dissolving polysulphides thicken Dissolving polysulphides thicken the the electrolyte, dendrite growth, electrolyte, thermal management 80 % cathode expansion, corroneeded, safety, durable seals needed sivity of Li Satellites Electronic vehicles, load leveling, peak shaving

Although LiS accumulators have been patented 5 years earlier than NaS cells [59, 60], their installation is more challenging. Due to the high reactivity of lithium, the operation with liquid metallic components is not desirable. The current research focus for LiS and NaS lies on the lowering of the operating temperature in order to reduce the cost for thermal shielding and thermal management. For the low-temperature accumulators, the use of a liquid electrolyte is

3.2 The present

51

necessary to guarantee sufficient ion diffusion [61]. In general, the low-temperature cells use the same mechanism as the high-temperature cells. Furthermore, the use of all possible polysulphide steps down to Na2S needs to be realized to increase the possible capacity of the cells. A big challenge in NaS or LiS low-temperature accumulators is the so-called polysulphide shuttle mechanism, which leads to self-discharging. The separator of the low-temperature cells is permeable to the long-chained S molecules, which then react with Na at the anode, and not at the cathode as necessary for discharging. The resulting shorter polysulphides diffuse back to the cathode and react to longer chain polysulphides [62]. The suppression of this process is indispensable to increase the capacity of these cells. 3.2.6 Metal fluoride accumulator To increase the voltage of a cell, the partial reactions need to have high standard potentials. Thus, the use of fluorine with a potential of 2.87 V could be a good way to reach this goal. Similar to the use of Li+ ions in lithium accumulators, the F– ions work as mobile species for metal fluoride accumulators: x F– + M′ M Fx + x e–

M′ Fx + x e– M + x F–

M′ Fx + M′

M′ + M Fx

Table 3.6 lists the cell parameters as well as advantages and disadvantages of the metal fluoride accumulator. Table 3.6: Technical data of the metal fluoride accumulators [26, 65], strongly dependent on the used metal. Anode Cathode Electrolyte

Ca, Li, La, Ce MnF2, CoF3, CuF2, BiF3, KBiF4, SnF2 LiPF6

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

1.8 ... 3.6 2100 ... 5200 760 ... 1640

Advantages Disadvantages

High cell voltage, high capacity Generation of reactive F2, low power, low cycle life, low ionic conductivity of the solid electrolyte (reversibility only for T > 150  C) Not yet applicable

Applications

40

52

3 Battery concepts: The past, the present, and research highlights

3.2.7 Metal–air battery Not only the cell voltage is a crucial parameter, also the energy density and specific energy need to be improved for future applications. One approach is to exclude the cathode from the cell itself and to have permanent supply of active material from the environment, e. g. by air. This idea was patented three times in 1969/1970 [66–68]. Today, these cells have been established especially as small portable cells for hearing devices. An overview of the cell parameters as well as advantages and disadvantages of the metal-air battery is given in Table 3.7.

Table 3.7: Technical data of metal–air batteries [25, 26, 69, 70], strongly dependent on the used metal. Anode Cathode Electrolyte

Metal, e. g. Zn, Al, Mg, Na O2 KOH

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) with/without air Self-discharge (% per month) Cycle number

1.3 ... 3.4 100 ... 200 200 ... 3,400/1300 ... 11,100 225 ... 1,000 0 ...

Advantages Disadvantages

High volumetric energy density, low self-discharge Dependent on environment, dendrite growth, regulation of O2 flow Button cells

Applications

Different metals M are suitable as anode material, and some of them even permit the construction of rechargeable cells, like zinc and iron. To allow the air to react within the cell, the current collector of the cathode needs to be permeable to air and porous:

4M O2 + 2H2O + 4e–

4M+ + 4e– 4OH–

4M + O2 + 2H2O

4M(OH)

3.2.8 Mercury battery The mercury battery was invented in 1884 [71] and commercialized in 1941 by the founder of Duracell, Samuel Ruben [72]. Table 3.8 lists the cell parameters as well as

3.2 The present

53

advantages and disadvantages of the mercury battery. Due to the hazardousness of mercury, this kind of battery is banned since 1992 by an EU directive [73]. Today, the mercury battery is completely replaced, mostly by silver oxide–zinc batteries and metal–air batteries: Zn + 2OH– HgO + H2O + 2e–

→ →

Zn(OH)2 + 2e– Hg + 2OH–

Zn + HgO + H2O

→

Zn(OH)2 + Hg

Table 3.8: Technical data of the mercury battery [25]. Anode Cathode Electrolyte

Zn HgO KOH

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

1.35 320 ... 400 85 ... 105

Advantages Disadvantages Applications

Very constant voltage Hazardous due to mercury Button cells, especially photometers (PX625), hearing aids

0

3.2.9 Molten salt accumulator The molten salt accumulator was invented in 1985 within the Zeolite Battery Research Africa (ZEBRA) project [74, 75]. Later, the meaning of the abbreviation was changed to Zero Emission Battery Research Activity, to lay the focus on the properties of the cell. The cell parameters as well as advantages and disadvantages of the molten salt accumulator are listed in Table 3.9. The cell operates with a liquid sodium anode (Tm ðNaÞ = 98  C) and a solid NiCl2 cathode (Tm ðNiCl2 Þ = 1001  C), in a temperature range of 250  C to 350  C [75]. Alternatively, FeCl2 can be used as cathode material, resulting in similar cell properties [76, 77]:

2Na 2Na+ + NiCl2

2Na+ + 2e– 2NaCl + Ni

2Na + NiCl2

2NaCl + Ni

54

3 Battery concepts: The past, the present, and research highlights

Table 3.9: Technical data of the molten salt accumulator [25, 26, 75, 78]. Anode Cathode Electrolyte

Liquid Na Solid NiCl2 Solid β′′ -Al2O3 or liquid NaAlCl4

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

2.59 60 ... 75 100 0 > 2,500

Advantages

High energy density, high specific energy, high safety, no selfdischarge, no cell damage by over-/undercharging Thermal isolation necessary, increasing internal resistance during discharge Electric vehicles, stationary applications

Disadvantages Applications

3.2.10 Nickel–cadmium accumulator The nickel–cadmium accumulator has prevailed over the nickel–iron accumulator, due to the improved temperature application range, the higher energy density, and specific energy, as well as lower self-discharge rates, although calendar life and cycle life are shortened [25], see also Table 3.10. Until the 1990s, the nickel–cadmium accumulator was the most used rechargeable cell. Afterward, the lithium ion and nickel–metal hydride technologies became dominant. In 2009, an EU directive prohibited the use of the cadmium cells (except for emergency applications) due to their toxicity [79]: Cd + 2OH– 2NiO(OH) + 2H2O + 2e–

Cd(OH)2 + 2e– 2Ni(OH)2 + 2OH–

0.81 V 0.49 V

2NiO(OH) + Cd + 2H2O

2Ni(OH)2 + Cd(OH)2

1.30 V

Table 3.10: Technical data of the nickel–cadmium accumulator [25, 26]. Anode Cathode Electrolyte

Cd NiO(OH) KOH

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

1.30 50 ... 150 40 ... 60 10 2,000 (continued )

3.2 The present

55

Table 3.10 (continued )

Advantages Disadvantages

Applications

Robust concerning deep discharge, high Cycle number, excellent long-time storage, low maintenance Voltage increase at the end of a cycle up to 1.6 V, voltage drops due to side products, toxicity of Cd, low energy density, corrosive electrolyte, temperature management needed Emergency lighting, emergency power supply [79], electrical vehicles, lighting and ventilation of trains, mine lamps

3.2.11 Nickel–metal hydride accumulator Research on the nickel–metal hydride accumulator (NiMH) started in the 1970s [80, 81]. Today, the NiMH type is one of the most important accumulators, due to its compatibility to the alkaline cells and its nontoxicity, see Table 3.11. The working principle corresponds to a (de)intercalation of hydrogen in the anode and cathode. To guarantee the rechargeability of the cell, the oxidation of the metal at the end of the discharge has to be avoided. Thus, the anode is oversized and the cathode determines the capacity of the cell: M H + OH– NiO(OH) + H2O + e– M H + NiO(OH)

M + H2O + e– Ni(OH)2 + OH–

0.83 V 0.49 V

M + Ni(OH)2

1.32 V

Table 3.11: Technical data of the nickel–metal hydride accumulator [25, 26, 82]. Anode Cathode Electrolyte

Metal alloy or metal–hydride powder, e. g. La0.8Nd0.2Ni2.5Co2.4Si0.1 Ni(OH)2 KOH

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

1.32 135 ... 250 60 ... 90 20 600

Advantages Disadvantages

Low resistance, high energy density, good cycle life Overcharging ! water electrolysis, H2 generation, overheating, sensitive to wrong poling, memory effect Electrical vehicles, small portable devices

Applications

56

3 Battery concepts: The past, the present, and research highlights

3.2.12 Silver oxide battery and accumulator The silver oxide battery reaches back to the first Voltaic pile [6], which was the basis for many experiments in the early field electrochemistry [8, 9], see Section 1. These cells were not rechargeable. The silver content makes these cells expensive; however, they are cost-effective in the form of button cells [25]: Zn + 2OH– AgO + H2O + 2e–

ZnO + H2O + 2e– Ag + 2OH–

Zn + AgO

ZnO + Ag

Due to the high energy density and specific energy of this cell, research for a rechargeable version started to obtain a better balance between product life and cost, see also Table 3.12: Zn + 2OH– AgO + H2O + 2e–

Zn(OH)2 + 2e– Ag + 2OH–

Zn + AgO + H2O

Zn(OH)2 + Ag

Table 3.12: Technical data of the silver oxide battery and accumulator [25, 83, 84].

Anode Cathode Electrolyte

Battery Zn AgO KOH

Accumulator Zn AgO KOH

Voltage (V) Energy density (Wh/l) Specific energy (Wh/kg) Self-discharge (% per month) Cycle number

1.55 500 130 30

0.85 80 ... 100 > 450

Further concepts of liquid metal batteries deal with the replacement of magnesium by lithium, due to its lower melting point (Tm ðLiÞ = 150  C) [115]. Still, the melting point of antimony is very high, but the addition of Pb lowers the melting point to 250  C while keeping the cell voltage almost constant at 0.85 V, see Table 3.15 [115]. Thus, the electrolytes entail the working temperature of 450  C, due to a melting point of 430  C:

Li Li + e + Sb / Pb

Li+ + e– Li in Sb / Pb

Li + Sb / Pb

Li in Sb / Pb

+

–

64

3 Battery concepts: The past, the present, and research highlights

3.3.5 Metal–air accumulator Based on the nonrechargeable metal–air batteries, the lithium–air accumulator was invented in 1996 [70, 116]. Due to the external cathode and the extremely high resulting specific energy (theoretically 13 kWh=kg, see Table 3.16), this type of accumulator is still in focus of recent research [117, 118]: 4Li O2 + 4e– 4Li + O2

4Li+ + 4e– 2O2– 2Li2O

2Li + O2

Li2O2

Table 3.16: Technical data of the lithium-air [25, 117, 118] and sodium-air accumulators [119–121]. Anode Cathode Electrolyte

Li Air Polymer mix with LiPF6

Na Air NaSO3CF3, NaPF6

voltage (V) Specific energy (Wh/kg) Cycle number

3.0 3000 > 2000

2.27 1100 > 300

Advantages

Flat discharge profile, environmentally friendly, long storage life Dependent on environment, low discharge rates, irreversible production of Li2CO3 and LiO2, drying out, safety concerns Vehicles, grid backup

No superoxide radicals, lower overpotentials compared to Li-air

Disadvantages

Applications

The sodium–air accumulator is inspired by the lithium–air accumulator, which has the disadvantages of large overpotentials during charge and discharge as well as the creation of highly reactive superoxide radicals O2– leading to the decomposition of the cell [119]. Both problems can be solved by replacing lithium by sodium and aiming for the formation of NaO2 as reaction product, as sodium superoxide is stable in contrast to LiO2. Thus, the chemical reactions within the sodium–air accumulator are different to the lithium–air accumulator. The formation of NaO2 is kinetically favoured (one electron transition) compared to Na2O2. Additionally, the NaO2 is thermodynamically more stable: Na O2 + e–

Na+ + e– O2–

Na + O2

NaO2

3.3 Research highlights

65

3.3.6 Paintable accumulator The paintable accumulator is designed for flexible and versatile cell solutions. In contrast to the tiny 3D cell, the paintable accumulator is designed for the application on buildings. Every part of the cell is produced as a paint and the accumulator is produced by multistep spray painting [122]. The spraying is performed by temperatures between 95  C and 120  C. Starting point is a nonconducting layer. Onto this layer, the cathode current collector (single-walled nanotubes SWNT) is sprayed, followed by the cathode (LiCoO2). Subsequently, the separator made of polymer is applied, followed by the anode (Li4Ti5O12) and the anode current collector (copper). Finally, the cell is soaked with the electrolyte. Table 3.17 lists the cell parameters of the paintable accumulator. It is crucial to use the correct temperature and drying steps before applying the next layer: Li7Ti5O12 3CoO2 + 3Li+ + 3e–

Li4Ti5O12 + 3Li+ + 3e– 3LiCoO2

Li7Ti5O12 + 3CoO2

Li4Ti5O12 + 3LiCoO2

Table 3.17: Technical data of the paintable accumulator, the specific energy is the sum of the given half cell values [122]. Anode Cathode Electrolyte

Li7Ti5O12 LiCoO2 LiPF6

Voltage (V) Specific energy (Wh/kg) Cycle number

2.5 562 > 60

3.3.7 Super-iron accumulator For certain applications, a high cell voltage is crucial. One way to increase the voltage is a multielectron transition during the cell reaction, instead of the one electron transition in lithium ion accumulators realized by high-valent elements, see Section 3.3.1. In 1999, the use of Fe(VI) salts in the form of ferrate salts (M FeO4, e. g. M = K2, Na2, Ba) as cathode material was proposed, due to their highly oxidized Fe basis, multiple electron transfer, and high intrinsic energy [123], see Table 3.18: 3Zn + 10OH– 2M FeO4 + 5H2O + 6e– 2M FeO4 + 3Zn

ZnO + 2ZnO22– + 5H2O + 6e– Fe2O3 + 10OH– + 2M2+ Fe2O3 + ZnO + 2M ZnO2

66

3 Battery concepts: The past, the present, and research highlights

Table 3.18: Technical data of the super iron accumulator [123, 124]. Anode Cathode Electrolyte

Zn ferrate salts, e. g. K2FeO4, BaFeO4 KOH

Voltage (V) Specific energy (Wh/kg) Cycle number

1.75 ... 1.85 420 ... 475 100 ... 200

3.3.8 Tin–sulphur–lithium accumulator The tin–sulphur–lithium accumulator is an approach to improve the properties of the LiS accumulator, such as voltage losses due to the shuttle mechanism and the formation of lithium dendrites [125]. Here, the final product of the cell reaction Li2S is intercalated into carbon and the electrolyte is replaced by a special gel polymer. This improvement prevents the dissolution of polysulphides and thus the shuttle mechanism. Another innovation is the use of tin as anode material, which is able to form alloys with lithium [125]. Table 3.19 lists the cell parameters of the tin–sulphur– lithium accumulator: 2.2Li2S / C 4.4Li + Sn / C + 4.4e– +

2.2Li2S / C + Sn / C

2.2S + C + 4.4Li+ + 4.4e– Li4.4Sn + C 2.2S + C + Li4.4Sn + C

Table 3.19: Technical data of the tin-sulphur-lithium accumulator [125–127]. Anode Cathode Electrolyte

Sn Li2S Gel polymer CGPE

Voltage (V) Specific energy (Wh/kg) Cycle number

2 1200 80

3.3.9 Virus-enabled electrodes In most cells, the reaction rate is constrained by the surface area, where the reaction can take place. Increasing this area means to improve the cell properties.

3.3 Research highlights

67

Biological systems like viruses have molecular recognition sites, which allow them to construct nanostructures suitable for electrodes. In 2006, the coat protein of the M13 virus was combined with tetra glutamate [128, 129]. This functional group interacts with cobalt ions, thus enabling the growth on a nanowire. The Co3O4 nanocrystals mineralize uniformly around the virus (diameter of the nanocrystals: 15nm). The virus adapts well to the growth of a monodisperse, highly crystalline nanowire. The so-constructed nanowire enhances the properties of the cell compared to a conventional Li ion accumulator. By adding gold nanoparticles to the growth solution of the Co3O4 nanowires, the properties can even be improved (e. g. higher capacity, see Table 3.20): Li Co3O4 + Li+ + e–

Li+ + e– LiCo3O4

Li + Co3O4

LiCo3O4

Table 3.20: Technical data of the virus-enabled electrodes [26, 128–130]. Anode Cathode Electrolyte

Li LiCo3O4 LiPF6

Voltage (V) Specific energy (Wh/kg) Cycle number

1.1 1100 N/A

3.3.10 Water accumulator The water accumulator is a “Non-Pollution Power” (NoPoPo) dry cell and was invented for the use in emergency goods like torches [131]. The electrolyte is not included in the accumulator to prevent self-discharge, which allows a shelf-life of 10 years for these cells. The customer can add the electrolyte anytime with a pipette. Generally, the electrolyte is water or a water-containing liquid, like blood or urine. The electrolyte also serves as cathode, combined with a carbon current collector within the cell. The anode consists of a magnesium alloy, see Table 3.21. The reactions are reversible, allowing a recharging of the cell for up to five times. Mg 2H2O + 2e–

Mg2+ + 2e– H2 + 2OH–

Mg + 2H2O

Mg(OH)2 + H2

2.36 V – 0.83 V 1.53 V

68

3 Battery concepts: The past, the present, and research highlights

Table 3.21: Technical data of the water accumulator [131]. Anode Cathode Electrolyte

Mg H2, OH– H2O

Voltage (V) Specific energy (Wh) Cycle number

1.53 0.75 per AA cell 5

3.4 Conclusion and outlook Although electricity has already been used in the ancient world, the scientific investigation only started in the eighteenth century. Since then, different battery concepts were developed, some of them prevailed until today, like the Leclanché cell or the lead accumulator. Some others vanished due to safety and health concerns, like the mercury battery or the nickel–cadmium accumulator. The large amount of new battery concepts clearly illustrates the need of the modern society for permanent access to electrical energy. Complementary to the dominant Li ion technologies, it is important to further develop other concepts, especially approaches with high-valent and multivalent elements.

References [1]

[2] [3] [4]

[5] [6]

[7] [8]

Kanani N. The Parthian battery: electric current 2,000 years ago? Fachzeitschrift des VINI. 2004;7:167. https://web.archive.org/web/20100401051957/http://www.vini.de/Gahname/ Gahname7/Gahname7-Parthian%20Battery.pdf König W. Ein galvanisches Element aus der Partherzeit? Forsch Fortschr. 1936;14:155. Franklin B. Franklin Papers, Volume 3, 29 April 1749 to Peter Collinson. p. 352 § 18. http://franklinpapers.org/franklin//framedVolumes.jsp?vol=3&page=352a Geschichte des Geschlechts von Kleist – Muttrin-Damensche Linie. Ergänzungen zu III. 350 – Ewald Jürgen, Schilderungen Ewald Jürgens zu seinen Experimenten und Informationen zum Ausbildungsgang. http://www.v-kleist.com/FG/Muttrin/fg0220.htm Piccolino M. Animal electricity and the birth of electrophysiology: the legacy of Luigi Galvani. Brain Res Bull. 1998;46:381. https://doi.org/10.1016/S0361-9230(98)00026-4. Volta A. Schriften über die thierische Elektrizität. Calve. 1793. https://books.google.de/books? hl=de&lr=& id=IMm6hvoMMWkC&oi=fnd&pg=PA1&dq=alessandro+volta&ots= ek2UVzeKmf&sig=WRjGXfQJFj1bXq9INeIMRhWN_ u4#v=onepage&q=alessandro%20volta&f= false Volta A. On the electricity excited by the mere contact of conducting substances of different kinds. Lett J. B. Bart. 1800. https://www.biodiversitylibrary.org/item/60966#page/53/mode/1up. Nicholson W, Carlisle A, Cruickshank M. IV. Experiments in galvanic electricity. Philos Mag: Comprehend Various Branch Sci, Liberal Fine Arts Agric Manuf Commer. 1800;7:337. 10.1080/ 14786440008562593.

References

[9] [10] [11] [12]

[13] [14] [15] [16]

[17] [18]

[19]

[20] [21] [22] [23]

[24]

[25] [26] [27] [28]

[29]

69

Ritter JW. Ladungssäule. Voigts Magazin für den neuesten Zustand der Naturkunde. 1803;6:105. Riecke E. Messung der von einer Zamboni’schen Säule gelieferten Electricitätsmenge. Annal Physik. 1883;256:512. 10.1002/andp.18832561110. Faraday M. Experimental researches in electricity. Philos Trans R Soc. 1832;122:125. doi:10.1098/rstl.1832.0006. Daniell JF. An introduction to the study of chemical philosophy. Landmarks of science. London: J.W. Parker, 1843:504. https://books.google.co.uk/books?id=bRkPAAAAYAAJ&pg=PA504#v= onepage&q&f=true Grove WR. XXIV. On voltaic series and the combination of gases by platinum. Philos Maga. 1839;3:127. 10.1080/14786443908649684. Leclanché G. French patent, Nr. 71865. 1866. Bunsen RW. Über eine neue Construction der galvanischen Säule. Eur J Org Chem. 1841;38:311. 10.1002/jlac.18410380306. Planté G. Nouvelle pile secondaire d’une grande puissance. In: Comptes rendus hebdomadaires des séances de l’Académie des sciences, vol. 50, 1860:640. http://gallica.bnf.fr/ark:/ 12148/bpt6k3007r/f653.image Clark L. On a standard voltaic battery. Philos Trans R Soc London. 1874;164. 10.1098/ rstl.1874.0001. Jungner EW. Anordning vid elektroder för elektriska accumulatorer. Kungliga Patent- och Registeringsverket, Patent No. 8558. 1897. http://was.prv.se/spd/pdf/gchKQBqp% 2BEDP0tTkAJNoTg%3D%3D/SE8558.C1.pdf. Jungner EW. Sätt att på elektrolytisk väg förstora ytan af sådana metaller, hvilkas syreföreningar äro kemiskt olösliga i alkaliska lösningar, Kungliga Patent- och Registeringsverket, Patent No. 15567. 1899. http://was.prv.se/spd/pdf/ NiU5SxtgOeTSfAo8BkBrHw%3D%3D/SE15567.C1.pdf Edison TA. Reversible galvanic battery. United States Patent, No. 678722. 1901. https://www. google.com/patents/US678722 Mayer SW, McKenzie DE. Lightweight secondary battery. United States Patent, No. 3185590. 1965. https://www.google.com/patents/US3185590. SwinkelsD AJ. Lithium-chlorine battery. J Elect Soc. 1966;113:6. 10.1149/1.2423867. Mizushima K, Jones PC, Wiseman PJ, Goodenough JB. LixCoO2 (0 < x ≤ 1): a new cathode material for batteries of high energy density. Mater Res Bull. 1980;15:783. DOI: https://doi.org/10. 1016/0025-5408(80)90012-4. Partridge BA, Jenkins TR, McGuire M. Sodium-sulphur battery cells. United States Patent, No. 3982959. 1976. https://worldwide.espacenet.com/publicationDetails/biblio?FT =D&date= 19760928&DB=&locale=de_EP&CC=US&NR=3982959A&KC=A&ND=1#. Reddy TD. Linden’s handbook of batteries. 4th ed. McGrawHill. New York, NY: United States, 2011. 978-0071624213. Kurzweil P, Dietlmeier OK. Elektrochemische Speicher. Wiesbaden: Springer Vieweg, 2015. 10.1007/978-3-658-10900-4. Daniel C, Besenhard J. Handbook of battery materials, vol. 1, 2nd ed. Weinheim: Wiley-VCH, 2012. 978-3-527-32695-2. Kangro W. Verfahren zur Speicherung von elektrischer Energie. Deutsches Patent, Nr. 914264 (C). 1954. https://worldwide.espacenet.com/publicationDetails/biblio?locale=de_EP&CC= DE&NR=914264#. Kangro W, Pieper H. Zur Frage der Speicherung von elektrischer Energie in Flüssigkeiten. Electrochim Acta. 1954;7:435. 10.1016/0013-4686(62)80032-2

70

3 Battery concepts: The past, the present, and research highlights

[30] Hagedorn NH. NASA redox storage system development project. Final report, No. DOE/NASA/ 12726-24, NASA-TM-83677. 1984. https://www.osti.gov/scitech/biblio/6472995. [31] Thaller LH. Electrically rechargeable redox flow cell. United States Patent, No. 3996064. 1976. https://www.google.com/patents/US3996064. [32] Ponce de León C, Frías-Ferrer A, González-García J, Szánto DA, Walsh FC. Redox flow cells for energy conversion. J Power Sources. 2006 9;160(1):716–732. DOI: 10.1016/j. jpowsour.2006.02.095. [33] Skyllas-Kazacos M, Rychick M, Robins R. All-vanadium redox battery. United States Patent, No. 4786567. 1988. https://www.google.com/patents/US4786567 [34] Remick RJ, Ang PG. Electrically rechargeable anionically active reduction-oxidation electrical storage-supply system. United States Patent, No. 4485154. 1981. https://www.google.com/ patents/US4485154. [35] Bradley CS. Secondary battery. United States Patent, No. 312802. 1880. https://www.google. com/patents/US312802. [36] Putt RA. Assessment of technical and economic feasibility of zinc/bromine batteries for utility load leveling. Rolling Meadows, IL: Final Report Gould, Inc., 1979. https://www.osti.gov/ scitech/biblio/6106713. [37] Skyllas-Kazacos M, Chakrabarti MH, Hajimolana SA, Mjalli FS, Saleem M. Progress in flow battery research and development. J Electrochem Soc. 2011;158:55. 10.1149/1.3599565. [38] Wang W, Luo Q, Li B, Wei X, Liu L, Yang Z. Recent progress in redox flow battery research and development. Adv Funct Mater. 2013;23:970. 10.1002/adfm.201200694. [39] Skyllas-Kazacos M, Kazacos G, Poon G, Verseema v. Recent advances with UNSW vanadiumbased redox flow batteries. Int J Energy Res. 2010;34:182. [40] Zhao P, Zhang H, Zhou H, Chen J, Gao S, Yi B. Characteristics and performance of 10 kW class all-vanadium redox-flow battery stack. J Power Sources. 2006;162:1416. 10.1016/j. jpowsour.2006.08.016. [41] Musée Tudor. curriculum vitae on the museum’s webpage. https://web.archive.org/web/ 20160909200212/http://musee-tudor.lu/lebenslauf [42] VARTA. VARTA brand history (1887–1909). https://www.varta-automotive.com/en-gb/whyvarta/varta-brand-history/1887-1909 [43] Power Sonic. Rechargeable sealed lead acid battery. data sheet PS-260 2 Volt 6.0 AH 2016. http://www.power-sonic.com/images/powersonic/sla_batteries/ps_psg_series/2volt/ PS260.pdf. [44] Leadbetter J, Swan LG. Selection of battery technology to support grid-integrated renewable electricity. J Power Sources. 2012;216:376. 10.1016/j.jpowsour.2012.05.081. and references therein [45] Wang Y, Liu B, Li Q. Cartmell S, Ferrara S, Deng ZD, et al. Lithium and lithium ion batteries for applications in microelectronic devices: a review. J Power Sources. 2015;286:330. 10.1016/j. jpowsour.2015.03.164. [46] Botouhi A, Auger DJ, Propp K, Longo S, Wild M. A review on electric vehicle battery modelling: from lithium-ion toward lithium-sulphur. Renew Sustain Rev. 2016;56:1008. 10.1016/j. rser.2015.12.009. [47] Diouf B, Pode R. Potential of lithium-ion batteries in renewable energy. Renew Energy. 2015;375. 10.1016/j.renene.2014.11.058. [48] Kraytsberg A, Ein-Eli Y. Higher, stronger, better... A review of 5 Volt cathode materials for advanced lithium-ion batteries. Adv Energy Mater. 2012;2:922. 10.1002/ aenm.201200068. [49] Fergus JW. Recent developments in cathode materials for lithium ion batteries. J Power Sources. 2010;195:939. 10.1016/j.jpowsour.2009.08.089.

References

71

[50] Zhang WJ. A review of the electrochemical performance of alloy anodes for lithium-ion batteries. J Power Sources. 2011;196:13. 10.1016/j.jpowsour.2010.07.020. [51] de las Casas C, Li W. A review of application of carbon nanotubes for lithium ion battery anode material. J Power Sources. 2012;208:74. 10.1016/j.jpowsour.2012.02.013. [52] Kalhoff J, Eshetu GG, Bresser D, Passerini S. Safer electrolytes for lithium-ion batteries: State of the art and perspectives. Chem Sus Chem. 2015;8:2154. 10.1002/cssc.201500284 [53] Alias N, Mohamad AA. Advances of aqueous rechargeable lithium-ion battery: A review. J Power Sources. 2015;274:237. 10.1016/j.jpowsour.2014.10.009. [54] Ordoñez J, Gago EJ, Girard A. Processes and technologies for the recycling and recovery of spent lithium-ion batteries. Renew Sustain Energy Rev. 2016;60:195. 10.1016/j.rser.2015.12.363 [55] Sharma RA, Wright WJ, Murie RA. Feedthrough assembly for lithium-iron sulphide cell. United States Patent, No. 4061841. 1977. https://www.google.com/patents/US4061841 [56] Energizer. Product Datasheet Energizer L91 (Ultimate Lithium), http://data.energizer.com/ pdfs/l91.pdf [57] Buß K, Wrobel P, Doetsch C. Global distribution of grid-connected electrical energy storage systems. Int J Sustain Energy Plann Manage 2016;9:31.10.5278/ijsepm.2016.9.4 [58] Ellis BL, Nazar LF. Sodium and sodium-ion energy storage batteries. Curr Opin Solid State Mater Sci. 2012;16:168. 10.1016/j.cossms.2012.04.002 [59] Danuta H, Juliusz U. Electric dry cell and storage battery. United States Patent, No. 3043896. 1962. https://www.google.com/patents/US3043896. [60] Kummer J, Weber V. A sodium-sulphur secondary battery. SAE Technical Paper, No. 670179 1967. 10.4271/670179. [61] Xin S, Yin YX, Guo YG, Wan LJ. A high-energy room-temperature sodium-sulphur battery. Adv Mater. 2014;26:1261. 10.1002/adma.201304126. [62] Tudron FB, Akridge JR, Puglisi VJ. Lithium-sulphur rechargeable batteries: Characteristics, state of development, and applicability to powering portable electronics. Sion Power Corporation. 2004. http://www.sionpower.com/pdf/articles/PowerSources2004.pdf. [63] Nazar LF, Cuisinier M, Pang Q. Lithium-sulphur batteries. Lithium Batteries Beyond. 2014;39:436. 10.1557/mrs.2014.86. [64] Kopera J. Sion Power’s lithium-sulphur batteries power high altitude pseudo-satellite flight. Sion Power Company Press Release. 2014. http://www.sionpower.com/pdf/articles/Sion_ Power_Zephyr_Flight_Press_Release_2014.pdf. [65] Reddy MA, Fichtner M. Batteries based on fluoride shuttle. J Mater Chem. 2011;21:17059. 10.1039/C1JM13535J. [66] Katsoulis EG, Prescia JJ. Metal/air battery. United States Patent, No. 3518123. 1970. https://www.google.com/patents/US3518123. [67] Moos AM. Metal/air batteries. United States Patent, No. 3531327 1970. https://www.google. com/patents/US3531327. [68] Landi HP, Voorhies JD. Metal-air battery including fibrillated cathode. United States Patent, No. 3462307. 1969. https://www.google.com/patents/US3462307. [69] Shao Y, Ding v, Xiao J, Zhang J, Xu W, Park S, et al. Making Li-air batteries rechargeable: Material challenges. Adv Funct Mater. 2013;23:987. 10.1002/adfm.201200688. [70] Abraham KM, Jiang Z. A polymer electrolyte-based rechargeable lithium/oxygen battery. J Electrochem Soc. 1996;143:1. 10.1149/1.1836378. [71] Clarke CL. Galvanic battery. United States Patent, No. 298175. 1884. https://www.google.com/ patents/US298175. [72] Fowler G. Samuel Ruben, 88, an inventor noted for electrochemical work. New York Times. 20 July 1988. http://www.nytimes.com/1988/07/20/obituaries/samuel-ruben-88-an-inventornoted-for-electrochemical-work.html.

72

3 Battery concepts: The past, the present, and research highlights

[73] Richtlinie 91/157/EWG des Rates vom 18. März 1991 über gefährliche Stoffe enthaltende Batterien und Akkumulatoren. EU Directive 91/157/EEC. 1991. http://eur-lex.europa.eu/legalcontent/DE/TXT/HTML/?uri=CELEX:31991L0157&from=EN. [74] Coetzer J, Galloway RC, Bones RJ, Teagle DA, Moseley PT. Electrochemical cell. United States Patent, No. 4546055. 1985. https://www.google.com/patents/US4546055. [75] Coetzer J. A new high energy density battery system, J Power Sources 1986;18:377. 10.1016/ 0378-7753(86)80093-3. [76] Bones RJ, Galloway RC, Coetzer J, Teagle DA. Electrochemical cell. United States Patent, No. 4612266. 1986. https://www.google.com/patents/US4612266. [77] Moseley PT, Bones RJ, Teagle DA, Bellamy BA, Hawes RW. Stability of beta alumina electrolyte in socium/FeCl2 (zebra) cells. J Electrochem Soc 1989;136:1361. 10.1149/1.2096922. [78] Rummich E. Energiespeicher: Grundlagen, Komponenten, Systeme und Anwendungen. Renningen: Expert-Verlag. 2011. 978-381-692736-5. [79] Europäisches Parlament. Sammlung, Behandlung und Recycling von Altbatterien und Altakkumulatoren. 2006. http://www.europarl.europa.eu/sides/getDoc.do?language= de&type=IM-PRESS&reference=20060628BRI09328&secondRef=ITEM-003-de. [80] Beccu K. Accumulator electrode with capacity for storing hydrogen and method of manufacturing said electrode, United States Patent, No. 3669745 1972. https://www.google.com/ patents/US3669745. [81] Beccu K. Negative electrode of titanium-nickel alloy hydride phases. United States Patent, No. 3824131 1974. https://www.google.com/patents/US3824131. [82] Energizer. Nickel metal hydride (NiMH) – Handbook and application manual. http://data.energizer.com/pdfs/nickelmetalhydride_appman.pdf. [83] Duracell. Silver oxide. Datasheet. 2009. https://web.archive.org/web/20091220201115/ http://www.duracell.com/ procell/chemistries/silver.asp. [84] DiCicco M. NASA research helps take silver-zinc batteries from idea to the shelf. NASA press release, 2016. https://www.nasa.gov/directorates/spacetech/spinoff/feature/Silver-Zinc_ Batteries [85] Glover D, Kozawa A, Schumm B. Handbook of manganese dioxides, battery grade. Int Battery Mater Asso. IC Sample Office, 1989. [86] Crowley CA, Langdon WM, Louzos DV. Battery cells. United States Patent, No. 2921110. https://www.google.com/patents/US2921110. [87] Mehta SA, Bonakdarpour A, Wilkinson DP. Impact of cathode additives on the cycling performance of rechargeable alkaline manganese dioxide-zinc batteries for energy storage applications. J Appl Elecrochem. 2017;47:167. 10.1007/s10800-016-1034-1. [88] Whitlock M. Panasonic oxyride editorial review – The revolution in battery power. On: Techlore. com, http://techlore.com/article/panasonic-oxyride-editorial-review-revolution-battery-power. [89] AccuCell. Der Aufbau der neuen AccuCell Batterien, http://cdn-reichelt.de/documents/daten blatt/D400/AccuCell_Datenblatt.pdf [90] Muldoon J, Bucur CB, Gregory T. Quest for nonaqueous multivalent secondary batteries: magnesium and beyond. Chem Rev. 2014;114:11683. 10.1021/cr500049y. [91] Nishi Y. Lithium ion secondary batteries; past 10 years and the future. J Power Sources. 2001;100:101. 10.1016/S0378-7753(01)00887-4. [92] Mayers MZ, Kaminski JW, Miller III TF. Suppression of dendrite formation via pulse charging in rechargeable lithium metal batteries. J Phys Chem C. 2012;116:26214. 10.1021/ jp309321w. [93] Aurbach D, Zinigrad E, Teller H, Dan dP. Factors which limit the cycle life of rechargeable lithium (metal) batteries. J Electrochem Soc. 2000;147:1274. 10.1149/1.1393349.

References

73

[94] de Jongh PE, Blanchard D, Matsuo M, Udovic TJ, Orimo S. Complex hydrides as room-temperature solid electrolytes for rechargeable batteries. Appl Physics A. 2016;122:1. 10.1007/ s00339-016-9807-2. [95] Kharton VV. Solid state electrochemistry I: Fundamentals, materials and their applications. John Wiley & Sons, 2009. 978-3-527-32318-0. [96] Guduru RK, Icaza JC. A brief review on multivalent intercalation batteries with aqueous electrolytes. Nanomater 2016;6:41. 10.3390/nano6030041. [97] Muldoon J, Bucur CB, Oliver AG, Sugimoto T, Matsui M, Kim HS, et al. Electrolyte roadblocks to a magnesium rechargeable battery. Energy Environ Sci. 2012;5:5941. 10.1039/C2EE03029B. [98] Lapidus SH, Rajput NN, Qu X, Chapman KW, Persson KA, Chupas PJ. Solvation structure and energetics of electrolytes for multivalent energy storage. Phys Chem Chem Phys. 2014;16:21941. 10.1039/C4CP03015J. [99] Liu M, Rong Z, Malik R, Canepa P, Jain A, Ceder G, et al. Spinel compounds as multivalent battery cathodes: A systematic evaluation based on ab initio calculations. Energy Environ Sci. 2015;8:964. 10.1039/C4EE03389B. [100] Li Q, Chen J, Fan L, Kong X, Lu Y. Progress in electrolytes for rechargeable Li-based batteries and beyond, Green Energy & Environ. 2016;1:18. 10.1016/j.gee.2016.04.006. [101] Anantharamulu N, Koteswara Rao K, Rambabu G, Vijaya Kumar B, Radha V, Vithal M. A wideranging review on NASICON type materials. J Mater Sci. 2011;46:2821. 10.1007/s10853-011-5302-5. [102] Smart LE, Moore EA. Solid state chemistry: An introduction. CRC press, 2005. 978-1-43984790-9. [103] Alamo J, Roy R. Crystal chemistry of the NaZr2(PO4)3, NZP or CTP, structure family. J of Mater Sci. 1986;21:444. 10.1007/BF01145507 [104] Alamo J. Chemistry and properties of solids with the [NZP] skeleton. Solid State Ionics. 1993:63:547. 10.1016/0167-2738(93)90158-Y. [105] Ellis BL, Nazar LF. Sodium and sodium-ion energy storage batteries, Curr Opin Solid State Mater Sci. 2012;16:168. 10.1016/j.cossms.2012.04.002. [106] Hong HY. Crystal structures and crystal chemistry in the system Na1+xZr2SixP3–xO12. Mater Res Bull. 1976;11:173. DOI: 10.1016/0025-5408(76)90073-8. [107] Sun K, Wei T-S, Ahn BY, Seo JY, Dillon SJ, Lewis JA. 3D printing of interdigitated Li-ion microbattery architectures. Adv Mater. 2013;25:4539. 10.1002/adma.201301036 [108] Lewis J, Dillon S, Sun K, Ahn BY, Wei T-S. Three-dimensional (3d) electrode architecture for a microbattery. United States Patent Nr. 20160126558. https://www.google.com/patents/ US20160126558 [109] Zhao B, Ran R, Liu M, Shao Z. A comprehensive review of Li4Ti5O12-based electrodes for lithium-ion batteries: The latest advancements and future perspectives. Mater Sci Eng R. 2015;98:1. DOI: 10.1016/j.mser.2015.10.001. [110] Human JP. Fuel cell. United States Patent No. 2012/0208104. 2012. https://www.google.com/ patents/US20120208104 [111] Agruss B. Regenerative Battery. United States Patent No. 3245836. 1966. https://www.google. com/patents/US3245836 [112] Shen Y, Zikanov O. Thermal convection in a liquid metal battery. Theor Comput Fluid Dyn. 2015;30:275. 10.1007/s00162-015-0378-1 [113] Bradwell DJ, Kim J, Sirk AHC, Sadoway DR. Magnesium-antimony liquid metal battery for stationary energy storage. J Am Chem Soc. 2012;134:1895. 10.1021/ja209759s [114] Leung P, Heck SC, Arnietszajew T, Mohamed MR, Conde MB, Dashwood RJ, et al. Performance and polarization studies of the magnesium-antimony liquid metal battery with the use of in-situ reference electrode. RSC Adv. 2015;5:83096. 10.1039/C5RA08606J

74

3 Battery concepts: The past, the present, and research highlights

[115] Wang K, Jiang K, Chung B, Ouchi T, Burke PJ, Boysen DA, et al. Lithium-antimony-lead liquid metal battery for grid-level energy storage. Nature. 2014;514:348. 10.1038/nature13700 [116] Abraham KM. Solid polymer electrolyte-based oxygen batteries. United States Patent No. 5510209. 1996. https://www.google.com/patents/US5510209 [117] Liu T, Leskes M, Yu W, Moore AJ, Zhou L, Bayley PM, et al. Cycling Li-O2 batteries via LiOH formation and decomposition. Science. 2015;350:530. 10.1126/science.aac7730 [118] Wang J, Li Y, Sun Y. Challenges and opportunities of nanostructured materials for aprotic rechargeable lithium-air batteries. Nano Energy. 2015;2:443. 10.1016/j.nanoen.2012.11.014 [119] Hartmann P, Bender CL, Vračar MS, Dürr AK, Garsuch A, Janek J, et al. A rechargeable roomtemperature sodium superoxide (NaO2) battery. Nat Mater. 2013;12:228. 10.1038/nmat3486 [120] Sun W, Yang Y, Fu Z-W. Electrochemical properties of room temperature sodium air batteries with nonaqueous electrolyte. Electrochem Commun. 2012;16:22. 10.1016/j.elecom.2011.12.019 [121] Adelhelm P, Hartmann P, Bender CL, Busche M, Eufinger C, Janek J. From lithium to sodium: cell chemistry of room temperature sodium–air and sodium–sulphur batteries. Beilstein J Nanotechnol. 2015;6:1016. 10.3762/bjnano.6.105 [122] Singh N, Galande C, Miranda A, Mathkar A, Gao W, Reddy ALM,et al. Paintable battery. Sci Rep. 2012;2:481. 10.1038/srep00481 [123] Licht S., Wang B, Ghosh S. Energetic Iron(VI) Chemistry: The Super-Iron Battery. Science 1999:1039–42. DOI: 10.1126/science.285.5430.1039.285. [124] Licht S, Tel-Vered R. Rechargeable Fe(III/VI) super-iron cathodes. Chem Commun. 2004;0:628. 10.1039/B400251B [125] Hassoun J, Scrosati B. A high performance polymer tin sulphur lithium ion battery. Angewandte Chem. 2010;49:2371. 10.1002/anie.200907324 [126] Hassoun J, Fernicola A, Navarra MA, Panero S, Scrosati B. An advanced lithium-ion battery based on a nanostructured Sn/C anode and an electrochemically stable LiTFSi-Py24TFSI ionic liquid electrolyte. J Power Sources. 2010;185:574. 10.1016/j.jpowsour.2009.07.046 [127] Hassoun J, Sun Y-K, Scrosati B. Rechargeable lithium sulphide electrode for a polymer tin/ sulphur lithium-ion battery. J Power Sources. 2011;196:343. 10.1016/j.jpowsour.2010.06.093 [128] Nam KT, Kim D-W, Yoo PJ, Chiang C-Y, Meethong N, Hammond PT, et al. Virus-enabled synthesis and assembly of nanowires for lithium ion battery electrodes. Science. 2006;312:885. 10.1126/ science.1122716 [129] Lee YJ, Lee Y, Oh D, Chen T, Cede G, et al. Biologically activated noble metal alloys at the nanoscale: For lithium ion battery anodes. Nano Lett. 2010;10:2433. 10.1021/nl1005993 [130] Cabana J, Monconduit L, Larcher D, Palacín MR. Beyond Intercalation-Based Li-Ion Batteries: The State of the Art and Challenges of Electrode Materials Reacting Through Conversion Reactions. Advanced Materials 2010:E170–E192. DOI: 10.1002/adma.201000717 22. [131] Japan Trend Shop. NoPoPo eco water-powered AA batteries. http://www.japantrendshop.com/ nopopo-eco-waterpowered-aa-batteries-p-546.html

4 Battery Materials 4.1 Computational analysis and identification of battery materials Falk Meutzner, Tina Nestler, Matthias Zschornak, Pieremanuele Canepa, Gopalakrishnan S. Gautam, Stefano Leoni, Stefan Adams, Tilmann Leisegang, Vladislav A. Blatov and Dirk C. Meyer Abstract: Crystallography is a powerful descriptor of the atomic structure of solid-state matter and can be applied to analyse the phenomena present in functional materials. Especially for ion diffusion – one of the main processes found in electrochemical energy storage materials – crystallography can describe and evaluate the elementary steps for the hopping of mobile species from one crystallographic site to another. By translating this knowledge into parameters and search for similar numbers in other materials, promising compounds for future energy storage materials can be identified. Large crystal structure databases like the ICSD, CSD, and PCD have accumulated millions of measured crystal structures and thus represent valuable sources for future data mining and big-data approaches. In this work we want to present, on the one hand, crystallographic approaches based on geometric and crystal-chemical descriptors that can be easily applied to very large databases. On the other hand, we want to show methodologies based on ab initio and electronic modelling which can simulate the structure features more realistically, incorporating also dynamic processes. Their theoretical background, applicability, and selected examples are presented. Keywords: crystallography, electrochemistry, Voronoi–Dirichlet partitioning, bond valence sum, density functional theory

4.1.1 Introduction The exponential growth of computer-processing power described by the empirical “Moore’s Law”, formulated in 1965 by Gordon Moore, has led to increasingly faster processors. This enabled powerful and very diverse computational methods including algorithms with a higher demand of computational power. They are becoming widely used today for tackling problems in materials science and engineering. Big-data approaches as well as multiphysics modelling employing finite elements all the way to quantum mechanical modelling can nowadays virtually simulate and test thousands of materials at a time. In science as well as in industry, these methods are becoming This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Meutzner, F., Nestler, T., Zschornak, M., Canepa, P., Gautam, G. S., Leoni, S., Adams, S., Leisegang, T., Blatov, V. A., Meyer, D. C. Computational analysis and identification of battery materials. Physical Sciences Reviews [Online] DOI: 10.1515/psr-2018-0044 https://doi.org/10.1515/9783110493986-004

76

4 Battery Materials

increasingly important to predict functional materials with enhanced properties: In Ref. [1], Ceder and Persson state that “yet materials science is on the verge of a revolution”. With respect to the development and discovery of new battery materials, computational methods are becoming a main driver. Hence, they will be a key enabler for the storage of large amounts of renewable energies and thus its further installation, since low cost and durable systems with preferably higher energy densities are needed. In order to use the potential to develop materials for ground-breaking technologies, the availability of infrastructure for High-Performance Computing (HPC) is the key factor [2]. In 2011, the Materials Genome Initiative (MGI) was established in the US, promising a renaissance of the American manufacturing industry by at least doubling the pace of discovery, development, and deployment of new materials so that they are introduced to the market at a fraction of the cost [3]. Several other national computing clusters as well as the Partnership for Advanced Computing in Europe (PRACE) has thus been founded or extended in recent years in order to offer computing and data management resources and services [2, 4–6]. In former days, the only way to find and explore new materials was to synthesise and test an enormous amount of different compounds based on the researcher’s experience [7].1 Once a variety of successful compounds and structures has been found by this trial-and-error-technique, advanced materials can be accessed by atomic substitution or doping. Even though material engineering can, for instance, be performed by experimental high-throughput combinatorial screening nowadays [8, 9], this approach is still extremely time-consuming and will rather lead to evolutionary than revolutionary findings. Additionally, the hope to be able to transfer compounds or at least structural motifs from the lithium-ion-battery to high-valent2 battery materials such as magnesium [10] and aluminium [11] often failed, demonstrating the need to start from scratch. A more time-efficient route would rather be to employ computational methods. They are already powerful enough to eliminate significant parts of the guesswork, as they can predict many properties relevant for battery materials before they are synthesised in the lab. For instance, by scaling material computations over supercomputing clusters [4], Ceder’s group has predicted several new battery materials, which were then synthesised and tested in the lab. For instance, on this basis, Pellion Technologies was founded for the development and commercialisation of high-energy-density

1 As stated in [1, 3], the Massachusetts Institute of Technology has found that the commercialisation of a successful material from lab-scale takes an average of 15–20 years. This long and resource-tying process costs companies and research institutes billions of US Dollars [1]. High-throughput computational materials design is capable to dramatically reduce costs and time spent on this. 2 In the scientific literature, the term multivalent is widely spread. We propose to avoid this term since it originally meant ions of multiple valence states (such as Cr). A synonym would be polyvalent. We suggest the term high-valent or highly valent and use it in this manuscript. Many high-valent cations are also multivalent but cations mostly regarded for future electrochemical energy storage devices are not multivalent (in this sense), like Al or Mg.

4.1 Computational analysis and identification of battery materials

77

rechargeable magnesium-ion batteries for applications that range from portable electronics to electrified vehicles [12, 13]. Besides the design and evaluation of new compounds, the search among already known materials by means of high-throughput theoretical methods appears highly promising, since for example the most successful cathode materials for lithium-ion batteries have been known for years, before they were considered and validated for the use in batteries.3 This fact motivated several studies to screen structural databases such as the Inorganic Crystal Structure Database (ICSD) for new battery materials and already lead to new discoveries of such compounds (e.g. [14–19]). Furthermore, in combination with already established or future available experimental characteristics, the data created during the screening process can lead to a deeper insight into the fundamental relationship between the structure and functional properties of battery materials (e.g. shown in [20, 21]), which helps to optimise and design the next generation of battery materials. Based on preliminary data mining, the setting up of structure-property relations is a valuable productive approach commonly used in crystallography and materials science. According to Merriam-Webster [22], data mining is described as “the practice of searching through large amounts of computerised data to find useful patterns or trends”, using pattern recognition technologies as well as statistical and mathematical techniques (see [23] for details). Eventually, by applying the outcome of continuing data mining in an iterating way, the so-called machine-learning is achieved, which will decisively accelerate our knowledge growth [24]. Nowadays, databases are increasing in size and complexity, eventually even coining the term big data, which is stated to become a key basis of competition, productivity growth, innovation, and consumer surplus [25]. For instance, the collection of data thus allows drawing conclusions from materials with the same function but very different crystallographic structures. The more data is generated and analysed, the more reliable the results become and the more reliable predictions can be derived utilising this data. By applying the knowledge gained through data mining, different data sets may be identified that comprise compounds with the same functionality but have not yet been analysed for it. Several examples based on data mining for targeting battery materials are known: electrolytes [19, 26, 27], cathodes [4, 28], and electrodes (chapter 4.1). 3 The rhombohedral (trigonal) structure of Li0.5CoO2, for instance, was first described in 1958 in Ref. [205], according to the datasets found within the Inorganic Crystal Structure Database (ICSD) [79]. This material was investigated regarding its magnetic properties, elucidating the concept of doubleexchange interaction. The structure was already well-known but put in a very different context in 1980 – 22 years later – when it was first proposed as an interesting positive electrode candidate by the group of John B. Goodenough [206]. For LiFePO4 the numbers are even further apart. Its use for Li-ion batteries was first proposed in 1997, again by the group of John B. Goodenough [207] but the material itself is already known as a mineral since 1834 [208]. J. N. Fuchs even obtained Li from this mineral through acidic dissolution 1 year later [209]. According to the datasets in the ICSD, the structure was first solved in 1938 [210] whereas the first synthesised structure is from 1977 [211].

78

4 Battery Materials

Besides battery materials and in particular the “Materials Project” [4], other projects and initiatives (see in Ref. [9]) also make use of this approach: the Harvard Clean Energy Project [29], which has focused on photovoltaic materials, AFLOWlib [30], which addresses the electronic structure of inorganic compounds to uncover thermoelectric and scintillator materials as well as magnetic materials for energy and spintronics applications, and OQMD, which specialises in thermodynamic properties [31]. These screenings target different material properties and apply diverse methods. An overview on the figures of merit that can be evaluated today for electrodes and electrolytes and the corresponding applicable theoretical methods is given in Figure 4.1.1. Theoretical capacity

Chemical composition & valence state Thermodynamic data

Insertion voltage Electrode

DFT energy calculation

Electronic conductivity

DFT electronic structure calculation

Volume change

DFT structural relaxation Thermodynamic data

Structural stability DFT energy calculation Thermodynamic data

Electrode & electrolyte

Chemical stability DFT interfacial simulations Geometry analysis; Bond valence

Ionic conductivity DFT + NEB; MD; MC

Electrolyte

Electrochemical window

DFT electronic structure calculation

Figure 4.1.1: Relevant properties to forecast the performance of electrode and electrolyte materials that can be simulated, and theoretical methods used to calculate them (DFT: density functional theory, NEB: nudged elastic band, MD: molecular dynamic; MC: Monte Carlo). Scheme was adapted from Ref. [35].

For instance, to select candidates for new solid electrolytes, high ionic conductivity is the most important criterion besides structural and electrochemical stability. For insertion-type electrode materials, the ionic conductivity in the electrodes limits the rate capability [21]. Typically, a balance has to be found between these, as high ion mobility also tends to reduce the structural stability. Precise simulations of migration barriers with density functional theory (DFT) employing the nudged elastic band (NEB) method [32] for thousands of structures, however, would take years or even decades.

4.1 Computational analysis and identification of battery materials

79

Even then, the migration barriers will only allow for a rough prediction of the absolute room temperature conductivities, which strongly depend on the often non-equilibrated defect concentrations and distributions in real materials. Practically, the reliability of DFT predictions also depends on whether all potentially relevant ion migration pathways, e.g. those between nominally unoccupied sites, have been included in the considerations. Direct screening of the ion dynamics by ab initio molecular dynamics (MD) simulations at relevant temperatures still appears out of reach and will anyways not solve the problems of defect influences. For polycrystalline materials, the microstructure often limits the practically achievable conductivity. This, as well as the lower variability of entropy in strongly disordered glasses, contributes to the finding that the prediction of absolute conductivities from local structure models via a determination of migration barriers is more reliable for glasses than for crystalline compounds [33]. Since the significantly higher effort in determining migration barriers by DFT cannot ensure a significantly more precise prediction of absolute conductivities anyways, some studies undertake computationally less expensive and intrinsically less accurate estimations of the migration barriers first, such as the geometry-based Voronoi-Dirichlet Partitioning [19] or bond valence sum (BVS) methods [34–36], like bond valence site energies (BVSE) [16, 37–39], where part of the saved computational cost can be spent on treating more complex local structure models and on ensuring that no pathway is overlooked. Further methods comprise e.g., Hirshfeld and procrystal analysis [40, 41]. It has to be noted, however, that the applicability and forecasting power of these methods is generally limited in comparison to DFT and in particular to dynamic approaches such as MD and Monte Carlo (MC) simulations. Voronoi-Dirichlet partitioning and BVS, or BVSE, are not applicable to alloys and in order to utilise them for other amorphous compounds, a representative local structure model of the amorphous materials has to be generated first. Since this involves rather time-consuming methods such as Reverse MC fitting of neutron and synchrotron diffraction data [42–45], a screening of amorphous materials appears not feasible yet. Additionally, all the listed approaches only simulate on an atomic or molecular scale. In order to realistically predict the material’s performance in the battery, other so-called multiscale techniques are needed in the future, which incorporate microstructural defects and describe the interplay between the individual materials and the electrical and physical behaviour of the cell as a whole. One example for a possible algorithm is the application of finite element methods as implemented for example in Comsol’s multiphysics programme [46–48]. By inputting a number of physical and chemical parameters, the operating cell can be simulated. On the other hand, with respect to thiophosphate-based solid electrolytes in lithium all-solid-state batteries, the bulk ion conductivity is not rate limiting anymore. In fact, the grain boundary resistance as well as the interface resistance to the electrode dominates the cell performance ([49], for a theoretical description see chapter 4.6). However, it is not possible to take these phenomena into account by screening methods, yet, since we are just on the way to fully understand and describe them.

80

4 Battery Materials

Besides the typical criteria for good electrodes and electrolytes, which are discussed in the chapter 4.6, the simulation of the volume change is of interest in the case of insertion or phase transition electrodes. A small volume change is generally related to a longer cycle life of lithium-ion batteries [51] and even more important for the improvement of all-solid-state batteries. To sum up, the use of computational methods, in particular high-throughput technologies and big-data approaches, appears promising for identifying and simulating materials with dedicated or novel properties in order to realise new concepts for electrochemical energy storage. This review gives an overview about relevant methods developed for deepening the understanding of and finding novel battery materials. The presented approaches can be applied for both predicting the properties of a specific compound and screening structure databases for potential battery materials. They cover the simulation of different properties and also different accuracy levels and thus computational effort. For the screening of crystalline solid electrolytes and intercalation materials, the authors thus suggest using these methods in succession in order to build a “filter”; that means e.g. first Voronoi-Dirichlet Partitioning, followed by BVbased methods and finally DFT and MD (Figure 4.1.2). Promising candidates are picked

Crystal structure data base

STRUCTURES

Potential ion migration paths

VDP

Migration barrier (estimation)

BVSE Activation energy for ion conduction Electronic band gap

DFT

Experimental validation of electric conductivities Figure 4.1.2: Scheme of the suggested screening approach for crystalline materials with fast ionic transport. Simulation methods with different accuracy levels and thus computational effort, as indicated by the clock symbol, are performed in succession.

4.1 Computational analysis and identification of battery materials

81

out first to be evaluated more precisely with increasingly time-consuming but more accurate methods. This sequential use thus allows a fast approach to reliably screen thousands of compounds. In the following, the respective methodologies are presented and their features and characteristics are highlighted.

4.1.2 Voronoi-Dirichlet partitioning 4.1.2.1 Introduction The classic description of crystallography focuses on space groups and embedded atoms on Wyckoff sites. Since its introduction, many different descriptions have been presented in the crystallographic literature, that is e.g. the assignment of a threedimensional periodic arrangement of atoms to a unit cell belonging to one of the 230 space groups [52], the determination of the chemical topology as independent one-, two-, and three-dimensional interpenetrating nets [53–55], or the partitioning of the atomic environments by coordination polyhedra or clusters, e.g. Frank-Kasper and other polyhedra [56–59]. Thus, atomic arrangements can be analysed from a solely geometric point-of-view. Comparing the arrangement of bonded atoms in thousands of existing compounds, geometrical and topological analyses reveal relationships – recurring structural motifs – between numerous crystal structures even though they may not be directly related from a space group point-of-view. In crystal chemistry, atoms are usually described as hard spheres or point-like objects that are positioned in a 3D space. Therefore, these atomic arrangements can be geometrically understood as point arrangements in space. Voronoi diagrams or Dirichlet domains have been widely used in geometrical analyses to describe the relationships and domains of single points in point arrangements. Especially in computer algorithms for the retrieval of data in databases, collision control, and the analysis of clustering, Voronoi diagrams have been extensively used. Further applications are e.g. described by Aurenhammer [60]. The polyhedron created through Voronoi-Dirichlet partitioning is called “Voronoi-Dirichlet polyhedron” (VDP) [61, 62] and has been interpreted crystalchemically since 1927 [63], with interest intensifying since around 1995 [64, 65]. Voronoi-Dirichlet partitioning geometrically subdivides a space filled with points (atoms). This VDP is generated4 for a given point i (the central point) in an assembly of its n neighbouring points j by constructing planes perpendicular to and midway on all line segments ij that connect the central point to all other points. The smallest polyhedron created in this way is the VDP of the point i. Therefore, the VDP pinpoints towards the nearest neighbours of a point because the planes created between close

4 For computation of Voronoi-Dirichlet partitioning, a faster and more efficient “gift wrapping” algorithm is used [65].

82

4 Battery Materials

points are closest to the central point. Each additional point that lies inside this construct is closer to point i than to any other point j. Figure 4.1.3 shows a twodimensional representation of the VDP. For atomic structures, each point is identified with an atom or ion (for the sake of readability, the terms atom and ion will be subsumed to a generic atom). These polyhedra therefore describe a certain volume (or domain) in a structure that can be assigned to a certain atom, as each point in this domain is closer to this atom than to any other atom (Figure 4.1.3). The idea of these domains of action (“Wirkungsbereiche”) was already proposed by Niggli in 1927 and lays thus the foundation of topological crystal chemistry [63]. For periodic real space lattices, the VDP is thus equivalent to the Wigner-Seitz elementary cell and the application of the same construction method to reciprocal lattices yields the first Brillouin zone. More details of the history and the development of Voronoi-Dirichlet partitioning, which can in its simplest form be traced back to René Descartes’ work [66], can be found in [65, 67, 68]. The following sections will subsume and refer to many of the authors’ findings already published in Refs [19, 27].

Figure 4.1.3: Left-hand side: two-dimensional representation of a Voronoi diagram. Segments between each blue dot are drawn and in the middle, a perpendicular line is constructed. The smallest polyhedron created around a given reference dot is its Voronoi-Dirichlet polyhedron (e.g. the area shaded in red). This figure was created with [69]. Right-hand side: Construction of a VDP for a bodycentred cubic structure (blue atoms) with the corresponding vertices (grey) that signify voids in the structure.

4.1.2.2 The Voronoi-Dirichlet approach The Voronoi-Dirichlet approach, carried out with the programme package “ToposPro” as described in [27, 54, 70], is a sophisticated crystal-chemically motivated and geometrically performed, high-throughput crystallographic analysis that can be easily applied to large crystal structure databases. The VDP of an atom in a structure is

4.1 Computational analysis and identification of battery materials

83

characterised by its volume, faces, vertices, and edges [65]. Volumes can be used to describe the size and shape of this atom in the crystal structure in the case of homogeneous environment of an atom (e.g. oxygen environment of metal atoms in oxides). The size of a face of a VDP is proportional to the strength of the chemical bonding to a neighbouring atom for a given atomic pair (e.g. cation-anion like Mg2+ and O2−). The closer a neighbour, the larger a face and the higher the attractive force between the central atom and its neighbour and thus the stronger the bond. By finding the atoms that share the largest faces of a VDP, the nearest neighbours are identified. According to Ref. [65], the face sizes can also be described by solid angles Ω – specifying the percentage of the face projected onto a unit sphere. The vertices of a VDP describe possible voids in the structure because they are farthest away from all surrounding atoms. Each vertex is connected to neighbouring vertices through VDP edges. These edges are, in analogy to the vertices, the farthest away from their constructing atoms. They can thus be regarded as channels between these voids (see Figure 4.1.3 and Figure 4.1.4).

1

2

3

4

5

Figure 4.1.4: Graphical demonstration of the Voronoi-Dirichlet approach applied to Na β-alumina (ICSD-# 9144). From the structure (1) all Na-ions (light blue) are ignored (2) and the VDP of all remaining atomic sites (selection in red) are calculated (3). For each of these VDPs vertices, a further VDP (green) is constructed (4) and compared to data-mined values in order to generate the conduction path (blue) (5). Figure reproduced with permission from [F. Meutzner, W. Münchgesang, T. Leisegang, R. Schmid, M. Zschornak, M. Ureña de Vivanco, A. P. Shevchenko, V. A. Blatov, and D. C: Meyer: “Identification of solid oxygen-containing Na-electrolytes: An assessment based on crystallographic and economic parameters”, Crystal Research and Technology 52, 1600223 (2017).], © 2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

By constructing secondary VDP for each single vertex of the primary VDP and the respective surrounding atoms, the space that is assigned to this vertex/void can be determined. Thus, atoms can be identified that could theoretically occupy this space. If neighbouring voids can host the same atom, there will be a VDP edge connecting them. By analysing the space between the constituting atoms, the edge can be evaluated as a potential diffusion path between these two voids. Also, a tertiary VDP can be used in the channel to determine the passage size of a migrating ion, allowing a more generalised description of migration space for mobile ions.

84

4 Battery Materials

The information-analytical system “ToposPro” is used for constructing VDP. It is a multifunctional tool for crystallographic analysis, which enables the application of VDP analysis and special topological methods for crystal structures and whole crystallographic databases through built-in database-handling routines [54, 71]. ToposPro is designed to process long lists of crystal structures in batch mode, perform statistical analysis of the structural descriptors, and find relations between crystal structures of different chemical composition and complexity. It also allows the user to compute a number of topological parameters of crystal structures and store the values of various geometrical and topological characteristics, i.e. structure descriptors, in knowledge databases. For the analysis of ionic conductivity, the most important values are rSD, G3, and rChan, as well as the solid angles of the void- and edge-constituting atoms [14, 70]. rSD is the radius of a spherical domain that has the same volume as the considered VDP, which is a measure of the atomic size in the structure. G3 is the second moment of inertia of the VDP, which is a measure for the sphericity of the VDP; the smaller this value, the more spherical the VDP [65]. rChan is the channel radius and describes the width of the bottleneck between two voids. Additional special software enables to find the correlations between these descriptors and other structural properties, including physical properties using the statistical methods of regression and multifactor analysis. It also includes algorithms for the heuristic analysis and modelling of physical properties of substances (ionic conductivity, volatility) and allows for searching for structural relationships at different levels of the crystal organisation. ToposPro5 is an integrated interactive software environment functioning in the Windows operating system. The algorithms used in ToposPro are copyrighted by the Samara Centre for Theoretical Materials Science (SCTMS) and based on the theoretical models of graph and polyhedral representations of atoms, molecules, and compounds. As described in [27], intercalation as a topochemical reaction can be easily described as the addition of an ion on an empty or partially occupied crystallographic site. The host structure stays more or less unchanged during insertion and removal of the intercalated species. For the whole crystal to accommodate a maximum number of ions, diffusion between these sites needs to be possible. In this way, crystalline solid ionic conductors can be described similarly to intercalation hosts. This idea is the basis of the application of the Voronoi-Dirichlet approach to describe and find new solid electrolytes [70]; it has been applied for Li-ion as well as Na-ion conducting oxides [14, 19, 27]. Based solely on geometry, it offers a very fast algorithm and allows the analysis of thousands of compounds within an hour

5 ToposPro is available for free at http://topospro.

85

4.1 Computational analysis and identification of battery materials

regarding possible voids, there sizes and connections, and the resulting topology of the potential conduction network. The most important factor for both voids and channels is their significance, which is connected to two conditions [70]: 1. Determination: In the case of cationic conduction, both voids and channels need to be determined only by anions. If one of the constituting atoms is a cation, Coulomb repulsion will give an energy barrier too large for another cation to enter this site or channel. 2. Comparison: The geometrical data of the voids and channels is compared to a set of expected values. That is, rSD and G3 need to be at least as large as/smaller than the data-mined values for the mobile ion in question and its environment, respectively. For the channel radius, data-mined average distances need to be considered. All the important steps for this methodology, including future testing of the newly identified materials, are summarised in Figure 4.1.5.

5

1 Na1.6AI7O11 P63/mmc

# 03

2

3

4 6

7

Figure 4.1.5: Graphical summary of all necessary steps for the Voronoi-Dirichlet approach. The starting point is the use of a database (1) for data mining (2). The Voronoi-Dirichlet partitioning (3) is then applied and the data-mined values (4) are applied to determine potential ion conductors/intercalation hosts (5). Experimental work (6) should finally clarify if the material qualifies for electrochemical storage technology (7). Figure reproduced with permission from [F. Meutzner, W. Münchgesang, N. A. Kabanova, M. Zschornak, T. Leisegang, V. A. Blatov, and D. C. Meyer: ‘On the Way to New Possible NaIon Conductors: The Voronoi–Dirichlet Approach, Data Mining and Symmetry Considerations in Ternary Na Oxides’, Chemistry – A European Journal 21, 16601–16608 (2015)], © 2015 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim.

86

4 Battery Materials

The disadvantages of VDP analysis are, similar to BV methods discussed in the next section, that no dynamic process can be analysed directly, since VDP deals only with static crystal structures. A second disadvantage is that the VDP analysis is based on the assumption of completely ionic bonds. This may be a permissible simplification for oxygen-containing materials, due to the high electronegativity of O. Lesselectronegative elements are, however, more easily polarisable, forming bonds of a higher covalency that tend to shield charges. Due to the geometrical fundament of this methodology, very complex conduction patterns may be generated, especially in low-symmetry compounds that are difficult to interpret. The same holds for large intra-structural voids that the algorithm may tessellate into a multitude of vertices very close to each other. Some algorithms have been added to the programme in the recent years to allow, for instance, void-merging and the calculation of VDP in the channels between voids. Generally speaking, even though the geometry can be seen as an expression of the energetic relations in the crystal structure, energetic calculations offer a deeper understanding than sole geometry. For instance, since there is a preferential bond length between atoms (depending on the chemical surrounding), a void becomes less favourable if it is too large or too small. The most important feature of a good ion-diffusing material is its low-deviation energetic landscape for the pathway of the diffusing ion. This energetic landscape is not directly accessible by evaluating geometrical volume sizes in crystals. Further developments of the VDP methodology could therefore comprise the calculation of energetic parameters and an optimisation towards the applicability on the cation-conduction at hand.

4.1.2.3 Example 1: Li-ion conductors In a first work, it was shown that the VDP method is an excellent tool to analyse crystalline solid electrolytes [70]. It reveals the ionic conduction channels of wellknown cationic conductors, their dimensionalities, and allows a differentiation between so-called significant and probabilistic voids and channels. The latter may describe channels that could be activated at higher energies (temperatures). These works were expanded by screening the ICSD for compounds containing at least both Li and O and analysing all void-networks for their possible Li-ion conduction [14]. These materials were both ternary and quaternary compounds with the third and fourth element being any other chemical element. Furthermore, compounds with disorder on the O-sites were excluded yielding databases for ternary and quaternary structures with 822 and 1,349 entries, respectively. All voids’ and channels’ sites and sizes within were identified and evaluated according to the principles introduced above. The topologies of these networks were furthermore analysed with ToposPro. Within the results, a total of 26 compounds have been predicted as suitable Li-ion conductors, while 126 structure types were found, already described in the literature as solid electrolytes (Table 4.1.1).

4.1 Computational analysis and identification of battery materials

87

Table 4.1.1: Substances with possible Li-ion conductivity and their ICSD collection code [14]. “Dim” stands for dimensionality of the conduction network. Formula

ICSD-# Dim Formula

Li3AuO3 Li2Al2Si4O12 Li3BiO3 Li5BiO5 LiBUO5 Li3CuO3 Li3CuSbO5 Li3Er(NO3)5(NO3) LiFe(SeO3)2

15113 98845 85072 203031 67114 4201 51392 401554 75554

1D 2D 1D 1D 1D 1D 1D 1D 3D

Li0.33MoO3 Li2Mo4O13 (HT) Li2Mo4O13 (LT) Li4Mo5O17 LiReO4 Li5SbO5 Li10Si2PbO10 Li2TeO3 Li2TeO4

ICSD-# Dim Formula 201959 4155 6134 85439 37118 203030 78326 4317 1485

1D 1D 3D 1D 1D 1D 3D 1D 3D

ICSD-# Dim

Li4TeO5 2403 1D α-Li2Te2O5 26451 1D β-Li2Te2O5 26452 1D Li2UO4 (Pnma) 200297 1D Li2UO4 (Fmmm) 20508 3D Li4UO5 20452 1D Li6UO6 48209 3D Li0.88U3O8 69846 2D

4.1.2.4 Example 2: Na-ion conductors For the long-term, large-scale energy storage in the 100 MW range the Na-S battery technology accounts for the largest market share [72, 73]. A Na-S battery, invented in the 1960s [74], was enabled by the discovery of β-alumina, which shows a high conductivity for Na-ions at the operating temperature of 300 °C [75]. This material works as separator as well as solid electrolyte in the Na-S battery. Due to the high operation temperatures of high-temperature Na-S batteries and the intense reaction between Na and S at these temperatures, only ceramic, glass-ceramic, or glass electrolytes can be used [76–78]. New solid electrolyte materials are necessary for decreasing the operation temperature and thus to increase energy density and safety. In Refs. [27] and [19], it was shown that VDP analysis is a suitable and highly promising approach for the identification of novel (Na+) ion conducting materials. It was firstly shown that by using the Voronoi-Dirichlet approach, well-known Na-ion conductors could be clearly identified and, moreover, new promising candidates have been determined. A data mining was first carried out to find the most recurring geometrical values within compounds containing at least Na and O. As described in Ref. [65], VDP analysis works only for same-anion-neighbourhoods due to the problem of the division coefficient. Hence, we decided to search within O-containing compounds, since the bestknown Na-ion conductors are oxides. All compounds comprising at least these two chemical elements were filtered from the ICSD 16/1 first [79]: 11,004 out of 183,804 entries. After all duplicates6 were deleted (9,152 remaining), all compounds whose chemical formula did not match the composition of the atomic positions7 were deleted

6 Duplicates share the same space group, number of atoms, composition, and unit cell volume. 7 This is a comparison between the recorded chemical composition of the cif-file and the composition calculated from the occupancies and multiplicities of the atomic sites (of each occupied crystallographic orbit).

88

4 Battery Materials

92

166

164

1,360

150

substitutional disorder

144

3,500

5,541

ICSD 183,804

2,041

1,550

data mining

1,915

Li /H co

du pl ic at es fo ≠ rm co ula m po si tio n

1,852

3,611

172,800

er ith or O e n an N

nt ai ni Ω( ng vo id -M ch e) > an ne 5% rS lw D (v oi ith d) M 0D 0.2 is normally taken as an indication that there is something wrong in the crystal structure determination. For plausible structures the migration energy Emig and the pathway dimensionality dpath, as well as a rough range of expected room temperature ionic conductivities σRT for this structure are suggested. For undoped materials, where defect formation energies have to be considered, the order of magnitude of the Na+ ion conductivity is expected to be near the lower end of this range or even much lower, while a suitable doping might push the conductivity of this structure type up to the upper limit of this range. For more information on the calculation of the conductivities we refer to Ref. [111].

98 4 Battery Materials

108387

191947

Na7.85Al7.85Si8.15O32

Na2.62V2(PO4)2(O1.6F1.4)

160820 16899 77121 83012 92968

0.138 0.133 0.151 0.136 0.171

0.304

0.781

0.186

0.096

0.091

1D 3D 1D 3D (1D)

0.527 0.577 0.286 0.489 0.263

3D 3D 2D 3D 2D

0.188 3D

0.051 0.678 0.063 0.562 (0.461)

0.051 1D

−5.7 to −8.5b −6.2 to −9.4 −3.5 to −4.9 −5.3 to −7.9b −3.3 to −4.6c Determined at 300 °C High-pressure synthesis

Structure model wrong (very high GII); described as Na insertion compound High GII, all BV sums low. (early Rietveld powder structure) Determined at 500 °C

– (−2.8 to −3.7)

[128]

Determined at 600 °C; average structure

Superionicd

[131] [132] [133] [120] [121]

[130]

[129]

[128]

[127]

Non-conductive Proposed as Na+ ion conductor, but no conductivity data. As for other hollandite-type structures the high, strictly 1D Na+ ion mobility can only be utilised in nanocrystals. d Superionic Determined at 100 °C; average structure

Bond valence sum test shows that Al and Si have been mixed up in this structure refinement. After exchanging the sites the GII drops to the plausible range. However, this stable structure is found to be not a good conductor. b If this high-temperature phase can be retained at room temperature. c With BV parameters for Os6+ instead of Os5.5+. d To be ascertained by a local structure model.

a

Na2(Al2Si3O10) Na2TiOSi4O10 Na2Mg3Zn2(Si12O30) Na1.92(Al2Si3O10) Na1.78(Mg1.87Al0.13)(Si2O7)

47101

108334

Na7.85Al7.85Si8.15O32

Na5(MnO4)

79502

Na1.72(Cr1.71Ti6.29)O16

4.1 Computational analysis and identification of battery materials

99

100

4 Battery Materials

3D percolation A

1D percolation

s3

s2

EBVSE (eV)

0.38 s1

0.34

0.10 eV 0.15 eV

0.30

Na5/Na6

Na5/Na6

Na5/Na6

0.26 Na4

0.22 0

2

4 8 6 Migration pathway (A)

10

Figure 4.1.13: BVSE model of Na+ migration pathways in Na5YSi4O12 from the crystal structure model in ICSD-# 20271 shown as isosurfaces of constant BVSE projected on the a-b plane (left-hand side) and as energy-landscape diagram (right-hand side). Here, Na+ BVSE are specified relative to the energy of the lowest-energy Na site, the immobile Na3. The pathways of lowest migration barrier ≈0.10 eV among the partially occupied Na5/Na6 sites are predicted to be the same channels along c-direction identified by VDP (shown in orange). While such 1D channels alone would be highly vulnerable to blocking by point defects, the BVSE model shows that they are interconnected by perpendicular pathways Na5/Na6 to Na4 with an only slightly higher migration barrier of 0.15 eV explaining the experimentally observed fast ionic conductivity.

4.1.4 Density functional modelling and the materials project 4.1.4.1 Introduction The methods VDP and BVSE presented in the preceding sections to assess battery materials are based on the analysis of static crystal structures. However, ionic movement in a crystal is in general an interconnected dynamic process of atomic as well as electronic movement up to the collective migration of whole atomic units. Thus, the static energy barrier approximation of BVSE can be a rough estimation, only. Compared to other ab initio methods, DFT offers a fast quantum mechanical methodology to include the relaxational degrees of freedom and to model any crystalline material ab initio providing next to the migration barriers also the full electronic structure and an access to electronic conductivity parameters. The electronic structure determines most material properties of a crystal. Electrons are of quantum nature and in the quantum-dynamical picture, according to Heisenberg’s uncertainty relation, position and momentum of an electron can be simultaneously determined only up to a limit Δx·Δp ≥ ħ/2. In an eigenstate of certain quantum numbers, its position in space can therefore be described with a probability density. Quantum particles, even if they have a mass, show interference with each other like waves. They are indistinguishable from another. This realisation is the foundation of DFT. The general equation that describes a quantum mechanical system of electrons and nuclei in the non-relativistic case is the Schrödinger equation. In its stationary form with no explicit time dependence on the potential, it represents the energy

4.1 Computational analysis and identification of battery materials

101

^ relation as an eigenwert equation with energies Ei for the respective Hamiltonian H. With the many-body wavefunctions jΨi i $ Ψi (x1, x2, ..., xN, R1, R2, ..., RM) in dependence of 3(N + M) spatial coordinates x and R of the N electrons and M nuclei and N ^ jΨi i = Ei jΨi i. Since the nuclei are electronic spin states (x = {r,s}) as solutions it reads H several orders of magnitude heavier than the electrons, the Born-Oppenheimer approximation can be applied, solving the electronic terms in the potential of fixed nuclei positions (for details see e.g. [134]). Still, it requires immense computational efforts to take into account all distinct electron position vectors x1, as is done e.g. in Hartree-Fock calculations, and only systems very limited in size can be modelled. This limitation is overcome by the idea of a general electron density, as has been already used in the Thomas-Fermi model [135, 136] (1927) treating all electrons as a uniform electron gas or later by Slater [137] (1951) giving a local density expression for the Fermi hole and thus simplifying the non-local Hartree-Fock exchange. But the revolution of atomistic structure calculations on the quantum level took place only after the theorems of Hohenberg and Kohn [138] (1964), which prove that “the full many-particle ground state is a unique functional of ρ(r)” and that only the true ground state density ρ(r) yields the lowest total energy. These proofs gave rise to base quantum calculations on the presumption of the ground state electron density, which uniquely determines the system, the Hamiltonian as well as all other physical and chemical properties (e.g. band gap, electronic conductivity, elasticity, optical properties, etc.). So there is a unique ground state density ρ(r) and a unique functional F[ρ(r)] for all interactions, but unlike in Hartree-Fock not all energy expressions are known. Thus, in principle the electron density will be exact after total energy minimisation, but only if the Hamiltonian is correct. Being able to not only obtain a variety of physical properties of crystals from DFT but also to model the unique electron density ρ(r) of large structures from first principles and to directly correlate the results with experimentally determined electron densities, e.g. by X-ray diffraction, makes the method particularly attractive. In the authors’ view this access has caused the major breakthrough of the method in recent years. An approach to self-consistently solve the electronic Schrödinger equation and to minimise the error of the energy expressions within DFT, in particular of the kinetic term, was given by Kohn and Sham [139] (1965) introducing the concept of a noninteracting reference system with Kohn-Sham single particle orbitals ϕi ðr, sÞ, whose electron density equals the ground state density ρ(r) of the interacting system. The Kohn-Sham orbitals are kept orthogonal as solutions to the single particle equation. The functional of the total energy: 2 N N X N ðð  2 1  1X 1X  2 hϕi j∇ jϕi i + dr1 dr2 ϕi ðr1 Þ E½ ρðrÞ = – ϕ j ð r 2 Þ  2 i=1 2 i=1 j=1 r12 N ð M X X 2 ZK  – ϕi ðr1 Þ + EXC ½ ρðrÞ dr1 r K 1K i=1

102

4 Battery Materials

to be minimised includes the exact kinetic energy of the non-interacting reference system, the electron-electron coulomb interaction (distances r12), the stabilising electron-nucleus attraction (nucleus charge Z, distances r1K) and the exchange and correlation energy EXC containing besides exchange and correlation also the selfinteraction correction and the shortage in kinetic energy. The theory is exact up to this last functional which remains a challenge in today’s DFT development with more and more sophisticated expressions and approximations tackling higher and higher levels of accuracy, e.g. in band gap description, effective mass prediction, magnetic coupling, van der Waals interactions, atomisation, ionisation and reaction energies, and structure prediction, just to mention a few. Due to the decreased complexity by utilising a one-electron density ρ(r), it is nowadays possible to model systems of many hundreds of atoms from first principles calculations. These methods transfer material science to a next quality stage in general and in the field of Li-ion batteries in particular, as was already indicated in the introduction. Critical battery properties are accessible through these methods as well, i.e. ionic conductivity, phase stability with intercalation, and influence of defects and dopants [140]. In this respect, migration barriers can be calculated, e.g. by means of NEB methods, with the advantage that there are no empirical parameter reliability issues. A comprehensive example can be found in [140]. The theoretical techniques have reached such maturity that they have recently started to be used to perform high-throughput computational search of materials [141]. Computationally even more demanding ab initio MD is capable to reliable include temperature effects into absolute conductivity parameters. By computing properties on large databases of thousands of potential electrode materials, researchers can identify the most promising compounds to be targeted by follow-up experimental work. To date, DFT calculations are regularly performed in the optimisation and design of new materials, with demonstrations in a number of research fields, such as energy storage [142, 143], catalysis [144], energy conversion [145], pharmaceutical [146], and metallurgy [147]. The availability of large high-performance supercomputer infrastructures coupled with DFT, pre-/post-processing libraries [30, 31, 148, 149] and databases [29, 141, 150] provide the ingredients to compute thousands of compounds, augmenting the predicting capabilities of computational chemistry and materials science to search for new high-potential battery materials with improved energy density (both per mass and per volume) [151–156]. One strategy is to look for materials that can provide high cell voltages, but still lie within the practical stability windows of commercial electrolytes. The marriage of computational materials science with computer informatics is often referred to as high-throughput (HT) calculations. Figure 4.1.14 shows an example of workflow implemented in the Materials Project initiative [141] which uses the python materials genomics (pymatgen) [148] to process the input files for calculations, post-process the output files, and analyse the results.

103

4.1 Computational analysis and identification of battery materials

a)

i) compound Design PREDICTIONS

ideas

PARAMETERS

potential materials

promising, stable materials

AB2X3 DATA

AB2X3

promising materials

80% AB + BX3 AB2X3

iii) Stability & Synthesis

ii) Property Exploration

b) Structure or Data Input

Analyses / Results

VASP input /output io.Vaspio, borg CIF

Serializers

Python objects - Structure / Molecule - ComputedEntry -DOS -Bandstructure

Serialized pymatgen objetcts matproj_rest Materials Projects REST API

Phasediagram

Reaction calculator

analysis. reaction_calculator

Electronic Structure analysis and visualization

electronic_ structure

Applicatiospecific. e.g., battery properties

apps

Structure visualizations

vis

io.cifio

io.babelio OpenBabel formats

Phase diagrams

Structure / Data manipulations - transformations - alchemy -entries.compatibilty (GGA/GGA+U mixing)

Other analyses, e.g., Ewald summations, structuresymmetry and slimilarity, etc.

analysis, symmetry

Figure 4.1.14: (a) Schematic of the workflow employed for performing high-throughput calculations and the resultant analysis in Materials Project (taken from [APL Materials 1, 011002 (2013)]; used in accordance with the Creative Commons Attribution (CC BY) license. (b) Structure of the python materials genomics (pymatgen) software which is used to power the Materials Project database taken from Ref. [148], (10.1016/j.commatsci.2012.10.028).

104

4 Battery Materials

4.1.4.2 The materials project The Materials Project aims to remove guesswork from materials design in a variety of applications by computing properties of all known materials [3]. Since 2011 a collaboration of leading researchers participates in this project in order to accelerate the computer-driven materials revolution [1]. The goal of this collaboration is to build free, open-access databases containing the fundamental thermodynamic and electronic properties of all known inorganic compounds. To date, basic properties of nearly all of the approximately 35,000 crystalline inorganic materials known to exist in nature were calculated as well as the properties of another few thousand that exist only in theory. Finally, with high-throughput DFT methods so far more than 67,000 inorganic compounds have been collected from which more than 3,600 are intercalation electrode materials and more than 16,100 are conversion type electrode materials for lithium-ion batteries. Different parameters can be accessed: structural (e.g. lattice parameters), electronic (e.g. voltage), and mechanical (e.g. elastic tensor). So far, more than 5,000 scientists have registered for access to the database containing this information. Thus, the experimental research can be targeted to the most promising compounds from computational data sets. It enables researchers to work out scientific trends in materials properties by data mining procedures. In the following paragraphs, we discuss some of the properties accessible via high-throughput calculations in the field of materials for energy storage applications to go beyond experimental improvement of voltages in materials through modifying the materials composition, utilising well established rules of thumb, the so-called “chemical intuition”. 4.1.4.3 Open-circuit cell voltage The open-circuit intercalation (conversion) voltage is set by the difference in the chemical potential, ΔμX, of the intercalation species (e.g. X = Li, Na, or Mg) in anode (e.g. Li metal in Li-ion batteries) and cathode material (e.g. LiCoO2):

V= –

=– 1 V zF

½μcathode – μanode  X X , zF

Zn2 ½μcathode – μanode dnX = – X X n1

(4:1:2)

ΔGreaction , ðn2 – n1 ÞzF

(4:1:3)

where F is the Faraday constant and z is the electron transfer number. Since the cell device operates at constant temperature and pressure, the change of Gibbs free energy is solely set by the change the of X content ΔnX between the anode and cathode weighted by the respective μX [157]. Therefore, integrating eq. (4.1.2) over the intercalation reaction (e.g., Lin1 CoO2 + ðn2 – n1 ÞLianode ! Lin2 CoO2 ) furnishes the

4.1 Computational analysis and identification of battery materials

105

open-circuit voltage as function of the Gibbs free energy change of the redox reaction for delivering n2 – n1 equivalents of the intercalating species X. Note that eq. (4.1.3) is the basis of the well-known Nernst equation. We note that DFT, in its various exchange-correlation functional flavours [139, 158–160] is limited only to internal ground state energies (and not free energies). Important thermal effects, such as configurational and vibrational entropy, are not captured. In general, at room temperature (and below), entropic and thermal contributions are negligible for many crystalline materials with high melting point, thus enabling us to approximate ΔGreaction by ΔEreaction . In DFT, the approximation of the exchange-correlation functional used can significantly vary redox reaction energetics and the predictability of open-circuit voltages. For example, the local density approximation (LDA) and the generalised gradient approximation (GGA) of the exchange-correlation fail to predict the intercalation voltages in highly correlated transition metal oxides due to spurious self-interaction errors [139, 158]. The self-interaction – the interaction of each electron with itself – underestimates the total energy of the reduced transition metal oxide, resulting in an underestimation of the intercalation voltage. In cathode materials with highly localised d or f electrons, such as LiFePO4 and LiMn2O4, the self-interaction error becomes particularly important. Two strategies commonly used to improve the voltage prediction within DFT are adding an energy penalty in the form of a Hubbard U for specific orbital occupations (formally referred to as DFT + U) [160, 161], or adding a fraction of the exact Hartree-Fock exchange to the GGA exchange-correlation, as implemented in hybrid functionals [159, 162–166]. Since hybrid functionals are significantly more computationally expensive (at least, in the pseudopotential-planewave formalism) the DFT + U approach is commonly adopted in HT calculations. Previous studies have shown significant accuracy in DFT + U voltage predictions with respect to experimental observations [143]. A graphical summary of voltage calculations for battery materials is presented in Figure 4.1.15.

4.1.4.4 Dynamics of ions in energy storage materials Apart from identifying cathode materials with high-energy densities, it remains equally important to identify cathodes and solid electrolytes with good intrinsic ionic transport. This is ultimately crucial in determining the power density of a given battery infrastructure [168, 169]. Particularly in the case of high-valent batteries, ionic mobility has been a major impediment in developing a wide range of cathode materials, as is available in Li-ion and Na-ion systems [170–172]. Typically, chemical intuition would suggest that ionic motion within “dense” or “closepacked” frameworks will be significantly poorer compared to “light” or “open” frameworks [173], but the most important factor for ionic diffusion is the chemical environment of the mobile ion and its evaluation during the migration process:

106

4 Battery Materials

(a)

(b) 3.5

6

3.0

4

800 Wh/kg

3

Voltage (V)

Voltage (V)

5

2.5 1/2

2.0

5/8 2/3 3/4

1.5

7/8

1.0

600 Wh/kg

0.5 2

100

0.0 0.0

200

Capacity (mAh/g)

0.2

(c) 1 electron transfer

0.4 0.6 X in NaXVO2

1.0

2 electron transfer Intercalation Conversion

Li

Mg

Mg

Ca

Ca

Zn

Zn

0.0

0.8

0.5

1.0

1.5

2.0

Voltage (V)

2.5

3.0

3.5

4.0

-2

-1

0

1

2

3

4

Voltage (V)

Figure 4.1.15: Examples of voltage calculations that have been applied to study the screening of several (a) Li intercalation in poly-anion materials as taken from Figure 4 of Ref. [156], (b) the benchmark of computed voltage curves against experiments for Na intercalation in VO2 (Reprinted figure with permission from [167] [A. J. Toumar, S. P. Ong, W. Davidson Richards, S. Dacek, and G. Ceder, Physical Review Applied 4, 064002 (2015).] Copyright 2015 by the American Physical Society. http://dx.doi.org/10.1103/ PhysRevApplied.4.064002), and (c) the competition of intercalation and conversion reactions in Li and multi-valent cathode materials (Reprinted from Chemical Reviews, 117, P. Canepa, G. S. Gautam, D. C. Hannah, R. Malik, M. Liu, K. G. Gallagher, K. A. Persson, and G. Ceder, Odyssey of Multivalent Cathode Materials: Open Questions and Future Challenges, 4287–4341, Copyright 2017, with permission from Elsevier, http://pubs.acs.org/doi/abs/10.1021/acs.chemrev.6b00614; further permissions related to the figure should be directed to the ACS).

even chemically “unfavourable” paths may show a flat energetic profile, if the environment stays “unfavourable” during the whole conduction path [174, 175], highlighting the non-intuitive behaviour of ionic motion and the importance of computations as a predictive tool for screening good candidates.

107

4.1 Computational analysis and identification of battery materials

4.1.4.5 Ionic mobility Ionic migration in periodic solids can be treated within the framework of transition state theory [176], a popular theoretical framework employed in the study of kinetics of chemical reactions (see Figure 4.1.16). Note that ions in solids will need to move via symmetrically equivalent sites, typically limited by a transition state with an energetic barrier, to ensure macroscopic diffusion. Given an energy barrier for migration (Em), the diffusivity of a given ion in a solid is written in an Arrhenius equation (eq. (4.1.4)) as:  D = f a2 v exp

a)

–

 Em , kT

Transition

(4:1:4)

b)

i–1

Fi

Initial

i

⊥

Fi Energy (eV)

NEB Fi

NEB

Fi

Em MEP Initial

τi

SII Fi

Final

i+1

NEB

Final

Path Distance (%)

Figure 4.1.16: (a) Typical energy profile during ionic migration in a solid. The energy difference between the initial (and final) state and the transition state gives the energy barrier (Em) for ionic migration. (b) Schematic of the force projection scheme used during the nudged elastic band minimisation scheme to find a minimum energy pathway (MEP) [179] (Reprinted from [D. Sheppard, R. Terrel, and G. Henkelman: “Optimization methods for finding minimum energy paths”, The Journal of Chemical Physics 128, 134106 (2008).], with the permission of AIP Publishing).

where f, a, v, k, t indicate the correlation factor, distance between symmetrically equivalent sites, vibrational frequency, Boltzmann constant, and temperature, respectively [177]. Note that (Em) is the most important intrinsic quantity that affects ionic diffusivity in a given solid and needs to be accurately determined, theoretically or experimentally. Also, the magnitude of (Em) is dependent on the (accurately determined) energy difference between the transition state and the stable initial (or final) state (Figure 4.1.16). The first step in calculating the energy of the transition state is accurately determining the transition state. Within the realm of ionic migration in solids, the transition state corresponds to a specific geometric position along the migration pathway, which is the saddle-point in the potential energy surface, i.e., the transition state is at an energymaximum along the migration pathway but is also at an energy-minimum in comparison to other possible transition states (see Figure 4.1.16). While “slowest ascent” and “drag”

108

4 Battery Materials

methods have been used to determine transition states in chemical reactions before [32], the NEB method [179] has emerged as the most reliable theoretical framework to obtain precise transition states. The NEB method involves creating a “chain-of-states” or a discrete set of “images”, often referred to as a “band”, along the migration pathway and subsequently relaxing the band using a force projection scheme (Figure 4.1.16) [179]. Apart from accurately determining the transition state, the force projection in NEB facilitates the calculation of the energies along the images used to represent the migration pathway, which is useful to understand the migration mechanism. While NEB is used to identify the transition state, DFT is nominally used to sample the potential energy surface (and related forces) within which the NEB is employed. Previous experimental studies have confirmed that DFT-based NEB can predict migration barriers with great accuracy, as illustrated in a variety of Li-ion, Naion, and Mg-ion conductors (and shown in Figure 4.1.17) [175, 180–186] and in a few Li-ion cathodes (such as LiFePO4) [187]. Notably, in cathode systems which contain d or f transition metals with highly correlated electrons, DFT + U is normally used in voltage calculations to reduce selfinteraction errors (as described above). However, previous studies have reported convergence difficulties in NEB calculations based on DFT + U, attributed to the metastability of the electronic states along the migration path [188–190]. During ionic migration in cathodes, the corresponding electron typically migrates across transition metal atoms, which is treated in adiabatic fashion within DFT + U due to the metastability of electron occupation, leading to poor convergence. In general, the barriers for electronic migration are lower compared to ionic migration in electrodes [162, 191], suggesting that the ionic migration is the rate-limiting step in cathodes, especially in high-valent systems [172, 188]. Also, previous studies that have indeed converged NEB calculations within DFT + U have reported insignificant differences in the barriers predicted versus DFT [192]. Recent experimental studies that qualitatively agree with DFT-NEB migration barrier predictions, especially in high-valent systems [189, 193–195], further validate the reliability of the DFT-NEB method. Given the complexity of combining NEB and DFT, high-throughput studies that calculate migration barriers over hundreds of compounds are yet non-existent. Prior computational studies have relied on the identification of computationally inexpensive metrics or parameters that may be used to screen for fast ionic diffusers [144, 174]. Significant efforts have also been made in “accelerating” the identification of the transition state via better construction of the initial set of images used in the NEB [196], leading to the identification of a new class of potential high-valent cathodes [197]. Constructing a theoretical model, based on the NEB or otherwise, that allows for high-throughput calculations of migration barriers will eventually aid not only the battery community in the identification of new cathodes and ionic conductors but also in allied fields such as catalysis, fuel cells, and semi-conductors [144, 198]. If an ionic conduction model shall be investigated in dependence on temperature for a bulk material, the degrees of freedom of the electronic density distribution may

4.1 Computational analysis and identification of battery materials

109

(a) 600 DFT migration barrier (meV)

Li9S3N 500 400

MgSc2S4 Na10SnP2S12

300 Li10SiP2S12

200

Li10GeP2S12

100 0

Li10SiP2S12 LLRZO

0

100 200 300 400 500 Experimental migration barrier (meV)

600

(b)

400 300 200 100 NaV2O4 (Discharged) NaV2O4 (Charged)

0 –100

Relative Energy (meV)

Relative Energy (meV)

500

300 250 200 150 100 50 0 –50

MgV2O4 (Charged) MgV2O4 (Discharged)

(c) Start

End

Pathfinder relaxation

MEP c b a

Li Fe P O

Initial linear path

Iterative linear optimization

10–1

10–2

Static diffusion potential (nondimensional)

Converged path

100

10–3

Figure 4.1.17: (a) Migration barriers calculated using DFT for various Li-ion, Na-ion, and Mg-ion conductors are plotted against corresponding barriers reported experimentally. LLRZO indicates Li7.08La2.96Rb0.04Zr2O12. Dashed black line indicates parity. (b) Sample migration barrier calculations reported in NaV2O4 (green) and MgV2O4 (blue) post-spinel structures taken from Figure 3 of Ref. [193]. (Published by The Royal Society of Chemistry), which have also been investigated experimentally [194]. (c) Schematic of an electronic charge density based scheme to “speed up” the identification of the transition state using NEB, as demonstrated in the LiFePO4 cathode [196].

110

4 Battery Materials

be reduced to the utilisation of two- or three-particle potentials. This is realised by MD modelling which is described in the following.

4.1.5 Molecular dynamics In MD, a trajectory is computed by iterative calculations of velocities and positions of a system in the future, one time increment apart from their values at a given time frame (Figure 4.1.18). In principle, a trajectory is set to visit any point in configuration space. In practice, large and possibly important regions thereof may remain unexplored within a typical, finite time MD simulation. While highly probable (metastable) regions are easily accessible, low probable regions may remain partially or even completely unsampled. Reasoning in terms of Gibbs free energy, if two metastable regions (state A and state B) are separated by high activation energy barriers (ΔG# >> kBT), a trajectory will mostly linger in either A or B and only occasionally visit any intermediate region. Depending on the severity of the activation energy barrier, said regions may not be explored at all. This means that any activated process may pose a challenge to MD simulations, making it intrinsically difficult to explore microscopic mechanisms of any process containing free energy barriers.

r

ρ(pt,rt)

ρ(p0,r0) t0

t Δt

p

Figure 4.1.18: Trajectory (orange dashed line) computed in molecular dynamics simulation. Positions (r) and momenta (p) are propagated in finite time-steps (Δt).

Correctly describing ionic migration pathways within battery materials belongs to this class of problems. Different from purely diffusive processes, a cation-hopping from one site to an adjacent one implies overcoming internal free energy barriers. As pointed out in the last section, approaches based on total energy calculations as DFT or pair potentials [199, 200] can estimate energy barriers based on pathway-interpolations from site to site, followed by energy minimisation. Typical activation energies for the wide-spread battery material LiFePO4 lie in the range (0.27–0.55) eV ([3133–6382] K) [199, 200]. In standard MD simulations, a significant overdriving of simulation parameters, here temperature would be required to sporadically observe inter-site particle (cation) jumps.

4.1 Computational analysis and identification of battery materials

111

To address this problem related to the time scale of MD simulations, in recent years, different methods have focused on acceleration approaches. So-called collective variables (CV) were introduced [201, 202], as a means to selectively accelerate events, which would not occur in plain MD approaches. A CV represents a bundle of internal degrees of freedom of a system [203]. During an accelerated MD simulation, a CV markedly changes its value as a function of the progress of the simulation. In the example above, CV would have distinguishable values in state A and state B, and the progress of a CV-driven MD simulation would be marked by a sharp change of the CV value on moving from A to B. This approach has been successful in an increasing number of chemical systems from materials science to biological systems. For batteries though, the need to describe the collective motion of mobile cations within relatively rigid framework scaffoldings, makes the selection of efficient CVs less intuitive. To efficiently collect details of cation jumping mechanisms, without overdriving temperature, the natural separation of frequencies between mobile, light cations and rigid, heavier framework atoms can be exploited. With the aim of elucidating pathways of cation (Li, Na) migration, our method [204] proceeds as follows: Instead of changing (overdriving) the overall temperature of the system, we “warm up” the mobile particles only by transferring a variable amount of kinetic energy from the slow (non-diffusive), heavier particles to the more mobile, lighter cations. For this purpose, velocity distributions are generated, which are not typical for the ensemble average, causing Li or Na cations to move quicker. As such, the chances of escaping local or global minima are increased. Along this line, the number of reactive events (jumps) is enhanced which results in shortened simulation times, lower nominal simulation temperature, and improved sampling efficiency [204]. A MD simulation accelerated this way can be implemented by introducing a perturbation of the velocities of each Li/Na cation at time t, by randomly choosing new values from a Gaussian distribution centred on the actual velocity of the target atom. Therein, the warming up effect is achieved by broadening the Gaussian halfwidth. Perturbations are introduced under strict conservation of total linear and angular momenta. The perturbed snapshot is then propagated in MD simulations and the velocities are perturbed again at a later stage, until a sufficient number of events are collected [204]. After an initial latency, the system enters a regime where jumps are observed with a constant frequency, which allows for the estimation of mean paths and importance of different types of cation translocation modes. Using this approach, a large statistic of jump events can be collected. In particular, collective events of cation translocation become accessible. This is different from approaches based on lattice energies, where single particle jumps are typically considered. To illustrate the method, we compare the dynamics of particle translocation (Li and Na) in the olivine battery materials LiFePO4 and NaFePO4. In Figure 4.1.19,

112

4 Battery Materials

splines are used to represent Li and Na displacements over a period of time of 500 ps. Correlated particle jumps are observed. The most common mechanistic pattern is represented by Li/Na migrations along the [010] channels, vertically in Figure 4.1.19. Episodically, jumps in the orthogonal direction [001], horizontally in Figure 4.1.19 are observed. This is an important result of considering collective cation dynamics, and is different from what can be predicted based on lattice energy calculations only. Indeed, [001] jumps where indicated elsewhere as “forbidden” [200], due to a too demanding activation energy barrier. However, thanks to collective movements, proximity of another charged particle can accelerate (i.e. catalyse) Li or Na jumps. Otherwise demanding barriers can be crossed this way.

[001]

[010]

Figure 4.1.19: Displacements of Li and Na cations in the olivine materials LiFePO4 (left) and NaFePO4 (right), respectively. The Li/Na motions are represented as splines over a period of 500 ps (T = 600 K). While events of cation translocation are more frequent along [010], the easy olivine axis, jumps along [001] are part of the overall mechanism.

This approach is extremely valuable towards collecting mechanisms of cation translocation within battery materials, both existing and predicted. In existing materials, MD results can indicate relevant pathways along which cations actually move. Combined with different techniques (Umbrella sampling, kinetic MC), rate constants and diffusion constants can be reliably calculated, subsequently. This provides a very solid basis for the prediction of cation translocation efficiency in predicted battery materials. Migration paths as predicted from the VDP and BVSE methodologies can be weighted according to the probability of cation migration along different pathways.

4.1 Computational analysis and identification of battery materials

113

4.1.6 Summary Crystallography is a powerful tool for the analysis of (crystalline) solid-state materials. Since the model of diffusion can be described very well by crystallographic means, the electrochemical processes in batteries and accumulators concerning the energy conversion from chemical to electrical energy can be described, as well. This way, new materials can be predicted with promising electrochemical functions. We presented a new methodology that seeks to optimise the time-efficiency of such a search by first utilising computationally less expensive approaches and applying increasingly complex calculations to these resulting promising materials. We have focussed on VoronoiDirichlet partitioning, a method analysing solely the geometry of crystal structure; BVSE calculations, combining the static crystal structure geometry with energetic parameters based on data mining; density functional theoretical simulations to evaluate the principal diffusion step(s) ab initio; and MD simulations to model the dynamics of the ion conduction. In this order, fewer crystal structures are calculated with more timeconsuming methodologies that allow gathering more in-depth information based on energetic parameters. Crystallographic databases like the ICSD, CSD, and the PCD represent a growing and increasingly reliable foundation for data mining, big-data approaches, and machine-learning. The more and the better the data collected within, the more promising functional materials may be identified, since the probability to find compounds with those crystallographic features necessary for high ionic diffusivity is increasing with the amount of reported crystal structures in crystallographic databases. In the very near future, materials scientists will use high-throughput computing together with big-data approaches, topological methods, and database screening for identifying and designing novel best-suited materials and complete battery systems. The authors of this contribution follow the view of G. Ceder [1], who already stated “that this will lead to technologies that will reshape our world – breakthroughs that will transform computing, eliminate pollution, generate abundant clean energy and improve our lives in ways that are hard to imagine today”. Acknowledgements: FM, TN, MZ, TL, and DCM are grateful for financial support of the Federal Ministry of Education and Research (CryPhysConcept (03EK3029A) and R2RBattery (03SF0542A)). PC is grateful to the Ramsey Memorial Trust for the provision of his Ramsey Fellowship. SL acknowledges support from the UK Research Council for using work in the paper that was undertaken by a student under Project No. EP/ M50631X/. SL also thanks ARCCA Cardiff for computational resources. SA would like to thank National Research Foundation, Prime Minister’s Office, Singapore for support under its Competitive Research Programme (NRF-CRP 10-2012-6) and NUS for support under the “NUS Centre for Energy Research” grant. VAB is grateful for financial support of the Russian Megagrant (14.B25.31.0005), and the Russian Science Foundation (16-13-10158).

114

4 Battery Materials

References [1] [2] [3] [4] [5] [6] [7] [8] [9]

[10] [11] [12] [13] [14]

[15] [16]

[17]

[18]

[19]

[20] [21] [22] [23]

Ceder G, Persson K. The Stuff of Dreams. Sci Am. 2013;309:36–40 Strategische weiterentwicklung des hoch- und höchstleistungsrechnens in Deutschland. Germany: Deutscher Wissenschaftsrat, 2012 Fueling discovery by sharing. Nat Mater. 2013;12:173. DOI: 10.1038/nmat3594. Material’s Project, https://materialsproject.org, cited 04.05.2017. ‘Partnership for Advanced Computing in Europe’, http://www.prace-ri.eu (cited 09. March 2017). Boosting materials modelling. Nat Mater. 2016;15:365. DOI: 10.1038/nmat4619. Jahan A, Ismail MY, Sapuan SM, Mustapha F. Material screening and choosing methods – a review. Mater Des. 2010;31:696–705. Mayr LM, Bojanic D. Novel trends in high-throughput screening. Curr Opin Pharmacol. 2009;9:580–588. Green ML, Choi CL, Hattrick-Simpers JR, Joshi AM, Takeuchi I, Barron SC, et al. Fulfilling the promise of the materials genome initiative with high-throughput experimental methodologies. Appl Phys Rev. 2017;4:011105. Levi E, Levi MD, Chasid O, Aurbach D. A review on the problems of the solid state ions diffusion in cathodes for rechargeable Mg batteries. J Electroceram. 2009;22:13–9. Elia GA, Marquardt K, Hoeppner K, Fantini S, Lin R, Knipping E, et al. An overview and future perspectives of aluminum batteries. Adv Mater. 2016;28:7564–79. Van Noorden R. The rechargeable revolution: a better battery. Nature. 2014;507:26–8. ‘Company overview of Pellion Technologies Inc.’, Bloomberg. Available at:http://www.bloomberg. com/research/stocks/private/snapshot. asp?privcapId=114339859. Accessed: 04 May 2017. Anurova NA, Blatov VA, Ilyushin GD, Blatova OA, Ivanov-Schitz AK, Dem’yanets LN. Migration maps of Li+ cations in oxygen-containing compounds. Solid State Ionics. 2008;179:2248–54. Hautier G, Fischer C, Ehrlacher V, Jain A, Ceder G. Data mined ionic substitutions for the discovery of new compounds. Inorg Chem. 2011;50:656–63. Avdeev M, Sale M, Adams S, Rao RP. Screening of the alkali-metal ion containing materials from the Inorganic Crystal Structure Database (ICSD) for high ionic conductivity pathways using the bond valence method. Solid State Ionics. 2012;225:43–6. Gao J, Chu G, He M, Zhang S, Xiao R, Li H, Chen L. Screening possible solid electrolytes by calculating the conduction pathways using bond valence method. Sci China Phys Mech Astron. 2014;57:1526–36. Xiao R, Li H, Chen L. High-throughput design and optimization of fast lithium ion conductors by the combination of bond-valence method and density functional theory. Sci Rep. 2015;5:14227. Meutzner F, Münchgesang W, Leisegang T, Schmid R, Zschornak M, Ureña De Vivanco M, Shevchenko AP, Blatov VA, Meyer DC. Identification of solid oxygen-containing Na-electrolytes: an assessment based on crystallographic and economic parameters. Crystal Res Technol. 2017;52:1600223. Wang Y, Richards WD, Ong SP, Miara LJ, Kim JC, Mo Y, Ceder G. Design principles for solid-state lithium superionic conductors. Nat Mater. 2015;14:1026–31. Wong LL, Chen H, Adams S. Design of fast ion conducting cathode materials for grid-scale sodium-ion batteries. Phys Chem Chem Phys. 2017;19:7506–23. Data mining, Merriam Webster: https://www.merriam-webster.com/dictionary/data% 20mining (cited 04.05.2017). Larose DT. Discovering knowledge in data: an introduction to data mining. Hoboken, NJ: Wiley & Sons Inc., 2014.

References

115

[24] Sendek AD, Yang Q, Cubuk ED, Duerloo K-A, Cui Y, Reed EJ. Holistic computational structure screening of more than 12000 candidates for solid lithium-ion conductor materials. Energy Environ Sci. 2017;10:306–20. [25] Manyika J, Chui M, Brown B, Bughin J, Dobbs R, Roxburgh C, et al. Big data: the next frontier for innovation, competition, and productivity. Brussels, San Francisco: McKinsey Global Institute, McKinsey & Company, 2011. [26] Qu X, Jain A, Rajput NN, Cheng L, Zhang Y, Ong SP, et al. The electrolyte genome project: a big data approach in battery materials discovery. Comput Mater Sci. 2015;103:56–67. [27] Meutzner F, Münchgesang W, Kabanova NA, Zschornak M, Leisegang T, Blatov VA, Meyer DC. On the way to new possible Na-ion conductors: the Voronoi–Dirichlet approach, data mining and symmetry considerations in ternary Na oxides. Chem – Eur J. 2015;21:16601–8. [28] Ghadbeigi L, Sparks TD, Harada JK, Lettiere BR. Data-mining approach for battery materials. 2015 IEEE Conference on Technologies for Sustainability (SusTech), 2015: 239–44. [29] Hachmann J, Olivares-Amaya R, Atahan-Evrenk S, Amador-Bedolla C, Sánchez-Carrera RS, GoldParker A, Vogt L, Brockway AM, Aspuru-Guzik A. The Harvard clean energy project: large-scale computational screening and design of organic photovoltaics on the world community grid. J Phys Chem Lett. 2011;2:2241–51. [30] Curtarolo S, Setyawan W, Wang S, Xue J, Yang K, Taylor RH, et al. AFLOWLIB. ORG: a distributed materials properties repository from high-throughput ab initio calculations. Comput Mater Sci. 2012;58:227–35. [31] Saal JE, Kirklin S, Aykol M, Meredig B, Wolverton C. Materials design and discovery with highthroughput density functional theory: the open quantum materials database (OQMD). Jom. 2013;65:1501–9. [32] Jónsson H, Mills G, Jacobsen KW. Nudged elastic band method for finding minimum energy paths of transitions. In: Berne BJ, Ciccotti G,, Coker DF, editor(s). Classical and quantum dynamics in condensed phase simulations. Singapore, New Jersey, London, Hong Kong: World Scientific, 1998:385–404. [33] Adams S, Swenson J. Predictability of ion transport properties from the structure of solid electrolytes. Ionics. 2004;10:317–26. [34] Adams S. Modelling ion conduction pathways by bond valence pseudopotential maps. Solid State Ionics. 2000;136:1351–61. [35] Ling S-G, Gao J, Xiao R-J, Chen L-Q. High-throughput theoretical design of lithium battery materials. Chin Phys B. 2016;25:018208. [36] Xiao R, Li H, Chen L. Candidate structures for inorganic lithium solid-state electrolytes identified by high-throughput bond-valence calculations. J Materiomics. 2015;1:325–32. [37] Adams S, Prasada Rao R. Transport pathways for mobile ions in disordered solids from the analysis of energy-scaled bond-valence mismatch landscapes. Phys Chem Chem Phys. 2009;11:3210–6. [38] Adams S, Prasada Rao R. Understanding ionic conduction and energy storage materials with bond-valence-based methods. In: Brown ID, Poeppelmeier KR, editor(s). Bond Valences. Berlin Heidelberg: Springer , 2014:129–59. [39] Sale M, Avdeev M. 3DBVSMAPPER: a program for automatically generating bond-valence sum landscapes. J Appl Crystallogr. 2012;45:1054–6. [40] Filsø MØ, Turner MJ, Gibbs GV, Adams S, Spackman MA, Iversen BB. Visualizing lithium-ion migration pathways in battery materials. Chem – Eur J. 2013;19:15535–44. [41] Filsø MØ, Eikeland E, Iversen BB. Procrystal analysis as a tool for the visualization of ion migration pathways. AIP Conference Proceedings 2016;1765:020010.

116

4 Battery Materials

[42] Adams S, Swenson J. Determining ionic conductivity from structural models of fast ionic conductors. Phys Rev Lett. 2000;84:4144. [43] Adams S, Swenson J. Bond valence analysis of transport pathways in RMC models of fast ion conducting glasses. Phys Chem Chem Phys. 2002;4:3179–84. [44] Adams S, Swenson J. Bond valence analysis of reverse Monte Carlo produced structural models; a way to understand ion conduction in glasses. J Phys Condens Matter. 2005; 17:S87–101. [45] Müller C, Zienicke E, Adams S, Habasaki J, Maass P. Comparison of ion sites and diffusion paths in glasses obtained by molecular dynamics simulations and bond valence analysis. Phys Rev B. 2007;75:014203. [46] Cai L, White RE. Mathematical modeling of a lithium ion battery with thermal effects in COMSOL Inc. Multiphysics (MP) software. J Power Sources. 2011;196:5985–9. [47] COMSOL Multiphysics® 3.5a USER’S Guide. 2008a. COMSOL Inc. [48] COMSOL Multiphysics® 3.5a Reference Guide. 2008b. COMSOL Inc. [49] Takada K, Ohno T. Experimental and computational approaches to interfacial resistance in solid-state batteries. Front Energy Res. 2016;4:10. [50] Meyer DC, Leisegang T, editor(s). Electrochemical storage materials: from crystallography to manufacturing technology. Oldenbourg: De Gruyter Publishing House, 2018 [51] Nishijima M, Ootani T, Kamimura Y, Sueki T, Esaki S, Murai S, et al. Accelerated discovery of cathode materials with prolonged cycle life for lithium-ion battery. Nat Commun. 2014;5:4553. [52] Fuess H, Hahn T, Wondratschek H, Müller U, Shmueli U, Prince E, et al., editors. International tables for crystallography. Berlin: Springer, 2004. [53] Batten SR, Robson R. Interpenetrating nets: ordered, periodic entanglement. Angew Chemie Int Ed. 1998;37:1460–94. [54] Blatov VA, Shevchenko AP, Proserpio DM. Applied topological analysis of crystal structures with the program package ToposPro. Cryst Growth Des. 2014;14:3576–86. [55] Blatov VA, Proserpio DM. Periodic-graph approaches in crystal structure prediction. In: Oganov AR, editor(s). Modern methods of crystal structure prediction. Weinheim: Wiley VCH, 2011:1–28. [56] Frank FT, Kasper J. Complex alloy structures regarded as sphere packings. I. Definitions and basic principles. Acta Crystallogr. 1958;11:184–90. [57] Frank FT, Kasper J. Complex alloy structures regarded as sphere packings. II. Analysis and classification of representative structures. Acta Crystallogr. 1959;12:483–99. [58] O’Keeffe M, Hyde BG. Crystal structures I. Patterns and symmetry. Washington, DC: Mineralogical Society of America, 1996. [59] Steurer W, Dshemuchadse J. Intermetallics: structures, properties, and statistics. Oxford: Oxford University Press, International Union of Crystallography Monographs on Crystallography, 2016 [60] Aurenhammer F. Voronoi diagrams – a survey of a fundamental geometric data structure. ACM Comput Surv. 1991;23:345–405. [61] Voronoi GF. Nouvelles applications des paramètres continus à la théorie des formes quadratiques. Deuxième mémoire. Recherches sur les parallélloèdres primitifs. J Reine Angew Math. 1908;134:198–287. [62] Dirichlet GL. Über die Reduktion der positiven quadratischen Formen mit drei unbestimmten ganzen Zahlen. J Reine Angew Math. 1850;40:209–27. [63] Niggli P. Die topologische Strukutranalyse. I. Z. Kristallogr Cryst Mater. 1927;65: 391–415.

References

117

[64] Blatov VA, Shevchenko AP, Serenzhkin VN. Crystal space analysis by means of Voronoi– Dirichlet polyhedra. Acta Crystallogr A. 1995;51:909–16. [65] Blatov VA. Voronoi–Dirichlet polyhedra in crystal chemistry: theory and applications. Crystallogr Rev. 2004;10:249–318. [66] Descartes R. Principia philosophiae. Amsterdam: Apud Ludovicum Elzevirum, 1644. [67] Liebling TM, Pournin L. Voronoi diagrams and Delauney triangulations: ubiquitous Siamese twins. Doc Math Extra Volume Optim Stories. 2012;2012:419–31. [68] Blatov VA, Serezhkin VN. Stereoatomic model of the structure of inorganic and coordination compounds. Russ J Inorg Chem. 2000;45:S105–22. [69] Voroglide. Fernuni Hagen. http://www.pi6.fernuni-hagen.de/GeomLab/VoroGlide. Accessed 01 Mar 2017. [70] Blatov VA, Ilyushin GD, Blatova OA, Anurova NA, Ivanov-Schits AK, Dem’yanets LN. Analysis of migration paths in fast-ion conductors with Voronoi–Dirichlet partition. Acta Crystallogr B. 2006;62:1010–8. [71] Blatov VA, Shevchenko AP. Registration of the computer program ‘ToposPro’, Russian certificate No. 2013619016, 2013. [72] Global materials market for alternative energy storage mechanisms. Frost & Sullivan, 2011:9833–9. [73] Buss K, Wrobel P, Doetsch C. Global distribution of grid connected electrical energy storage systems. Int J Sustainable Energy Plann Manag. 2016;9:31–56. [74] Kummer JT, Neill W. Battery having a molten alkali metal anode and a molten sulfur cathode. Publication date: 26. November 1968, US Patent US3413150 A. [75] Yao Y-F, Kummer JT. Ion exchange properties of and rates of ionic diffusion in beta-alumina. J Inorg Nucl Chem. 1967;29:2453–75. [76] Hueso KB, Armand M, Rojo T. High temperature sodium batteries: status, challenges and future trends. Energy Environ Sci. 2013;6:734–49. [77] Daniel C, Besenhard J. Handbook of battery materials, vol. 1, 2nd ed. Weihnheim: Wiley-VCH; 2011. [78] Dunn B, Kamath H, Tarascon J-M. Electrical energy storage for the grid: a battery of choices. Sci. 2011;334:928–35. [79] Belsky A, Hellenbrandt M, Karen VL, Luksch P. New developments in the inorganic crystal structure database (ICSD): accessibility in support of materials research and design. Acta Crystallogr B. 2002;58:364–9. [80] Whittingham MS, Huggins RA. Measurement of sodium ion transport in beta alumina using reversible solid electrodes. J Chem Phys. 1971;54:414–6. [81] Song S, Duong HM, Korsunsky AM, Hu N, Lu L. A Na+ superionic conductor for room-temperature sodium batteries. Sci Rep. 2016;6:32330. [82] Beyeler HU, Shannon RD, Chen HY. Ionic conductivity of single-crystal Na5YSi4O12ʹ. Appl Phys Lett. 1980;37:934–5. [83] Glöser S, Espinoza LT, Gandenberger C, Faulstich M. Raw material criticality in the context of classical risk assessment. Resour Policy. 2015;44:35–46. [84] Brown ID, Poeppelmeier KR, editors. Bond valences. Berlin Heidelberg: Springer, 2014. [85] Pauling L. The principles determining the structure of complex ionic crystals. J Am Chem Soc. 1926;51:1010–26. [86] Brown ID. Recent developments in the methods and applications of the bond valence model. Chem Rev. 2009;109:6858–919. [87] Baur WH. Bond length variation and distorted coordination polyhedra in inorganic crystals. Trans Am Crystallogr Assoc. 1970;6:125–55.

118

4 Battery Materials

[88] Donnay G, Allmann R. How to recognize O2−, OH−, and H2O in crystal structures determined by X-rays. Am Mineral. 1970;55:1003–15. [89] Brown ID, Altermatt D. Bond-valence parameters obtained from a systematic analysis of the inorganic crystal structure database. Acta Crystallogr B. 1970;41:244–7. [90] Adams S. From bond valence maps to energy landscapes for mobile ions in ion-conducting solids. Solid State Ionics. 2006;177:1625–30. [91] Preiser C, Loesel J, Brown ID, Kunz M, Skowron A. Long-range Coulomb forces and localized bonds. Acta Crystallogr B. 1999;55:698–711. [92] Brown ID. The chemical bond in inorganic chemistry: the bond valence model. Oxford: Oxford University Press, 2002:27 [93] Garrett JD, Greedan JE, Faggiani R, Carbotte S, Brown ID. Single-crystal growth and structure determination of Ag16I12P2O7. J Solid State Chem. 1982;42:183–90. [94] Zhou Y, Adams S, Rao RP, Edwards DD, Neiman A, Pestereva N. Charge transport by polyatomic anion diffusion in Sc2(WO4)3. Chem Mater. 2008;20:6335–45. [95] Yashima M, Sekikawa T, Sato D, Nakano H, Omoto K. Crystal structure and oxide-ion diffusion of nanocrystalline, compositionally homogeneous ceria–zirconia Ce0.5Zr0.5O2 up to 1176K. Cryst Growth Des. 2013;13:829–37. [96] Cabana J, Ling CD, Oró-Solé J, Gautier D, Tobias G, Adams S, Canadell E, Palacin MR. Antifluorite-type lithium chromium oxide nitrides: synthesis, structure, order, and electrochemical properties. Inorg Chem. 2004;43:7050–60. [97] Mazza D. Modeling ionic conductivity in Nasicon structures. J Solid State Chem. 2001; 156:154–60. [98] Fedotov SS, Kabanov AA, Kabanova NA, Blatov VA, Zhugayevych A, Abakumov AM, Khasanova NR, Antipov EV. Crystal structure and Li-ion transport in Li2CoPO4F high-voltage cathode material for Li-ion batteries. J Phys Chem C. 2017;121:3194–202. [99] Ouerfelli N, Guesmi A, Mazza D, Madani A, Zid MF, Driss A. Synthesis, crystal structure and mono-dimensional thallium ion conduction of TlFe0.22Al0.78As2O7ʹ. J Solid State Chem. 2007;180:1224–9. [100] Safanama D, Sharma N, Rao RP, Brand HE, Adams S. Structural evolution of NASICON-type Li1 +xAlxGe2-x(PO4)3 using in situ synchrotron X-ray powder diffraction. J Mater Chem A. 2006;4:7718–26. [101] Kan WH, Huq A, Manthiram A. Exploration of a metastable normal spinel phase diagram for the quaternary Li–Ni–Mn–Co–O system. Chem Mater. 2016;28:1832–7. [102] Adams S. Relationship between bond valence and bond softness of alkali halides and chalcogenides. Acta Crystallogr B. 2001;57:278–87. [103] Chen H, Adams S. Bond softness sensitive bond-valence parameters for crystal structure plausibility tests. IUCrJ. 2017;4:614–25. [104] Swenson J, Adams S. Application of the bond valence method to reverse Monte Carlo produced structural models of superionic glasses. Phys Rev B. 2001;64:024204. [105] Swenson J, Adams S. Mixed alkali effect in glasses. Phys Rev Lett. 2003;90:155507. [106] Hall A, Swenson J, Adams S, Meneghini C. Mixed mobile ion effect and cooperative motions in silver-sodium phosphate glasses. Phys Rev Lett. 2008;101:195901. [107] Adams S, Rao RP. High power lithium ion battery materials by computational design. Phys Status Solidi A. 2011;208:1746–53. [108] Adams S, Rao RP. Structural requirements for fast lithium ion migration in Li10GeP2S12ʹ. J Mater Chem. 2012;22:7687–91. [109] Wong LL, Chen HM, Adams S. Sodium-ion diffusion mechanisms in the low cost high voltage cathode material Na2+δFe2−δ/2(SO4)3ʹ. Phys Chem Chem Phys. 2015;17:9186–93.

References

119

[110] Nishitani Y, Adams S, Ichikawa K, Tsujita T. Evaluation of magnesium ion migration in inorganic oxides by the bond valence site energy method. Solid State Ionics. 2018;315:111–115.10.1016/j. ssi.2017.11.031. [111] Wong LL, Chen H, Adams S. Design of fast ion conducting cathode materials for grid-scale sodiumion batteries. Physical Chemistry Chemical Physics. 2017;19:7506–7523.10.1039/C7CP00037E. [112] Pearson WH, Lomax JF. X-ray crystal diffraction study of Zr,Na-betaʺ-alumina. Mate Res Soc Symp Proc. 1993;293:315–21. [113] Boilot JP, Collin G, Colomban P. Relation structure-fast ion conduction in the NASICON solid solution. J Solid State Chem. 1988;73:160–71. [114] Collin G, Comes R, Boilot JB, Colomban P. Disorder of tetrahedra in Nasicon-type structure – Part I. Na3Sc2(PO4)3: structures and ion-ion correlations. J Phys Chem. 1986;47:843–54. [115] Susman S, Delbecq CJ, Brun TO, Prince E. Fast ion transport in the Nasicon analog Na3Sc2(PO4)3: structure and conductivity. Solid State Ionics. 1983;9:839–44. [116] Chamkir M, El Jazouli A, De Waal D. Synthesis, crystal structure and spectroscopy properties of Na3AZr(PO4)3 (A = Mg, Ni) and Li2.6Na0.4NiZr(PO4)3 phosphates. J Solid State Chem. 2006;179:1883–91. [117] Zatrovsky IV. NASICON-type Na3V2(PO4)3ʹ. Acta Crystallogr E. 2010;66:pi12–12. [118] Maksimov BA, Belov NV. The high temperature X-ray analysis of the monocrystals Na5YSi4O12ʹ. Dokl Akad Nauk SSSR. 1981;261:623–7. [119] Rüscher CH, Gesing TM, Buhl J-C. Anomalous thermal expansion behaviour of Na8[AlSiO4]6 ̅ phase transition by untilting and contraction of TO4 units. Z (NO3)2-sodalite: P4̅3n to Pm3n Kristallogr Cryst Mater. 2003;218:332–44. [120] Barth TF, Posnjak E. Silicate structures of the cristobalite type: III. Structural relationship of high-cristobalite, alpha-carnegieite, and Na2SiO4ʹ. Z Kristallogr. 1932;81:376–85. [121] Baur WH, Joswig W. The phases of natrolite occurring during dehydration and rehydration studied by single crystal X-ray diffraction methods between room temperature and 923 K. Neues Jahrb Mineral Abh Monatsh. 1996;1996:171–87. [122] Kahlenberg V, Böhm H. Crystal structure of hexagonal trinepheline – a new synthetic NaAlSiO4 modification. Am Mineral. 1998;83:631–7. [123] Isasi J, Daidouh A. Synthesis, structure and conductivity study of new monovalent phosphates with the langbeinite structure. Solid State Ionics. 2000;133:303–13. [124] Williams JJ, Smith CW, Evans KE, Lethbridge ZA, Walton RI. Off-axis elastic properties and the effect of extraframework species on structural flexibility of the NAT-type zeolites: simulations of structure and elastic properties. Chem Mater. 2007;19:2423–34. [125] Saniz R, Freeman AJ. Pressure effects on the electronic properties and superconductivity of the beta-pyrochlore oxides: AOs2O6 (A = Na, K, Rb, Cs). Phys Rev B. 2007;72:024522. [126] Nishi F, Takéuchi Y. Cubic structure of sodium calcium germanate Na3.7Ca1.15Ge3O9ʹ. Acta Crystallogr C. 1988;44:1867–9. [127] Michiue Y, Watanabe M. NaxCrxTi8-xO16, priderite with sodium ions in the tunnel structural study for stability and Na ion transport. J Solid State Chem. 1995;116:269–99. [128] Vuli P, Kahlenberg V. On the high temperature behaviour of monoclinic trinepheline. Neues Jahrb Mineral Geol Palaeontol A. 2012;189:197–206. [129] Serras P, Palomares V, Rojo T, Brand HE, Sharma N. Structural evolution of high energy density V(3+)/V(4+) mixed valent Na3V2O2 x (PO4)2F3-2x (x = 0.8) sodium vanadium fluorophosphate using in situ synchrotron X-ray powder diffraction. J Mater Chem A. 2014;2:7766–79. [130] Brachtel G, Bukovec N, Hoppe R. Das erste Oxomanganat(III) mit Inselstruktur: zur Kenntnis von Na5(MnO4). Z Anorg Allg Chem. 1984;515:101–13. [131] Wang H-W, Bish DL. A P(H2O)-dependent structural phase transition in zeolite natrolite. Am Miner. 2008;93:1191–4.

120

4 Battery Materials

[132] Peacor DR, Bürger MJ. The determination and refinement of the structure of narsarsukite, Na2TiOSi4O10ʹ. Am Mineral. 1962;47:539–56. [133] Wohlfahrt A. Kristallchemie von verbindungen und mischkristallen mit millarit-struktur. Heidelb Geowiss Abh. 1998;92:1–91. [134] Zschornak M. 2015. dissertation, TU Bergakademie Freiberg. http://nbn-resolving.de/urn:nbn: de:bsz:105-qucosa-169703. [135] Thomas LH. The calculation of atomic fields. Math Proc Cambridge Philos Soc. 1927;23: 542–8. [136] Fermi E. Eine statistische methode zur bestimmung einiger eigenschaften des atoms und ihre anwendung auf die theorie des periodischen systems der elemente. Z Phys. 1928;48:73–9. [137] Slater JC. A simplification of the Hartree-Fock method. Phys Rev. 1951;81:385. [138] Hohenberg P, Kohn W. Inhomogeneous electron gas. Phys Rev. 1964;136:B864. [139] Kohn W, Sham LJ. Self-consistent equations including exchange and correlation effects. Phys Rev. 1965;140:A1133. [140] Johannes MD, Love CT, Swider-Lyons K. Calculations in Li-ion battery materials. In: Breitkopf C, Swider-Lyons K, editors. Springer Handbook of Electrochemical Energy. Berlin Heidelberg: Springer, 2016:313–28. [141] Jain A, Ong SP, Hautier G, Chen W, Richards WD, Dacek S, et al. Commentary: the materials project: a materials genome approach to accelerating materials innovation. APL Mater. 2013;1:011002. [142] Jain A, Shin Y, Persson KA. Computational predictions of energy materials using density functional theory. Nat Rev Mater. 2016;1:15004. [143] Urban A, Seo D-H, Ceder G. Computational understanding of Li-ion batteries. NPJ Comput Mater. 2016;2:16002. [144] Nørskov JK, Bligaard T, Rossmeisl J, Christensen CH. Towards the computational design of solid catalysts. Nat Chem. 2009;1:37–46. [145] Sokolov AN, Atahan-Evrenk S, Mondal R, Akkerman HB, Sánchez-Carrera RS, Granados-Focil S, et al. From computational discovery to experimental characterization of a high hole mobility organic crystal. Nat Commun. 2011;2:437. [146] Pyzer-Knapp EO, Suh C, Gómez-Bombarelli R, Aguilera-Iparraguirre J, Aspuru-Guzik A. What is high-throughput virtual screening? A perspective from organic materials discovery. Annu Rev Mater Res. 2015;45:195–216. [147] Sanvito S, Oses C, Xue J, Tiwari A, Zic M, et al. Accelerated discovery of new magnets in the Heusler alloy family. Sci Adv. 2017;14:e1602241. [148] Ong SP, Richards WD, Jain A, Hautier G, Kocher M, Cholia S, et al. Python materials genomics (pymatgen): a robust, open-source python library for materials analysis. Comput Mater Sci. 2013;68:314–9. [149] Larsen AH, Mortensen JJ, Blomqvist J, Castelli IE, Christensen R, Dułak M, et al. The atomic simulation environment – a Python library for working with atoms. J Phys Condens Matter. 2017;29:273002. [150] Curtarolo S, Setyawan W, Hart GL, Jahnatek M, Chepulskii RV, Taylor RH, et al. AFLOW: an automatic framework for high-throughput materials discovery. Comput Mater Sci. 2012;58:218–226. [151] Tarascon J-M, Armand M. Issues and challenges facing rechargeable lithium batteries. Nature. 2001;414:359–67. [152] Armand M, Tarascon J-M. Building better batteries. Nature. 2008;451:652–7. [153] Bruce PG, Freunberger SA, Hardwick LJ, Tarascon J-M. Li-O2 and Li-S batteries with high energy storage. Nat Mater. 2012;11:19–29.

References

121

[154] Larcher D, Tarascon J-M. Towards greener and more sustainable batteries for electrical energy storage. Nat Chem. 2015;7:19–29. [155] Kim S-W, Seo D-H, Ma X, Ceder G, Kang K. Electrode materials for rechargeable sodium-ion batteries: potential alternatives to current lithium-ion batteries. Adv Energy Mat. 2012;2:710–21. [156] Hautier G, Jain A, Ong SP, Kang B, Moore C, Doe R, Ceder G. Phosphates as lithium-ion battery cathodes: an evaluation based on high-throughput ab initio calculations. Chem Mater. 2011;23:3495–508. [157] Aydinol MK, Kohan AF, Ceder G, Cho K, Joannopoulos J. Ab initio study of lithium intercalation in metal oxides and metal dichalcogenides. Phys Rev B. 1997;56:1354. [158] Perdew JP, Burke K, Ernzerhof M. Generalized gradient approximation made simple. Phys Rev Lett. 1997;77:3865. [159] Heyd J, Scuseria GE. Hybrid functionals based on a screened Coulomb potential. J Chem Phys. 2003;118:8207. [160] Anisimov VI, Zaanen J, Andersen OK. Band theory and Mott insulators: Hubbard U instead of Stoner I. Phys Rev B. 1991;44:943. [161] Kulik HJ, Cococcioni M, Scherlis DA, Marzari N. Density functional theory in transition-metal chemistry: a self-consistent Hubbard U approach. Phys Rev Lett. 2006;97:103001. [162] Ong SP, Chevrier VL, Ceder G. Comparison of small polaron migration and phase separation in olivine LiMnPO4 and LiFePO4 using hybrid density functional theory. Phys Rev B. 2011;83:075112. [163] Becke AD. Density-functional exchange-energy approximation with correct asymptotic behavior. Phys Rev A. 1988;38:3098. [164] Lee C, Yang W, Parr RG. Development of the Colle-Salvetti correlation-energy formula into a functional of the electron density. Phys Rev B. 1988;37:785. [165] Becke AD. Density‐functional thermochemistry. III. The role of exact exchange. J Chem Phys. 1993;98:5648. [166] Adamo C, Barone V. Toward reliable density functional methods without adjustable parameters: the PBE0 model. J Chem Phys. 1999;110:6158. [167] Toumar AJ, Ong SP, Richards WD, Dacek S, Ceder G. Vacancy ordering in O3-type layered metal oxide sodium-ion battery cathodes. Phys Rev Appl. 2015;4:064002. [168] Simon P, Gogotsi Y, Dunn B. Where do batteries end and supercapacitors begin? Science. 2014;343:1210–1. [169] Kang K, Meng YS, Bréger J, Grey CP, Ceder G. Electrodes with high power and high capacity for rechargeable lithium batteries. Sci. 2006;311:977–80. [170] Whittingham MS. Ultimate limits to intercalation reactions for lithium batteries. Chem Rev. 2014;114:11414–43. [171] Yabuuchi N, Kubota K, Dahbi M, Komaba S. Research development on sodium-ion batteries. Chem Rev. 2014;114:11636–82. [172] Canepa P, Gautam GS, Hannah DC, Malik R, Liu M, Gallagher KG, Persson KA, Ceder G. Odyssey of multivalent cathode materials: open questions and future challenges. Chem Rev. 2017;117:4287–341. [173] Yoo HD, Shterenberg I, Gofer Y, Gershinsky G, Pour N, Aurbach D. Mg rechargeable batteries: an on-going challenge. Energy Environ Sci. 2013;6:2265–79. [174] Rong Z, Malik R, Canepa P, Gautam GS, Liu M, Jain A, Persson K, Ceder G. Materials design rules for multivalent ion mobility in intercalation structures. Chem Mater. 2015;27:6016–21. [175] Canepa P, Bo S-H, Gautam GS, Key B, Richards WD, Shi T, et al. High magnesium mobility in ternary spinel chalcogenides. Nat Commun. 2017;8:1759. [176] Pechukas P. Transition State Theory. Annu Rev Phys Chem. 1981;32:159–77.

122

4 Battery Materials

[177] Cabrera N, Mott NF. Theory of the oxidation of metals. Reports on Progress in Physics 1949;12:163–84. [178] Yashima M, Itoh M, Inaguma Y, Morii Y. Crystal Structure and Diffusion Path in the Fast LithiumIon Conductor La0.62Li0.16TiO3. J Am Chem Soc. 2005;127:3491–3495.10.1021/ja0449224. [179] Sheppard D, Terrel R, Henkelman G. Optimization methods for finding minimum energy paths. J Chem Phys. 2008;128:134106. [180] Miara LJ, Suzuki N, Richards WD, Wang Y, Kim JC, Ceder G. Li-ion conductivity in Li9S3N. J Mater Chem A. 2015;3:20338–44. [181] Ong SP, Mo Y, Richards WD, Miara L, Lee HS, Ceder G. Phase stability, electrochemical stability and ionic conductivity of the Li10±1MP2X12 (M = Ge, Si, Sn, Al or P, and X = O, S or Se) family of superionic conductors. Energy Environ Sci. 2013;6:148–56. [182] Kamaya N, Homma K, Yamakawa Y, Hirayama M, Kanno R, Yonemura M, et al. A lithium superionic conductor. Nat Mater. 2011;10:682–686. [183] Bron P, Johansson S, Zick K, Schmedt Auf Der Günne J, Dehnen S, Roling B. Li10SnP2S12: an affordable lithium superionic conductor. J Am Chem Soc. 2013;135:15694–7. [184] Richards WD, Tsujimura T, Miara LJ, Wang Y, Kim JC, Ong SP, Uechi I, Suzuki N, Ceder G. Design and synthesis of the superionic conductor Na10SnP2S12ʹ. Nat Commun. 2016;7:11009. [185] Miara LJ, Ong SP, Mo Y, Richards WD, Park Y, Lee J-M, Lee HS, Ceder G. Effect of Rb and Ta doping on the ionic conductivity and stability of the garnet Li7+2x–Y(La3–XRbx)(Zr2–YTay) O12 (0 ≤ x ≤ 0.375, 0 ≤ y ≤ 1) superionic conductor: a first principles investigation. Chem Mater. 2013;25:3048–55. [186] Kuhn A, Gerbig O, Zhu C, Falkenberg F, Maier J, Lotsch BV. A new ultrafast superionic Liconductor: ion dynamics in Li11Si2PS12 and comparison with other tetragonal LGPS-type electrolytes. Phys Chem Chem Phys. 2014;16:14669–74. [187] Malik R, Abdellahi A, Ceder G. A critical review of the Li insertion mechanisms in LiFePO4 electrodes. J Electrochem Soc. 2013;160:A3179–97. [188] Liu M, Rong Z, Malik R, Canepa P, Jain A, Ceder G, Persson KA. Spinel compounds as multivalent battery cathodes: a systematic evaluation based on ab initio calculations. Energy Environ Sci. 2015;8:964–74. [189] Gautam GS, Canepa P, Abdellahi A, Urban A, Malik R, Ceder G. The intercalation phase diagram of Mg in V2O5 from first-principles. Chem Mater. 2015;27:3733–42. [190] Gautam GS, Canepa P, Malik R, Liu M, Persson K, Ceder G. First-principles evaluation of multi-valent cation insertion into orthorhombic V2O5ʹ. Chem Commun. 2015;51: 13619–22. [191] Maxisch T, Zhou F, Ceder G. Ab initio study of the migration of small polarons in olivine LixFePO4 and their association with lithium ions and vacancies. Phys Rev B. 2006; 73:104301. [192] Xu B, Meng S. Factors affecting Li mobility in spinel LiMn2O4 – a first-principles study by GGA and GGA + U methods. J Power Sources. 2010;195:4971–6. [193] Hannah DC, Gautam GS, Canepa P, Rong Z, Ceder G. Magnesium ion mobility in post-spinels accessible at ambient pressure. Chem Commun. 2017;53:5171–4. [194] Sun X, Blanc L, Nolis GM, Bonnick P, Cabana J, Nazar LF. NaV1.25Ti0.75O4: a potential post spinel cathode material for Mg batteries. Chem Mater. 2018;30:121–8. [195] Mukherjee A, Sa N, Phillips PJ, Burrell A, Vaughey J, Klie RF. Direct Investigation of Mg intercalation into the orthorhombic V2O5 cathode using atomic-resolution transmission electron microscopy. Chem Mater. 2017;29:2218–26. [196] Rong Z, Kichaev D, Canepa P, Huang W, Ceder G. An efficient algorithm for finding the minimum energy path for cation migration in ionic materials. J Chem Phys. 2016;145:074112.

4.2 Electrodes: definitions and systematisation

123

[197] Rong Z, Xiao P, Liu M, Huang W, Hannah DC, Scullin W, Persson KA, Ceder G. Fast Mg2+ diffusion in Mo3(PO4)3O for Mg batteries. Chem Commun. 2017;53:7998–8001. [198] Freysoldt C, Grabowski B, Hickel T, Neugebauer J, Kresse G, Janotti A, Van De Walle CG. Firstprinciples calculations for point defects in solids. Rev Mod Phys. 2014;86:253. [199] Islam M, Driscoll D, Fisher C, Slater P. Atomic-scale investigation of defects, dopants, and lithium transport in the LiFePO4 olivine-type battery material. Chem Mater. 2005;17: 5085–92. [200] Morgan D, Van Der Ven A, Ceder G. Li conductivity in LixMPO4 (M = Mn, Fe, Co, Ni) olivine materials. Electrochem Solid-State Lett. 2004;7:A30–2. [201] Maragliano L, Vanden-Eijnden E. A temperature accelerated method for sampling free energy and determining reaction pathways in rare events simulations. Chem Phys Lett. 2006;426:168–75. [202] Laio A, Parrinello M. Escaping free-energy minima. PNAS. 2002;99:12562–6. [203] Abrams C, Bussi G. Enhanced sampling in molecular dynamics using metadynamics, replicaexchange, and temperature-acceleration. Entropy. 2014;16:163–99. [204] Boulfelfel SE, Seifert G, Leoni S. Atomistic investigation of Li+ diffusion pathways in the olivine LiFePO4 cathode material. J Mater Chem. 2011;21:16365–72. [205] Johnston W, Heikes R, Sestrich D. The preparation, crystallography, and magnetic properties of the LixCo(1-x)O system. J Phys Chem Solids. 1958;7:1–13. [206] Mizushima K, Jones P, Wiseman P, Goodenough JB. LixCoO2 (0 < x < 1): a new cathode material for batteries of high energy density. Mater Res Bull. 1980;15:783–9. [207] Padhi AK, Nanjundaswamy K, Goodenough JB. Phospho-olivines as positive-electrode materials for rechargeable lithium batteries. J Electrochem Soc. 1997;144:1188–94. [208] Fuchs JN. Über ein neues mineral (triphylin). Adv Synth Catal. 1834;3:98–104. [209] Fuchs JN. Vermischte Notizen. J Prakt Chem. 1835;5:316–24. [210] Björling CO, Westgren A. Minerals of the varuträsk pegmatite: IX. X-ray Studies on triphylite, varulito, and their oxidation products. Geol Foeren Stockholm Foerh. 1938;60: 67–72. [211] Yakubovich OV, Simonov MA, Belov NV. The crystal structure of a synthetic triphylite LiFe[PO4]. Sov Phys Dokl. 1977;22:347.

4.2 Electrodes: definitions and systematisation – a crystallographers view Falk Meutzner, Matthias Zschornak, Melanie Nentwich, Damien Monti and Tilmann Leisegang Abstract: Electrodes are, in combination with electrolytes and the active, reacting materials the function-giving materials in electrochemical energy storage devices. They are responsible for the transfer of electrons and provide the surface at which the electrochemical reactions take place. Those electrochemical reactions span the

This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Meutzner, F., Zschornak, M., Nentwich, M., Monti, D., Leisegang, T. Electrodes: definitions and systematisation – a crystallographers view. Physical Sciences Reviews [Online] 2018, 3. DOI: 10.1515/psr-2018-0043

124

4 Battery Materials

potential difference which drives the battery. We present a crystallographically inspired systematisation of all electrodes found in electrochemical storages that comprise inert and reactive electrodes, subdivided in active and passive electrodes, and solvation, mixed crystal, and phase transition electrodes, respectively. After the description of all electrode types we present a concise summary of battery chemistries and the applied electrode types. Keywords: crystallography, systematisation

electrochemistry,

electrode,

anode,

cathode,

4.2.1 Definitions The term electrode is a portmanteau word derived from the greek words “ηλεκτρόν” (electron), amber, and “ὁδός” (odos), way, that describes an electron-conducting material at which the electronic current is transferred to another, non-electron-conducting medium [1]. In one general setup, two electrodes are externally connected by an electron-conducting material while in-between, the non-electron-conducting medium such as a vacuum (e.g. electron tube), a gas phase (e.g. sputtering process, X-ray detector), a plasma, or an electrolyte is placed. Two sorts of electrodes can be distinguished: the cathode, which transfers negative charge to the medium (electron donator), and the anode which transfers positive charge to the medium, this way accepting negative charge from the medium (electron acceptor); or in the words of Faraday: “The cathode is that surface at which the current leaves the decomposing body, and is its positive extremity; the combustible bodies, metals, alkalies, and bases, are evolved there, and it is in contact with the negative electrode”. Electrodes are used in a variety of fields of interest such as biology [2], energy storage [3], and medicine [4]. The world’s largest electrodes can be found at the terminals of unipolar high-voltage direct-current power line submarine cables, for instance, beneath the Baltic Sea linking the power grids in Sweden and Germany [5]. With a capacity of 600 MW at 450 kV DC, copper (cathode) and titanium (anode) electrodes are utilised to transfer the electric current to the ground since the seawater is used as return conductor. The cathode consists of a bare copper ring with a 2 km diameter, whereas the anode consists of 40 titanium nets each with a surface of 20 m2. The most widespread electrodes are tubes mostly used for material characterisations (e.g. X-ray tubes and electron guns for microscopes), material preparation (e.g. magnetron sputtering), and chemical processes (e.g. electrolysis, fuel cells, and batteries) involving oxidation and reduction reactions also known as ‘redox’ processes. In the latter, electrodes essentially are solid materials like metals [6, 7], metal oxides [8, 9], carbon [10], or ceramics [11].

4.2 Electrodes: definitions and systematisation

125

Redox reactions are the foundation of electrochemical cells as they simply convert chemical energy into electric energy by means of chemical reactions of an active material present at the surface of both electrodes and acting as a ‘fuel’. At the anodic side, where the oxidation occurs, electrons are withdrawn from the active material, augmenting its oxidation state. At the cathodic side, where the reduction takes place, electrons are received resulting in a diminution of the other active material’s oxidation state. The electrode reaction is characterised by the presence of three phases known as the three-phase boundary: the electrode enabling electron transfer, the active material acting as reactant, and the electrolyte providing ionic mass transport between the electrodes. As will be shown later, the active material may be a part of the electrode. At the surface of an isolated electrode, both oxidation and reduction processes in the three-phase boundary are in a dynamical equilibrium. When two distinct electrodes are connected by means of an electrolyte, a resulting redox reaction is generally preferential promoting an imbalance of the chemical potentials of the educts on the one side and the products on the other. This imbalance is induced by a specific difference in free enthalpy, also called Gibbs free energy and the reaction will be considered as spontaneous only if its value is negative. The difference in free enthalpy induces an electric potential difference between the electrodes (see chapter 2). This potential difference defines the cell voltage for the open electric circuit. In this steady state, the electric potential difference counters the respective chemical potential difference and no current or mass transport takes place through the cell. This holds as well for a closed circuit as long as the externally applied voltage reflects this specific electric potential difference. The application of a voltage deviating from the equilibrium voltage will allow for electronic transport through the external circuit. In the case of a smaller voltage this electric current will counter the polarisation of the system, in the case of a higher voltage the polarisation is increased. The electric current will generate ionic current and mass transport through the electrolyte of the cell in either direction. The operation will modify the concentration of species present at the surface of each electrode. To understand the processes on a particle level one has to comprehend the species’ behaviours at each electrode (see Figure 4.2.1). Thereby, the electrochemical cell can be schematically split in two half-cells containing either the positive or negative electrode. The potential of the respective electrode with respect to the standard hydrogen electrode (SHE) [1] is the equilibrium electrode potential Eeq. If an external electric potential higher than Eeq is applied, electrons will travel from the electrode to the external circuit and oxidation processes are promoted. On the one hand, if one applies an external potential below Eeq, the opposite phenomena happens and reduction processes are favoured. Thus, regarding charge and discharge, the location of oxidation and reduction is not fixed in the setup of the electrochemical cell. Among electrochemical cells, secondary batteries are

126

4 Battery Materials

Discharge process

Charge process Source e– – +

e–

e–

Separator

Negative electrode Anode (oxidation)

e– M+ Electrolyte

M

M+ + e–

M+ + e–

M+

e–

e–

M+

M+

Positive electrode Anode (oxidation)

M+

Positive electrode Cathode (reduction)

M+

e–

Separator

+

Ammeter

Negative electrode Cathode (reduction)

e–

Electrolyte

M+ + e–

M

M

M

M+ + e–

qϕ 0 μ′ M

μM

qϕ

μM

qϕ eq +

μ ′M

qϕ eq +

qϕ –eq

μ″ M

qϕ –eq

μ″ M

Figure 4.2.1: Top: Schematic depiction of the electron flow, electrochemical reactions and the function of the respective electrodes. At the anode side, electrons are accepted from the nonmetal to the metal, while the cathode donates electrons from the metal to the non-metal. The non-metal in our case is the bold line in front of the electrode: the active material (see next section) which undergoes the electrochemical reaction. During the voluntary discharge reaction, the anode side becomes the electron source for the outer circuit (the accumulator form is hinted at with the solid black lines indicating the battery poles) while the cathode is the electron sink for the outer circuit. During the (non-voluntary) charge reaction, the polarity of the poles is reversed. At the bottom, the electric (dark blue) and chemical potentials (dark orange) are shown to illustrate the development of the electrode potentials ’ – and ’ + (solid) eq with respect to their particular equilibrium electrode potentials ’eq – and ’ + (dashed). In the discharge case, both the chemical and electric potentials promote the movement of the M+ ions from negative to positive electrode. In the charge case, the electromotive force of the set terminal potential promotes the reduction and oxidation at the opposite electrodes, inverting the discharge reactions. The M+ ions now move from the positive to the negative electrode, as do the electrons on the outer circuit.

reversible, enabling both discharge and charge reactions, in which the polarity of the electron flow is reversed. Therefore we strongly recommend terming the respective electrodes positive and negative electrode to avoid misunderstanding. During discharge, the negative electrode is connected to the positive electrode. Thus, based on the previous considerations, the oxidation reaction occurs at the negative electrode and the reduction at the positive electrode. Taking into account electric as well as chemical potential, the negative electrode carries a lower electrochemical potential than the positive. During the charging process, the opposed externally

4.2 Electrodes: definitions and systematisation

127

supplied voltage source is stronger than the source of the electrochemical cell voltage, the negative electrode is connected to a more negative and the positive electrode to a more positive electric potential. The locations of oxidation and reduction, the subsequent current and mass transport, as well as the respective electrochemical potential gradient are reversed. That said, it is admitted that by convention the electrode labelling is chosen based on the discharge reaction, hence the cathode and anode would be the positive and negative electrode, respectively. In summary, the exergonic galvanic reaction, the change of free enthalpy, and the respective chemical potential drive the reaction direction of the discharge process, whereas the externally applied voltage source and the respective electric potential drive the reaction direction of the charge process. In some cases, confusion may occur between physicists, chemists, and technologists regarding the definition and polarity of anode and cathode. There may be three reasons, we want to shortly address: 1. Technical versus particle current direction: The term cathode was coined by William Whewell in 1834 and introduced by Michael Faraday in his paper ‘Experimental Researches in Electricity’ [13]. It is derived from the greek κατοδοσ which means ‘to descend’ or ‘the way down’ and refers to the direction of the technical current flow in terms of the electric potential from the positive ‘downwards’ to the negative electric pole. In respect, the term anode is derived from ανοδοσ and means ‘to ascend’ or ‘the way up’. Electrons, however, are negatively charged and will thus move from the negative pole to the positive, whereas positively charged particles like cations behave vice versa. 2. Outer versus inner circuit: In actual, physical batteries, the anode side is labelled as negative electric pole of the battery and the cathode as positive, since electrons move from anode to cathode in the outer circuit attached to the battery. Inside the battery, the anode accepts electrons from the electrochemical reaction and resembles the destination of anionic migration. Because of this fact, the interface is often marked with a positive sign in discrepancy with the negative electric potential. On the other hand, the cathode offers electrons to the electrochemical reaction and resembles the destination of cationic migration, often marked with a minus sign. 3. Potential versus charge: Electrochemical cells may offer two different directions of current and mass transport: Charging (imposed action, electrolytic reaction) and discharging (voluntary action, galvanic reaction). During the latter, the anodic side is both the more negative electric pole of the reaction and the electrode with a lower electrochemical potential. The cathodic side is both the more positive pole and the electrode with the higher electrochemical potential. During the charging process, the gradient of the electrochemical potential between the electrodes is reversed, while the polarity of the electric potential remains constant.

128

4 Battery Materials

The electrodes of electrochemical storage devices need to provide primarily the desired redox reaction, while any other undesired or parasitic reaction, mostly implying electrolyte decomposition, should be avoided. However, degradation processes may lead to the formation of a passive layer made of decomposition products known as solid electrolyte interphase ionically permeable but electronically insulating, avoiding any further redox reaction at the electrolyte/electrode interface [14]. The electrochemical reaction occurs at the surfaces of the electrodes which should be as large and chemically active as possible. The electrodes itself act as electron conductors and thus the electronic conductivity should be sufficiently high. Additionally, the galvanic reaction should provide a high electrode potential difference between the accompanying electrodes (high battery voltage) and should not show immense volume changes. Furthermore, low price and economic benignity is preferential [11].

4.2.2 Systematisation There exist two very general types of electrodes in electrochemical energy storage devices. Either the active material constitutes the electrode or the active material is not part of the electrode. These two respective types will be subsequently called reactive and inert electrode. These terms have been chosen to express that inert electrodes, while enabling the redox reaction, will not change their physical properties – mainly density and volume – during the discharge and charge process. On the other hand, reactive electrodes will take active part in the reaction and alter physically and/or chemically. Further classification of reactive electrodes can be made with respect to compositional changes of the active mass and potential phase transition. Inert electrodes can be further subdivided regarding their catalytic participation into active and passive electrodes. For a comparison of the different electrode types see Figure 4.2.2 and the following descriptions.

4.2.2.1 Inert electrodes Inert electrodes allow electron transfer between the active material and the external circuit of the battery. The active material will be transported to or is on the surface of the electrode and electrochemically reacts there. For some cases, this reaction may be catalysed by the electrode, in which case the electrode is called active, while noncatalysing electrodes are called passive. Depending on the definition or the way the functionality of electrodes is interpreted, inert electrodes may also be called current collectors, as their only function is to conduct electrons from or to the interface with the electrolyte. The active material is sometimes also referred to as an electrode. For example, oxygen is sometimes referred to as the cathode in metal-air batteries. In the following sections, the reacting material will be called active material/mass only.

4.2 Electrodes: definitions and systematisation

129

Electrode Place of redox-reaction electron-conductor

Inert

Reactive

Active Mass ≠ Electrode

Active Mass = Electrode

Redox Reaction Of the active mass

Change of chemical compositon

No

Yes

Active

Passive

Solvation

Mixed Crystal

Phase Trans.

Catalysis & electron transfer Air cathode

Only electron transfer

Solution of ions into electrolyte Zinc anode

Solution of ions into electrode Li-ion battery

...after entrance of ions Lead-acid battery

Active mass

Sulphur cathode

Reacted active mass

Figure 4.2.2: Systematisation of electrodes and graphical representations of the different kinds of electrodes. The common process during discharge is depicted and examples for the different types of electrodes are given (blue italic). The abbreviation Phase Trans. means phase transition.

The capacity of inert electrodes depends only on the amount of active mass, which is not necessarily linked to the mass or volume of the cell. Metal-air or more specifically metal-oxygen batteries show very high energy densities (Li-O2 almost reaches the theoretical energy density of gasoline [15]), because the capacity depends only on the active mass of the anode, as oxygen is introduced to the system from the outside. In theory, the capacity of inert electrodes is infinite as long as fuel is provided from outside reservoirs. This is why in fuel cells, due to their two inert electrodes, no capacity or energy density is defined. Whenever fuel is available, electric energy is created. Nevertheless, in real systems this provision of infinite energy is not achievable. If parts of the electrode are physically removed or corrosion starts to lower the specific surface, mass transport and gained energy gradually decrease. Thus, inert electrodes must be shielded from degradation. Another big problem is clogging of the surface which again lowers the surface and thus the availability of a place for the redox reaction.

130

4 Battery Materials

4.2.2.1.1 Passive inert electrodes The widest-spread passive inert electrodes are carbonaceous because they offer very high porosity and thus surface, they are very good electronic conductors, inert to most solvents, stable at high temperatures, and relatively cheap. Especially in redox flow cells, carbon is in very widespread use. This is particularly true for the most advanced cells, i.e. the all-vanadium redox flow cell, the Regenesys-cell (brominepolysulfide), and the zinc-bromine cell [16, 17]. Further inert electrodes comprise sulphides and selenides, as well as nickel, lead, and titanium meshes [5, 18]. In sodium-sulphur cells the positive electrodes consist of highly porous carbon mats wetted by liquid sulphur as active mass. Since the sulphur phase itself has poor electronic conductivity as does the reaction product of the discharge [19], carbonaceous electrodes are also commonly used for cathodes in these cells. At passive inert electrodes, no catalytic activation of the chemical processes is necessary. It is however possible to enhance the performance at the interface by utilising catalysts. In this case, the passive inert electrode becomes an active inert electrode. 4.2.2.1.2 Active inert electrodes At the interface of an active inert electrode, the electrode material is temporarily taking part in the reaction process while remaining itself unaltered. Next to providing the electron transfer, it forms intermediate bonds to the reacting species and acts as a catalyst. In this respect, metal-air cathodes are active electrodes. In metal-air cells, the ionisation of oxygen can be catalysed by thin layers of noble metals on the surface of the mostly carbonaceous electrode in order to enhance the reaction rate [20]. Neither carbon nor the noble metal is consumed in this process, which is characteristic for catalysts. The electrochemistry of oxygen is very complex and the respective kinetics very slow [20]. Thus, in metal-air batteries, catalysts are utilised in order to allow for a reasonable power density. The kinetics is the main factor for these batteries to be primary batteries, because the reversal of the process – the creation of oxygen at the electrode – is energetically very difficult. Catalysts are currently being researched which allow a cyclability of the process up to tens of times. Two different catalysts can be used: one for the oxygen reduction reaction and one for the oxygen production during charging. Some catalysts also allow for both processes to be reasonably viable. Passive inert electrodes will usually profit largely from introducing catalytic reaction paths to the system. Especially carbonaceous electrodes, whose production can yield different surface morphologies, could strongly benefit from surface activation. This way, two important factors can be ruled out. On the one hand, the morphology is ‘normalised’, allowing usage of more low-quality and thus cheaper base material. This also enables more low-quality electrolytes to be used, if the catalysts are very specific. On the other hand, the efficiency is increased. Common surface

4.2 Electrodes: definitions and systematisation

131

modifications include the introduction of ions on the surface, as well as acid, and heat-treatments [21–23]. In general, the focus of inert electrode research is the optimisation of the chemical processes on the surface of the electrode. Therefore, surface engineering is aiming at understanding and optimising catalytic processes. 4.2.2.2 Reactive electrodes Reactive electrodes take active part in the electrochemical reactions inside of the cell not only in terms of electron transfer but of ionic mass transfer as well. This means, that the active mass needs to be electronically conductive in order to convey electrons to and from the place of the electrochemical reaction. The electrode‘s mass determines the capacity of the cell. The more material reacts, the more electric energy is created in total. This follows from the equations derived by Faraday: C = νe n F The capacity C [As] represents the amount of charge that can be stored in a battery [24] and is directly proportional to the amount of active mass n [mol] in formula units and the amount of electrons created/used for one formula unit of the active mass, νe . The proportionality factor is Faraday’s constant F [As mol−1] and expresses the capacity of one mole of electrons. The product of capacity and cell voltage reflects the stored energy of the battery. W=CU The capacity is usually normalised by the mass of the active mass, yielding the gravimetric capacity, which leads to the specific energy wm if multiplied by the open-circuit potential (OCP) of the cell. If normalised by volume the volumetric capacity multiplied by the OCP yields energy density wV : wm =

CU CU , wV = m V

Specific Energy and energy density are major factors for batteries. The higher these values, the more energy can be converted out of a certain amount of active material. For stationary applications (such as grid compensation) this is less crucial than for mobile applications: For electric cars, mass is very important (specific energy), while for mobile electronic applications like mobile phones and tablets volume is crucial.1 Generally speaking, for high specific energies, low-atomic mass elements should be used for batteries, while materials with high active material densities are sought for

1 Of course, for cars, lower volume is also very advantageous, as is a low mass for mobile electronics.

132

4 Battery Materials

high energy densities. Therefore, lithium, sodium, zinc, manganese, oxygen, and sulphur are commonly used elements in electrochemical energy storage devices. Especially lithium, which has the third-lowest atomic mass of all elements, is an interesting reaction partner, also because it has a very extreme redox-potential of – 3 V vs. SHE. Another interesting partner would be aluminium. Even though it might be much heavier, it allows for triple ionisation, which yields almost the same specific capacity as lithium and, due to its higher density, shows more than three times higher volumetric capacities. These considerations of basic electrochemistry should be kept in mind, when designing new positive electrode materials, as well. They should be as light and/or small as possible. Therefore, light elements and mostly structurally less complex structures are utilised (for more information see chapter 2). 4.2.2.2.1 Solvation electrodes Generally speaking, reactive electrodes can be primarily distinguished from each other by the question: Does the chemical composition of the electrode change during charging/discharging? If it does not, the electrode is simply dissolved into the electrolyte or active material is deposited onto the electrode’s surface. Through this solvation/deposition, the electrode chemistry stays untouched. Problems can arise for example by inhomogeneous deposition in the form of dendrites, which can lead to short-circuiting of the cell. In most cells that have solvation type electrodes, this electrode is the negative part of the cell. The most prominent representative is zinc, which is used as negative mass for example in alkaline batteries. The solvation process has the material solve into the electrolyte, surrounding the generated ions with solvent particles. In the case of Zn, these particles are usually hydroxide-ions which complexate the Zn2+-ions, drawing the ions out of the surface of the electrode [25]. This idea is the same for liquid sodium batteries that utilise sodium beta-alumina solid electrolytes. The solid electrolyte offers places inside of its crystal structure [26], into which sodium diffuses after ionisation and eventually migrates to the opposite side of the electrolyte where it then reacts with either sulphur or nickel chloride. The electro-deposition process works similar to the crystallisation process out of a melt. The crystallisation temperature (melting temperature) corresponds to the standard potential of the electrode. At electrode potentials lower than the standard potential (corresponding to lowering of temperature in this case), reduction will take place, which for the common case of metals corresponds to the electro-deposition process. In this picture, the so-called overpotential2 corresponds to an undercooling in the melt which acts as a kinetic barrier [27]. The gradient of the electrochemical potential through the electrolyte is thus the driving force for depositing and reducing ions from the solution. Similar to dendrite formation in undercooled liquids, dendrite 2 An overpotential is a shift of a specified potential because of shifting chemical potentials. For example, hydrogen-evolution in an electrochemical reaction may be impeded because of additives as compared to its potential in water on a platinum electrode.

4.2 Electrodes: definitions and systematisation

133

formation in electrochemical systems can occur. This effect is usually unwanted and can result in short-circuits and capacity fading. These problems cannot occur if the electrode is liquid for the whole charge/discharge cycle. From a crystallographic and materials science point of view, a dissolution/deposition does not change the crystal structure or the phase of the electrode but subtracts/ adds the same kind of material from/to the electrode’s surface. 4.2.2.2.2 Mixed crystal electrodes Reactive electrodes that undergo a change in chemical composition are the complement to solvation type electrodes. Here, the electrode is the solid solvent into which ions are dissolved in and out of incorporating a redox-reaction inside of the electrode. Mixed crystal electrodes will however not alter their phase during this process. From a crystallographic perspective, this means that the crystal structure and the respective space group with its specific symmetry elements does not change, but the atomic occupancies of some crystallographic sites are altered. This may manifest in two cases: either there is a crystallographic site which is only occupiable by the activemass-atom, or an ‘electrode-site-atom’ is substituted for an active-mass-atom. Both of these phenomena can be subsumed crystallographically as interstitial and substitutional point defects. This process thus resembles a concentration change in a solid solution. In the aforementioned solvation electrodes, the active mass is going in or out of solution with the electrolyte, while for the mixed crystal type, the active mass is going in or out of solution with the electrode, as well. Therefore, this type could also be called (solid) solution type. The term “mixed crystal” is adviced by the IUPAC gold book and better distinguishable from solvation than (solid) solution. Present-day commercial lithium-ion batteries (LIB) use mixed crystal electrodes on both the positive and the negative electrode side of the battery. The keyword for this process is intercalation. A chemical approach of systematisation for LIBs is presented in Ref [28].: three types of positive electrodes are distinguished: insertion, alloying, and conversion. In our system from a crystallographic point of view, insertion and alloying are subsumed as discussed above under mixed crystal electrodes. During the insertion process, empty sites inside of the crystal structure are used for the insertion of ions. Through alloying, the phase is not changed either, as inserted atoms substitute normally occupied sites after the electrochemical reaction. In a crystallographic interpretation, insertion would deal with interstitial voids and vacant sites, while the alloying process would be described by substitutional disordering. In both ways, the host’s crystal structure and the respective space group are not altered by the intercalation. During the conversion process, however, a new phase is created with a new crystal structure – we call these electrodes phase transition electrodes, as specified in the next paragraph. In the case of ‘insertion’, the sites responsible for the uptake of the mobile species percolate through the whole structure, forming different dimensionalities: 1D (lines),

134

4 Battery Materials

2D (layers), and 3D (networks of lines). The most typical examples for these for LIBs are olivines (lithium iron phosphate, LixFePO4 – 1D), layered oxides (lithium cobalt oxide, Li½+xCoO2 – 2D), and spinels (lithium manganese oxide, LixMn2O4 – 3D). Due to insertion, the generally achieved capacities are much lower than e.g. for bulk lithium (two orders of magnitude). The reasons for this are either solubility limits, volume expansion, and/or phase transitions and stability. Alloying electrodes on the other hand offer much higher capacities but tend more strongly to a higher volume increase [6]. 4.2.2.2.3 Phase transition electrodes From a crystallographic point of view, a phase transition means a change of the crystal structure, as commonly presented by means of phase diagrams of specific structures in a system with respect to varying concentrations of the constituents and other intrinsic parameters as e.g. temperature and pressure. For the short-range structure, this means a reordering of the atoms and structure motifs of the reactant. Hence, contrary to the process in mixed crystal electrodes, if an ion enters the bulk of the material and reaches its destination, the symmetry of the atomic positions changes around this ion, creating a new phase that may expand as more ions enter the electrode. An example of this process is found in ZEBRA-type batteries [29]. The positive electrode in the charged state is NiCl2 (space group R3m, trigonal). During discharge, Na+ ions from the negative electrode diffuse through the Na β’’-alumina electrolyte and the secondary (liquid) electrolyte (e.g. NaAlCl4) into the negative electrode, forming metallic Ni and NaCl (space group Fm3m, cubic), which is not soluble in NiCl2 or the secondary electrolyte, as can be seen in the phase diagram in Ref [30]. Another prime example for a phase transition process is the lead-acid battery. The negative electrode is pure lead (Pb, space group Fm3m, cubic), the positive electrode is lead oxide (α-PbO2, space group Pbcn, orthogonal and β-PbO2, space group P42 =mnm, tetragonal [31, 32]), and the electrolyte is composed of sulphur acid. During the discharge process, lead is oxidised eventually forming lead sulphate (PbSO4, space group Pnma, orthorhombic [33]) with sulphate ions from the solution, while the lead oxide is reduced, forming lead sulphate and water in combination with the sulphur acid from the electrolyte. Water molecules are consumed in this process.

4.2.3 Categorisation Given the different electrode types above, we want to present table 4.2.1 with a selection of battery types found in the literature. Electrodes can be further categorised when their morphology is taken into account (nanostructure). Hence, the manufacturing technology has an impact on the battery performance, too. This is discussed in more detail in Refs. [10, 34].

4.2 Electrodes: definitions and systematisation

135

Table 4.2.1: Summary of both industrially spread and more scientific battery concepts. Type stands for the battery type (primary, 1; secondary, 2; tertiary, 3), Prototype describes the prototype of this system, Electrodes categorises the positive and negative electrodes according to our proposal: (Passive, I; Solvation, S; Mixed Crystal, M; Phase transition, P), Mobile species describes the ions that move through the electrolyte during the electrochemical reaction. Type Alkali/Nickel Alkaline Zinc-chloride Zinc/carbon Zinc/mercury Zinc/silver oxide Nickel/zinc Nickel/iron Nickel/cadmium Nickel/metal hydride

1 1 1 1 1 2 2 2 2

Prototype

Electrodes

Mobile species

Zn-C Zn-C Zn-C Zn-Hg Zn-Hg Ni-Cd Ni-Cd Ni-Cd Ni-MH

SP SP SP SP PP SP SP PP PP

OH− H+ H+ OH− OH− OH− OH− OH− OH−

[8] [7] [7] [7] [7] [35] [36] [7] [7]

SP SP

Na+ Na+

[19] [19]

SM

Mg2+

[37]

SP SP SP

Ca5(CrO4)3+1 Li+ Li+

[38] [38] [38]

Molten Salt/Thermally activated batteries HT sodium-sulphur 2 HT Na-S Sodium nickel 2 Na-NiCl chloride (ZEBRA) Liquid metal battery 2 Liquid metal 1 CaCrO4 Ca/CaCrO4 1 CaCrO4 Li/FeS2 1 CaCrO4 Li/CoS2 Metal-air Zinc-air Aluminium-air Iron-air Lithium Lithium-ion accumulator Lithium sulphur accumulator Lithium-air nonaqueous Lithium-air aqueous Lithium/iron sulphide Miscellaneous Lead-acid accumulator Solid oxide fuel cell

Citation

1/2 1 1/2

Zn-air Zn-air Zn-air

SI SI SI

OH− OH− OH-

[20] [39] [40]

2

Li-ion

MM

Li+

[41]

2

Li-S

SP

Li+

[15]

2

Li-air

SI

Li+

[15]

1 1

Li-air Li-S

SI SP

Li+ Li+

[15] [7]

2

Pb-acid

PP

H+

[7]

3

Fuel cell

II

O2−

[42]

136

4 Battery Materials

Table 4.2.1 (continued ) Type Redox flow battery Metal fluoride Daniell cell

2/3 2 1

Prototype

Electrodes

Mobile species

Citation

Redox flow Me-F Daniell

II

H+

[16]

PP SS

FSO42−

[43] [7]

4.2.4 Summary An electrode is a body capable of conducting electrons and transferring them to another body. Since transfer can be either from the electrode to a non-electronconducting body and vice versa, two electrode types can be distinguished; the anode (electron acceptor) and the cathode (electron donator). In electrochemical cells applying both a cathode and an anode, a further discrimination is possible regarding the potential difference between these electrodes; the more negative potential electrode is called the negative electrode and the more positive one the positive electrode. In this manuscript, we proposed a further categorisation of electrodes concerning their crystallography and possible changes during the electron transfer and following electrochemical reactions. Inert electrodes are separate from active mass and do not undergo a chemical reaction themselves, while in reactive electrodes, the active mass is equal to the electrode and thus reacts electrochemically. Inert electrodes may have an additional catalyst to promote the desired reaction (active electrode, otherwise passive). If the reactive electrode undergoes only a mass change due to material dissolving into/ depositing from the electrolyte we call it a solvation electrode. Otherwise, if the active material is inserted into the electrode while the phase stays the same, we denote it mixed crystal electrode. If the phase changes after insertion, it is a phase transition electrode. Eventually, we assigned various battery chemistries to these categories to show the applicability of our systematisation. Funding: Financial support of the Federal Ministry of Education and Research (CryPhysConcept (03EK3029A) and R2RBattery (03SF0542A)) is gratefully acknowledged. DM acknowledges the EC for a H2020 MSCA-IF grant (contract number 743439).

References [1] [2]

Wiberg N. Lehrbuch der Anorganischen Chemie, 102. stark umgearbeitete und verbesserte Auflage. Berlin: Walter de Gruyter & Co, 2007. Takahashi Y, Shevchuk AI, Novak P, Babakinejad B, Macpherson J, Unwin PR, et al. Topographical and electrochemical nanoscale imaging of living cells using voltage-switching

References

[3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23]

[24] [25] [26] [27]

137

mode scanning electrochemical microscopy. Proc Nat Acad. 2012;109:11540–5. DOI: 10.1073/ pnas.1203570109. Tarascon JM, Armand M. Issues and challenges facing rechargeable lithium batteries. Nature. 2001;414:359–67. Leuthardt EC, Schalk G, Wolpaw JR, Ojemann JG, Moran DW. A brain-computer interface using electrocorticographic signals in humans. J Neural Eng. 2004;1:63–71. Sutton SJ, Lewin PL, Swingler SG. Review of global HVDC subsea cable projects and the application of sea electrodes. Int J Electr Power Energy Syst. 2017;87:121–35. Obrovac MN, Chevrier VL. Alloy negative electrodes for Li-Ion batteries. Chem Rev. 2014;114:11444–502. Reddy TD, editor(s). Linden’s handbook of batteries, 4th ed. McGraw Hill, 2011. Daniel C, Besenhard J. Handbook of battery materials, 2nd ed, vol. 1. Weihnheim: Wiley-VCH, 2011. Daniel C, Mohanty D, Li J, Wood DL. Cathode materials review. AIP Conf Proc. 2014;26–43 1597. Dresselhaus MS, Dresselhaus G. Intercalation compounds of graphite. Adv Phys. 2002;51: 1–186. Whittingham MS. Lithium batteries and cathode materials. Chem Rev. 2004;104:4271–301. Meyer DC, Leisegang T. Electrochemical storage materials: from crystallography to engineering. DE GRUYTER OLDENBOURG Publishing House, 2018. Faraday M. Experimental researches in electricity. Philos Trans Royal Soc London. 1832;122:125–62. Verma P, Maire P, Novák P. A review of the features and analyses of the solid electrolyte interphase in Li-ion batteries. Electrochim Acta. 2010;55:6332–41. Bruce PG, Freunberger SA, Hardwick LJ, Tarascon JM. Li-O2 and Li-S batteries with high energy storage. Nat Mater. 2012;11:19–29. Ponce de Leon C, Frías-Ferrer A, González-García J, Szánto DA, Walsh FC. Redox flow cells for energy conversion. J Power Sources. 2006;160:716–32. Skyllas-Kazacos M, Chakrabarti MH, Hajimolana SA, Mjalli FS, Saleem M. Progress in flow battery research and development. J Electrochem Soc. 2011;158:R55–79. Remick RJ, Ang PGP. Electrically rechargeable anionically active reduction – oxidation electrical storage-supply system. U.S. Patent No. 4,485,154 1984 27Nov. Lu X, Xia G, Lemmon JP, Yang Z. Advanced materials for sodium-beta alumina batteries: status, challenges and perspectives. J Power Sources. 2010;195:2431–42. Cheng F, Chen J. Metal–air batteries: from oxygen reduction electrochemistry to cathode catalysts. Chem Soc Rev. 2012;41:2172–92. Sun B, Skyllas-Kazakos M. Chemical modification and electrochemical behaviour of graphite fibre in acidic vanadium solution. Electrochim Acta. 1991;36:513–17. Sun B, Skyllas-Kazacos M. Modification of graphite electrode materials for vanadium redox flow battery application – I. Thermal treatment. Electrochim Acta. 1992;37:1253–60. Sun B, Skyllas-Kazacos M. Chemical modification of graphite electrode materials for vanadium redox flow battery application – part II. Acid treatments. Electrochim Acta. 1992;37:2459–65. Holze R. Anodes – materials for negative electrodes in electrochemical energy technology. AIP Conf Proc. 2014;1597:44–65. Powers RW, Breiter MW. The anodic dissolution and passivation of zinc in concentrated potassium hydroxide solutions. J Electrochem Soc. 1969;116:719–29. Zendejas MA, Thomas JO. Conduction mechanisms in solid electrolytes: Na+ beta-alumina. Phys Scr. 1990;1990:235–44. Diggle JW, Despic AR, Bockris JM. The mechanism of the dendritic electrocrystallization of zinc. J Electrochem Soc. 1969;116:1503–14.

138

4 Battery Materials

[28] Palacin MR. Recent advances in rechargeable battery materials: a chemist’s perspective. Chem Soc Rev. 2009;38:2565–75. [29] Galloway RC. A sodium/beta-alumina/nickel chloride secondary cell. J Electrochem Soc. 1987;134:256–7. [30] Bones RJ, Teagle DA, Brooker SD, Cullen FL. Development of a Ni, NiCl2 positive electrode for a liquid sodium (ZEBRA) battery cell. J Electrochem Soc. 1989;136:1274–7. [31] Hill RJ. The crystal structure of lead dioxides from the positive plate of the lead/acid battery. Mat Res Bull. 1982;17:769–84. [32] D’Antonio P, Santoro A. Powder neutron diffraction study of chemically prepared β-lead dioxide. Acta Cryst B. 1980;36:2394–7. [33] James RW, Wood WA. The crystal structure of barytes, celestine and anglesite. Proc Roy Soc A. 1925;109:598–20. [34] Cherkouk C, Nestler T. Cathodes – technological review. AIP Conf Proc. 2014;1597:134–45. [35] Parker JF, Chervin CN, Pala IR, Machler M, Burz MF, Long JW, et al. Rechargeable nickel–3D zinc batteries: an energy-dense, safer alternative to lithium-ion. Science. 2017;356:415–8. [36] Mulder FM, Weninger BMH, Middelkoop J, Ooms FGB, Schreuders H. Efficient electricity storage with a battolyser, an integrated Ni–fe battery and electrolyser. Energy Environ Sci. 2017;10:756–64. [37] Bradwell DJ, Kim H, Sirk AHC, Sadoway DR. Magnesium–antimony liquid metal battery for stationary energy storage. J Am Chem Soc. 2012;134:1895–7. [38] Guidotti RA. Thermal batteries: a technology review and future directions. Proc 27th Int SAMPE Tech Conf Abuquerque 1995;27:807–18. [39] Egan DR, Ponce De León C, Wood RJK, Jones RL, Stokes KR, Walsh FC. Developments in electrode materials and electrolytes for aluminium–air batteries. J Power Sources. 2013;236:293–310. [40] Narayanan SR, Surya Prakash GK, Manohar A, Yang B, Malkhandi S, Kindler A. Materials challenges and technical approaches for realizing inexpensive and robust iron–air batteries for large-scale energy storage. Solid State Ionics. 2012;216:105–9. [41] Winter M, Besenhard JO, Spahr ME, Novák P. Insertion electrode materials for rechargeable lithium batteries. Adv Mater. 1998;10:725–63. [42] Kirubakaran A, Jain S, Nema RK. A review on fuel cell technologies and power electronic interface. Renew Sustainable Energy Rev. 2009;13:2430–40. [43] Reddy AM, Fichtner M. Batteries based on fluoride shuttle. J Mater Chem. 2011;21:17059–62.

4.3 Nanostructured anode materials Goriparti Subrahmanyam, Ermanno Miele, Remo Proietti Zaccaria and Claudio Capiglia Abstract: Throughout the lithium ion battery (LIB) history, since they were mass produced by Sony in 1991, graphite-based materials have been the anode material of choice. There have been enormous efforts to search for ways of tapping higher energy with alternative anode materials to work in LIBs. Yet, those materials have always This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Subrahmanyam, G., Miele E., Zaccaria, R. P., Capiglia, C. Nanostructured anode materials Physical Sciences Reviews [Online] DOI: 10.1515/psr-2017-0149

4.3 Nanostructured anode materials

139

been subjected to detrimental mechanisms that hinder their applications in LIBs. Will nanotechnology and nanostructured anode materials change the energy storage technologies markedly in the future? Keywords: nanostructuring of anode materials, intercalation/deintercalation, alloy/ de-alloy, conversion anodes

4.3.1 Introduction The research community want to achieve effective energy storage strategies which are the key for the exploitation of alternative energy and thus for the replacement of fossil fuels and traditional energy sources. In this direction, rechargeable lithium ion batteries (LIBs) play a significant role due to their high gravimetric and volumetric energy, high-power density, long cycle life and low self-discharge. Furthermore, they have proved to be the most efficient energy storage strategy in a wide range of portable devices like cellular phones, laptops and digital electronics. However, the employment of LIBs in hybrid electric vehicles (HEVs), plug-in hybrid electric vehicles (PHEVs) and pure electric vehicles (PEVs) need about two to five times more energy density than the present lithium batteries technology can offer (250 Wh/kg). The increase of the energy density of lithium batteries can be achieved by either using high-voltage cathode active materials as electrodes or by developing high-capacity anode and cathode electrode materials. One of the main hindrances to design high-voltage cathode in LIBs is represented by the electrolyte decomposition that occurs at more than 4.2 V vs. Li/Li+. Hence, the development of electrolyte with wider electrochemical window stability is also essential in the realization of the next generation of devices that will operate at higher cell potential with appropriate lithium ion conductivity [1–3]. The state of the art for cathode materials for LIBs is represented by LiCoO2 [and their related LiNiMnCo (NMC), LiNiAlCo (NMA)], LiMn2O4 and LiFePO4, while graphite is definitely the most used anode owing to its excellent features, such as flat and low working potential vs. lithium, low cost and good cycle life. However, graphite allows the intercalation of only one lithium ion with six carbon atoms, with a resulting stoichiometry of LiC6 and thus an equivalent reversible capacity of 372 mAhg−1. In addition, the diffusion rate of lithium into carbon is less than 10–6 cm2 s−1, which results in batteries with low-power density. Hence, there is an urgency to replace graphite anodes to materials with higher capacity, energy and power density. Even though lithium metal holds one of the highest capacity among anode materials (3,860 mAhg−1), safety issues prevent the use of lithium as anode material in secondary batteries. In fact, dendrite formation on the lithium metal can cause short circuit between anode and cathode. Therefore, the path that leads to improved energy and power density of LIB has, as major challenge, the selection of suitable anode materials which can provide high

140

4 Battery Materials

2.0

Potential (V) vs Li/Li+

Titanium oxides

1.5 Oxides (MOx)

1.0 Phosphides (MPx) Nitrides (MNx)

0.5 Porous carbons

0

Sulfides (MSx)

Alloys (Si, Ge, Al, Sn, Sb etc.)

Lithium

Graphite

200

400

600

800

1000 1200 2000

3000

4000

5000

Capacity (Ah/Kg)

Figure 4.3.1: Overview of anode materials for applications in lithium ion batteries (from ref. 4). Reprinted with permission from ref. [4]

capacity and easy diffusion of lithium ions into the anode, along with good cycle life and safety concerns. Accordingly, our group has recently reviewed this subject [4]. Figure 4.3.1 shows the roadmap illustration of active anode materials for today’s used (graphite and titanium oxide) and next generation of lithium batteries. Many efforts have been done in the investigation of both carbon and non-carbon materials for high performances and high-capacity anode in LIBs. A short list must include the following: carbon nanotubes (CNTs) (1,100 mAhg−1), carbon nanofibers (450 mAhg−1), graphene (960 mAhg−1), porous carbon (800–1,100 mAhg−1), SiO (1,600 mAhg−1), silicon (4,200 mAhg−1), germanium (1,600 mAhg−1), tin (994 mAhg−1) and transition metal oxides (500–1,000 mAhg−1). Furthermore, metal sulphides, phosphides and nitrides might be also considered for anode purposes; in fact, they possess a specific capacity higher than 500 mAh g−1. However, high volume expansion, poor electron transport, capacity fading and low coulombic efficiency as well are the main limitations that have to be overcome before they can be used as effective anodes. In this sense, promising results and a bright perspective are offered by nanostructuring the above listed materials. Nanosize and tailored morphology represent the key feature capable of leading these innovative materials from theoretically relevant to an effective technological breakthrough. The expected advantages from using nanotechnology in LIBs can be listed as follows: i. Active materials with high surface to volume ratio. Intensified presence of active sites for lithium storage, hence increasing the specific capacity. Also, high surface area allows a high contact area with electrolyte and leads high lithium ion flux across the electrode/electrolyte interface ii. Electrochemical reactions, impossible to be triggered for bulk materials, can take place at the nanoscale (for example β-MnO2).

4.3 Nanostructured anode materials

141

iii. Improved lithium diffusion due to the reduction of its path length, hence batteries with enhanced power capability. iv. Higher electron transfer rates. In the next sections, we will discuss the state of the art of anode materials for LIBs; for sake of simplicity, we will classify the discussed innovative anode materials into three main groups, depending on their LIB performances and reaction mechanism (see Table 4.3.1): 1. Intercalation/deintercalation materials, such as carbon-based materials, porous carbon, CNTs, graphene, TiO2, Li4Ti5O12, etc.; 2. Alloy/de-alloy materials such as Si, Ge, Sn, Al, Bi, SnO2, etc.; 3. Conversion materials like transition metal oxides (MnxOy, NiO, FexOy, CuO, Cu2O, MoO2, etc.), metal sulphides, metal phosphides, metal nitrides (MxXy; here X = S, P, N).

4.3.2 Intercalation/Deintercalation materials 4.3.2.1 Carbon-based materials Carbon-based materials with various consistencies and morphologies have been recognized as appropriate anode materials for LIBs due to their features, such as ease of availability, stability in thermal and electrochemical environment, low cost and good lithium intercalation and deintercalation reversibility, i. e. the reversible chemical process that leads to lithium ion inclusion in layered carbon structures. A variety of carbon-based materials used as anode in LIBs are classified into two categories, according to the degree of crystallinity and carbon atoms stacking: (i) SOFT carbon (graphitizable carbons) where crystallites are stacked almost in the same direction and (ii) HARD carbon (non-graphitizable carbons) where crystallites have disordered orientation. In particular, the former is quite popular in the battery community. In fact, it shows an appropriate reversible capacity, i. e. (350–370) mAhg−1, long cycle life and good coulombic efficiency (more than 90 %). The reaction mechanism between lithium and graphite, following an intercalation/deintercalation process, has been extensively studied with various analytical techniques [22, 23]. Among the types of commercially available graphite is worth mentioning mesocarbon microbead, mesophase-pitch-based carbon fibre, vapour grown carbon fibre and massive artificial graphite. Despite their massive production and the relative low cost of the industrial processes, these classes of carbon materials have, as major issue, the low specific capacity (i. e. 372 mAhg−1), especially when employed for applications such as HEV, PHEV or PEVs. Hence, the use of graphitic carbon as anode is still limited to low-power devices like mobile phones and laptops. Therefore, further advances in lithium battery anode materials are necessary for the development of competitive performance electric vehicles, smart electric grid systems and, in general, portable

142

4 Battery Materials

Table 4.3.1: Overview of advantages and common issues on anode materials which are used in today’s lithium ion batteries or are under research and development. Active anode material

1. Intercalation/deintercalation materials A. Carbonaceous a. Hard carbons b. CNTS c. Graphene B. Titanium oxides a. LiTi4O5 b. TiO2

Theoretical capacity (mAh g-1) [Reference]

Common issues

– Good working potential – Low coulombic efficiency 1,116 [8, 9, 24, 25] – Low cost – High-voltage hysteresis 960 [10] – Good safety – High irreversible capacity 200–600 [5–7]

175 [11] 330 [11]

2. Al loy/de-alloy materials a. Silicon 4,212 [12] b. Germanium c. Tin d. Antimony e. Tin oxide f. SiO 3. Conversion materials a. Metal oxides( Fe2O3, Fe3O4, CoO, Co3O4, MnxOy, Cu2O/CuO, NiO, Cr2O3, RuO2, MoO2/ MoO3 etc.) b. Metal phosphides/ sulphides/nitrides (MXy; M= Fe, Mn, Ni, Cu, Co etc. and X= P, S, N)

Advantages

1,624 [13] 993 [14] 660 [15] 790 [16] 1,600 [31] 500–1,200 [6, 16–19]

– Extreme safety – Good cycle life – Low cost – High-power capability

– Very low capacity – Low energy density

– Higher specific capacities – High energy density – Good safety

– Large irreversible capacity – Huge capacity fading – Poor cycling

– High capacity

– Low coulumbic efficiency

500–1,800 [19–21] – High energy

– Low cost – Environmentally compatibility – High specific capacity

– Unstable SEI formation

– Large potential hysteresis – Poor cycle life

– Poor capacity retention – Low operation potential – Short cycle life and Low polarization – High cost of than counteroxides production

4.3 Nanostructured anode materials

143

high-power-consuming devices. One of the possibilities is the use of both soft and hard carbon materials, because of their high theoretical capacity. Presently, the research activity is focusing on CNTs, nanofibers and graphene as the most promising carbon based anode materials. The reduction in size and the unique shape of these structures introduce novel properties that will substantially improve the energy storage capacity in lithium batteries. For example, very recently, carbon nanorings (CNRs) with 20 nm outer diameters and 3.5 nm wall thicknesses, have shown outstanding performances as anode-active materials in lithium battery: high lithium uptake and larger capacity, i. e. more than 1,200 mAhg−1, and over a hundreds of cycles at the current density of 0.4 Ag−1. Even at the higher current rates of 45 Ag−1, it has been observed a capacity as high as 500 mAhg−1. The larger specific capacity and the high rate capability have been rationalized as due to the reduction of the diffusion distance and to the increased number of storage sites for lithium ions. These advantages are paradigmatic results of the developing of nanosized carbon-based materials. 4.3.2.1.1 Hard carbon Even though soft carbons are now the state of art for anode materials in LIBs providing an interesting theoretical capacity, some issues, especially the high-voltage hysteresis during the de-lithiation process and low capacity are hinders their use as anode materials. Instead, hard carbons have high reversible capacity (more than 500 mAg−1) in the potential range between (0–1.5) V vs. Li/Li+, and thus they represent a valid alternative to soft carbons. They were firstly developed in 1991 by Kureha Corporation, Japan. Recently, several groups have reported reversible capacity between (200–600) mAhg−1 using hard carbon as anode. These results are related to the porous nature of hard carbon and to the high number of graphene sheets with high surface area. However, hard carbons as anode have as drawbacks the low initial coulombic efficiency and the low tap density. Surface oxidation, fluorination and thin soft carbon or metal coating is some strategies pursued in order to improve the performance of hard carbons as anodes. Amongst, thin soft carbon layer surface coating results in good capacity as well as in improvement of the coulombic efficiency. Interesting results in this direction have been obtained through a systematic investigation on the effect of temperature on the carbon coating process. However, it is still needed to improve the capacity and the cycle life of hard carbon anodes. 4.3.2.1.2 Carbon nanotubes CNTs are highly ordered carbon nanostructures, realized through a self-assembly unidirectional growth process. CNTs can be classified in single-walled carbon nanotubes (SWCNTs) and multiwalled carbon nanotubes (MWCNTs) depending on the thickness and on the number of coaxial layers. Since their discovery in 1991, both SWCNTs and MWCNTs were extensively investigated both as anode material and as a composite, when used together with other active anode materials, owing to their superior electronic conductivity, good mechanical and thermal stability, adsorption

144

4 Battery Materials

and transport properties. The theoretical reversible capacity of CNTs has been evaluated to be 1,116 mAhg−1 for SWCNTs in LiC2 stoichiometry [24, 25]. This impressive value is by far the highest capacity among carbon-based active materials, and it has been attributed to the intercalation of lithium into stable sites located on the surface of pseudo-graphitic layers and inside the central tube as well. The research activity in CNTs as anodes is intense: several attempts were made with a variety of synthesis protocol and pretreatments such as acid treatment and ball milling surface modification. However, the achievement of high coulombic efficiency with CNTs remains challenging because of the presence of large structure defects and high-voltage hysteresis. In order to overcome these issues, research has been focused on the morphological features of CNTs, such as wall thickness, tube diameter, porosity and shape (bamboo shape, quadrangular cross-section, etc.). In order to improve the lithium storage capacity and cycle life in batteries, many researchers have been synthesising CNTs conjugated with a variety of nanostructured materials (Si, Ge, Sn, Sn-Sb) or metal oxides (MxOy; M = Fe, Mn, Ni, Mo, Mo, Cu, Cr). These hybrid systems result in CNTs with enhanced electrical conductivity and reduced volume changes during the charging and discharging processes. 4.3.2.1.3 Graphene Graphene consists in a honeycomb network of sp2 carbons bonded into two-dimensional sheets with nanometre thickness (single-atom thickness). Since its discovery in 2004, graphene has drawn much attention because of its admirable properties and versatility in a number of fields such as chemical, physical, biological and engineering sciences. Among its astonishing properties, we can recall good electrical conductivity, relevant mechanical strength, high values of charge mobility and high surface area (2,630 m2g−1) which make graphene a suitable material for electrodes in LIBs. In addition, it is theoretically proven that graphene can accommodate lithium ions on both sides of its surfaces which results in twice higher capacity than other carbon-based materials. Presently, the mainstream research activity suggests various hybrid graphene/ metals or semiconductors and graphene/metal oxides/phosphides as anode suitable candidates in LIBs. For instance, recently proposed hybrid systems are 2–3 nm SnO2 particles / nitrogen-doped graphene (1,220 mAhg−1 gravimetric capacity over 100 cycles) and silicon–carbon nanocables/reduced graphene oxide sheets. 4.3.2.2 Titanium-based oxides Titanium-based oxides have drawn significant interest among the lithium battery community because they allow the designing of operational devices with minor safety concern. Moreover, this class of active materials show other suitable features such as inexpensiveness, low toxicity, low volume change of (2–3)% on both lithium intercalation and deintercalation together with an excellent cycle life. However, they have low theoretical capacity, i. e. (175–m330) mAhg−1) and low electronic conductivity.

4.3 Nanostructured anode materials

145

The electrochemical performance and the lithium intercalation/deintercalation in titanium-based oxides mainly depend on their structure, morphology and size. In particular, it has been found that nanostructured titanium oxides lead to better capacity, longer cycle life and higher rate capability than the bulk materials. Hence, many researchers engaged in the development of anode materials based on titanium oxides. 4.3.2.2.1 Spinel Li4Ti5O12 Spinel Li4Ti5O12 (LTO) is considered the most appropriate titanium-based oxide for lithium storage because it exhibits excellent lithium ion reversibility at high operating potential (1.55 V vs. Li/Li+). Lithium intercalation/deintercalation in LTO occurs by the lithiation of spinel Li4Ti5O12 yielding rock salt type Li7Ti5O12. During the intercalation process, the spinel symmetry and structure remain almost unaltered. The high operating potential leads to extra safety characteristics: the formation of solid electrolyte interface (SEI) is mitigated and the development of dendrites, typical issue in carbon-based anodes, is avoided. However, low theoretical capacity (175 mAhg−1) and low electronic conductivity (1,600 mAhg−1). The main advantage of SiO is the ability to provide an electrochemical lithium alloying process which is an alternative to silicon. In particular, lithium oxygen coordination provides minimal volume change and, at the same time, lowers activation energy. Solid SiO is thermodynamically unstable at all temperatures. Therefore, it can be transformed into Si and SiO2 in a temperaturetriggered disproportionation reaction. As previously pointed out about silicon, SiO undergoes consistent changes in the volume during the lithiation and de-lithiation processes. To solve these problems and to improve the reversible capacity as well as the cycle retention, different approaches have been tested. The most promising are carbon coating, electrochemical reduction by lithium and synthesis of SiO nanoparticles. Among these, particle size reduction combined with carbon coating is an

150

4 Battery Materials

effective way to overcome the above issues by shortening the diffusion distance of lithium ions and thus to enhance the ionic and the electrical conductivities. In 2002, the author of this chapter demonstrated the effect of both the oxygen concentration and particle size on the cyclic life and the reversible capacity of SiOx [31]. The discharge capacities for 50 nm size particles of SiO0.8 and SiO1.1 were found to be 1,700 mAhg−1 and 750 mAhg−1, respectively. The capacity retention of SiO0.8 and SiO1.1 were 40 % and 93 % of the initial discharge capacity after 25 cycles, respectively. It turned out that lower oxygen content in SiOx allows for higher specific capacity associated to poorer cycle life. In addition to oxygen concentration, also particle size affects the anode performance: 30 nm size of SiO particles showed better capacity retention with high reversible capacity compared to SiO larger particles. Finally, it is interesting to note that in order to understand the important electrochemical reaction mechanism between SiO and lithium, various analytical techniques have been employed: X-ray photoelectron spectroscopy, NMR, highresolution TEM and electrochemical dilatometry, just to name a few.

4.3.3.3 Germanium Germanium (Ge) is an extensively studied active anode material also owing to its high lithium storage capability (1,623 mAhg−1) with Li22Ge5 as equivalent stoichiometry and reversible alloy/de-alloy reactions. Even though Ge is considerably more expensive and has lower capacity than silicon, it still has desirable advantages such as a narrow band gap (0.67 eV), a high intrinsic electron conductivity (104 times higher than silicon) and a higher capacity than graphite anode. This results in a lithium diffusion into Ge faster than in silicon which ensues higher rate capability and more efficient charge transport than silicon and graphite as well. Ge high-power capability is extremely important in advanced high-power density applications such as electric vehicles. However, as discussed for silicon, the practical usage of Ge as active electrode in LIB is also hindered by the dramatic volume change (~300 %) during lithium intercalation/deintercalation. Ge nanostructures, such as nanoparticles, nanowires and nanotubes, can effectively sustain the volume change and alleviate the pulverization in cycling more than bulk and microstructures. Further improvements have been observed with hybrid composite of Ge nanoparticles using conductive matrices, obtained through simple preparation routes, for example solid-state pyrolysis. Ge nanoparticles, with diameter between 5 nm and 20 nm, were encapsulated inside carbon nanospheres with diameters ranging from 50 nm to 70 nm. These obtained composite nanospheres exhibited highly reversible lithium storage along with high rate capability. In these systems, lithiation/de-lithiation happens through carbon that mitigates the volume changes and also avoids direct contact with the electrolyte. Similar behaviour has been observed in Ge nanoparticles deposited on SWCNTs by CVD.

4.3 Nanostructured anode materials

151

To improve the electrochemical performances of Ge, our group have decided to follow the approach of reducing the Ge active material size to the nanoscale. Furthermore, nanosize particles usually suffer from aggregation effect which manifests similar behaviour as bulk materials. Therefore, to reduce and prevent the aggregation of nanoparticles, we have used MWCNT composite. A simple approach via solvothermal method was used to produce Ge nanocrystals, with a diameter in the 4–10 nm range, and their Ge-MWCNTs. The resulting Ge-MWCNTs nanocomposite showed an improved cycling performance with higher capacity retention compared to pure Ge electrode and a discharge capacity of ~1,170 mAhg−1 with good capacity retention [32, 33]. Figure 4.3.5 shows the as-prepared Ge nanocrystal particles on the surface of MWCNTs. Further work has been done by our group to improve the performance of Ge as active material for LiB, and the results will be published soon. Ge-MWCNTs / dark field

Ge-MWCNTs

(a)

(b) 0.326 nm

5 nm

100 nm

(111)

10 nm (c)

Ge

de-lithiation lithiation

de-lithiation lithiation

60

20 1000

Coulombic efficiency

2000

de-lithiation lithiation

MWCNTs

3000

100 Ge-MWCNTs

Specific capacity (mAh g–1)

4000

0

0 0

10

20

30 Cycle number

40

50

60

Figure 4.3.5: Transmission electron microscopy images of (a) dark-field image of Ge–MWCNTs, (b) highresolution image of Ge nanocrystals on the surface of MWCNTs. The inset in (b) illustrates the lattice fringes and crystal structure of Ge and (c) galvanostatic charge–discharge capacities of Ge, Ge–MWCNTs, MWCNTs along with coulombic efficiencies (half-filled dots) at a current rate of 0.1C (from ref. 33).

4.3.3.4 Tin oxide Tin oxide (SnO2) was first developed by Fuji Photo Film corporation and received significant attention as anode in LIBs due to low work potential, i. e. 0.6 V vs. Li/Li+

152

4 Battery Materials

and high theoretical capacity. The electrochemical lithium alloying reactions can be summarized in a first partially irreversible step where SnO2 is reduced into Sn and lithium oxide, which is followed by the reversible Sn-lithium alloying/de-alloying reaction. The overall electrochemical process involues 8.4Li for one SnO2 formula unit with a corresponding theoretical capacity of 1,491mAhg−1, but it is reduced to 783 mAhg−1 when just the second reversible step with 4.4Li for SnO2 unit is taken into account. Furthermore, severe electrode degradation is observed because of consistent volume change (\gt200 %) upon cycling. Much attention has been paid to improve the cycle stability of SnO2 and to reduce the irreversible capacity caused by volume change. Porous nanostructures, nanocomposites, and hollow nanostructured SnO2 have been also proposed. Compared to bulk and SnO2 particles, it is possible to identify optimal pore size in the nanostructures capable to balance the volume changes during the lithium intercalation/deintercalation. Therein, the pores act as buffer for the large volume changes. Further developments were obtained by using composite carbon-based materials, such as carbon-coated SnO2, SnO2/nanofibers, SnO2/carbon nanoparticles, SnO2/ CNTs or SnO2/graphene.

4.3.4 Conversion materials In this section, we will provide an overview on the transition metal compounds such as oxides, phosphides, sulphides and nitrides (MxNy; M = Fe, Co, Cu, Mn, Ni and N = O, P, S and N) when utilized as anodes in the field of LIBs. Those materials were first proposed by Tarascon et al. in 2000 [34]. The electrochemical reaction mechanism involving these compounds and lithium implies the reduction and the oxidation of transition metals along with the composition and the decomposition of lithium compounds (LixNy; here N = O, P, S and N), respectively. Anodes based on these compounds exhibit high reversible capacities of (500–1,000) mAhg−1 owing to the fully utilization of the metal by the participation of a high number of electrons in the conversion reactions.

4.3.4.1 Iron oxide Iron-based oxides have been extensively studied for rechargeable lithium batteries because of their low cost, non-toxicity and high natural abundance. Iron oxides, both hematite (α-Fe2O3) and magnetite (Fe3O4), are capable of participating in reversible conversion reactions with lithium, with corresponding theoretical capacity of 1,007 mAhg−1 and 926 mAhg−1, respectively. However, iron oxides tend to exhibit poor cycling performance due to low electrical conductivity, low diffusion of lithium ions, expansion in volume and iron aggregation during

4.3 Nanostructured anode materials

153

charging–discharging. Therefore, many recent investigations have been focused on developing new methods for the preparation of iron oxide nanomaterials as well as modification of their size, shape and porosity. Further studies have been focused on methods to stabilize their structure and to improve the electrochemical kinetics and power capability, mainly by carbon coating or composites of α-Fe2O3 and Fe3O4.

4.3.4.2 Cobalt oxides In this section, we consider recent advances on cobalt oxide materials (Co3O4 and CoO) as anode for LIBs. Co3O4 and CoO show theoretical capacities of 890 mAhg−1 and 715 mAhg−1, respectively. A number of different forms for cobalt oxides have been reported. Porous nanostructures, nanosheets, nanocubes, nanowires and nanotubes have been prepared by various synthetic routes such as wet chemical, solid state, hydrothermal and microwave that allowed the tailoring of the electrochemical performance. Composites of cobalt oxides have been studied to buffer the volume changes and to prevent the detachment and the agglomeration of cobalt oxides during the charge–discharge process. Carbon-based composites, especially graphene, are intensively studied due to their desirable properties. Various other metal oxides, showing a conversion reaction mechanism with lithium, have also been studied. For example, NiO, MnOx, CuOx, MoOx and CrOx were extensively investigated as potential anode materials for LIBs and they showed miscellaneous physicochemical properties and large reversible capacities of around 500 mAhg−1

4.3.4.3 Metal phosphides Metal phosphides (MPx) have also been extensively studied as an anode for rechargeable lithium batteries. They can react with lithium both in a conversion reaction schema and in an intercalation/deintercalation reaction, depending on the nature of the transition metal and phosphorous-bonding stability upon electrochemical environment. It is possible to divide MPx into two groups. The first one involves the lithium intercalation/deintercalation without breaking the metal–phosphorous bond, known as insertion/de-insertion mechanism. The second group involves a conversion reaction mechanism. In such electrochemical reactions, the bonds between metal and phosphorous are broken, resulting in nanosized metal particles and Liphosphides. Copper-, cobalt-, iron-, nickel- and tin-based phosphides are usually considered to belong to the second group, i. e. conversion mechanism. Nevertheless, some reports showed that MPx could participate to both the intercalation and the conversion mechanism with respective to potential vs. Li/Li+. Metal phosphides can deliver high capacities between 500 and 1,800 mAhg−1. In addition, MPx as anodes show high

154

4 Battery Materials

degree of electron delocalization, which leads to low formal oxidation state and strong covalent bond of M-phosphorous. Another advantage is the lower intercalation potential of MPx than their oxide counterparts. However, MPx usually have low electrical conductivity and high volume changes upon charge–discharge cycling. The use of MPx as anodes deserves further exploration to overcome these drawbacks.

4.3.4.4 Metal sulphides and nitrides Another kind of materials which have been thoroughly investigated for anode applications in the LIBs field are the transition metal sulphides (MSx) and nitrides (MNx). Iron, molybdenum, tin, antimony, nickel, cobalt and tungsten have attracted major interest due to their high lithium storage capacity and structural advantages during the lithium intercalation/deintercalation process. As described in the previous sections, MSx and MNx reaction mechanisms with lithium involve the reduction to the metal and the formation of lithium–sulphur or lithium–nitride, respectively, through conversion reaction. In particular, CoS2 and CoS are the most promising composites exhibiting superior electrochemical performance with reversible larger capacities of 929 mAhg−1 and 835 mAhg−1 after 10 cycles, respectively. Besides metal sulphides, also metal nitrides have emerged as promising anodes for LIBs. Among them, Cu3N, VN, Co3N, CrN, Fe3N, Mn4N and Ni3N have especially been studied.

4.3.5 Conclusions The technological development of LIBs relies on the bright perspectives offered by nanoscience and nanotechnology. Novel techniques and deep insight into material science have offered the opportunity to design and synthesize appropriate nanostructured anode materials for the next generation of LIBs. Uniquely tailored physical and chemical properties of active anode materials allow achieving high lithium storage, high lithium ion flux at the interface, lower diffusion length for both lithium ions and electrons and minimal volume change during the charging/discharging. These exquisite features combined together led to high energy density and power devices. In particular, the advancement in material science will permit a better understanding of electrochemistry and operating mechanisms underlying the battery cell on anode, cathode and electrolyte sides with deeper insights down to the nanoscale and material structures. These advancements together with improved nanomanufacturing will allow designing and assembling the best combination of cell part depending on the application target. Nanomanufacturing is also promising a better integration of power supply into electronic devices. For instance, stationary batteries for electric grid stabilization will need high energy and fast response but no constrains on power densities. Such desirable features can

References

155

be offered by cell based on LTO- or TiO2-based anode material. Conversely, EV applications will need the higher energy densities offered by the utilization of high-energy density materials such as silicon (or Sn, Ge based) and/or their composite anode materials. However, to ensure the these anode materials towards high-energy density and long life in practical LIB applications, future research should focus on fundamental problems associated with nanostructuring of anode materials that are minimizing the anode/electrolyte reaction by surface protection and/or pre-lithiation of anode, and also increasing the tap density with various manufacturing techniques and, most importantly, reducing the cost of manufacturing technology towards large-scale production of nanostructured electrode materials.

References [1] [2] [3] [4]

[5]

[6] [7]

[8] [9] [10]

[11] [12] [13]

[14]

Schalkwijk WV, Scrosati B. Advances in lithium-ion batteries. Springer: Berlin, 2007. Yoshio M, Brodd RJ, Kozawa A. Lithium-ion batteries: science and technologies. Berlin: Springer, 2010 Xia X, Liu YH, Zhang J. Lithium-ion batteries: advanced materials and technologies. New York: CRC Press, 2016. Goriparti S, Miele E, De Angelis F, Di Fabrizio E, Proietti Zaccaria R, Capiglia C. Review on recent progress of nanostructured anode materials for Li-ion batteries. J Power Sources. 2014;257:421–3. Fujimoto H, Tokumitsu K, Mabuchi A, Chinnasamy N, Kasuh T. The anode performance of the hard carbon for the lithium ion battery derived from the oxygen-containing aromatic precursors. J Power Sources. 2010;195:7452–6. Yang J, Zhou X-Y, Li J, Zou Y-L, Tang J-J. Study of nano-porous hard carbons as anode materials for lithium ion batteries. Mater Chem Phys. 2012;135:445–50. Bridges CA, Sun X-G, Zhao J, Paranthaman MP, Dai S. In situ observation of solid electrolyte interphase formation in ordered mesoporous hard carbon by small-angle neutron scattering. J Phys Chem C. 2012;116:7701–11. Meunier V, Kephart J, Roland C, Bernholc J, Initio A. Phys Rev Lett. 2002;88:075506. Nishidate K, Hasegawa M. Energetics of lithium ion adsorption on defective carbon nanotubes. Phys Rev B. 2005;71:245418. Hou J, Shao Y, Ellis MW, Moore RB, Yi B. Graphene-based electrochemical energy conversion and storage: fuel cells, supercapacitors and lithium ion batteries. Phys Chem Chem Phys. 2011;13:15384–402. Chen Z, Belharouak I, Sun YK, Amine K. Titanium-based anode materials for safe lithium-ion batteries. Adv Funct Mater. 2013;23:959–69. Szczech JR, Jin S. Nanostructured silicon for high capacity lithium battery anodes. Energy & Env Sci. 2011;4:56–72. Rudawski NG, Yates BR, Holzworth MR, Jones KS, Elliman RG, Volinsky AA. Ion beam-mixed ge electrodes for high capacity li rechargeable batteries. J Power Sources. 2013;223: 336–40. Bruce PG, Scrosati B, Tarascon J-M. Nanomaterials for rechargeable lithium batteries. Angew Chem Int Ed. 2008;47:2930–46.

156

4 Battery Materials

[15] Park C-M, Kim J-H, Kim H, Sohn H-J. Li-alloy based anode materials for Li secondary batteries. Chem Soc Rev. 2010;39:3115–41. [16] Wang Z, Zhou L, Lou XW. Metal oxide hollow nanostructures for lithium-ion batteries. Advanced Mater. 2012;24:1903–11. [17] Jiang J, Li Y, Liu J, Huang X, Yuan C, Lou XW. Recent advances in metal oxide-based electrode architecture design for electrochemical energy storage. Advanced Mater. 2012;24:5166–80. [18] Prosini PP, Carewska M, Loreti S, Minarini C, Passerini S. Lithium iron oxide as alternative anode for li-ion batteries. Int J Inorg Mater. 2000;2:365–70. [19] Ji L, Lin Z, Alcoutlabi M, Zhang X. Recent developments in nanostructured anode materials for rechargeable lithium-ion batteries. Energy & Env Sci. 2011;4:2682–99. [20] Lai C-H, Lu M-Y, Chen L-J. Metal sulfide nanostructures: synthesis, properties and applications in energy conversion and storage. J Mater Chem. 2012;22:19–30. [21] Boyanov S, Annou K, Villevieille C, Pelosi M, Zitoun D, Monconduit L. Nanostructured transition metal phosphide as negative electrode for lithium-ion batteries. Ionics. 2008;14:183–90. [22] Inaba M, Yoshida H, Ogumi Z, Abe T, Mizutani Y, Asano M. In situ Raman study on electrochemical li intercalation into graphite. J Electrochem Soc. 1995;142:20–6. [23] Whitehead AH, Edström K, Rao N, Owen JR. In situ X-ray diffraction studies of a graphite-based Li-ion battery negative electrode. J Power Sources. 1996;63:41–5. [24] Schauerman CM, Ganter MJ, Gaustad G, Babbitt CW, Raffaelle RP, Landi BJ. Recycling single-wall carbon nanotube anodes from lithium ion batteries. J Mater Chem. 2012;22: 12008–15. [25] Zhao J, Buldum A, Han J, Ping Lu J. First-principles study of li-intercalated carbon nanotube ropes. Phys Rev Lett. 2000;85:1706–9. [26] Goriparti S, Miele E, Prato M, Scarpellini A, Marras S, Monaco S, et al. Direct synthesis Of carbon-doped Tio2–bronze nanowires As anode materials For high performance lithium-ion batteries. ACS Appl Mater Interfaces. 2015;7:25139–46 [27] Claudio C, Remo Proietti Z, Subrahmanyam G, Miele E, De Angelis F. Direct synthesis of carbondoped TiO2–bronze nanostructures as anode materials for high performance lithium batteries. PCT WO 2017/060407 A1 [28] Kasavajjula U, Wang C, Appleby AJ. Nano- and bulk-silicon-based insertion anodes for lithiumion secondary cells. J Power Sources. 2007;163:1003–9. [29] Chan CK, Peng H, Liu G, McIlwrath K, Zhang XF, Huggins RA, et al. High-performance lithium battery anodes using silicon nanowires. Nat Nanotechnol. 2007;3:31. [30] Miele E, Goriparti S, Messina GC, Prato M, Ansaldo A, Barone A, et al. Porous silicon as nanostructured anode material for lithium ion batteries. ECS Trans. 2014;62:25–34. [31] Yang J, Takeda Y, Imanishi N, Capiglia C, Xie JY, Yamamoto O. SiOx-based anodes for secondary lithium batteries. Solid State Ionics. 2002;152–153:125–9. [32] Goriparti S, Miele E, Scarpellini A, Marras S, Prato M, Ansaldo A, et al. Germanium nanocrystalsMWCNTs composites as anode materials for lithium ion batteries. ECS Trans. 2014;62: 19–24. [33] Goriparti S, Gulzar U, Miele E, Palazon F, Scarpellini A, Marras S, et al. Facile synthesis of GeMWCNT nanocomposite electrodes for high capacity lithium ion batteries. J Mater Chem. 2017;5:19721–8. [34] Poizot P, Laruelle S, Grugeon S, Dupont L, Tarascon JM. Nano-sized transition-metal oxides as negative-electrode materials for lithium-ion batteries. Nature. 2000;407:496.

4.4 Modification of cathode materials for Li batteries

157

4.4 Modification of cathode materials for Li batteries Tina Weigel and Florian Schipper Abstract: Beside the development of novel cathode materials with higher capacity and cyclability, the optimization of commercial used materials are necessary as well. Therefore the materials, as spinels, olivines or layered oxides can be modified to improve their electrochemical performance. Two promising technologies for material modification are doping or coating. Both approaches are introduced and explain by examples. Keywords: cathode material modification, coating, doping

4.4.1 Introduction The main materials for commercial used Li battery cathodes are spinels (LiM2O4, where M is a transition metal) with a Li diffusion in three dimension, olivines (LiMPO4) with one-dimensional Li diffusion and layered transition metal oxides (LiMO2) with a diffusion of Li in two dimensions (see Figure 4.4.1) [1]. Modification

b

b c

a

a)

c

a

b) Li M O P

c a

b

c) Figure 4.4.1: Illustration of the crystal structure of the main cathode materials for Li ion batteries: (a) LiM2O4 spinel structure (3D material), (b) LiMPO4 with olivine structure (1D material) and (c) LiMO2 with a layered structure (2D material). This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Weigel, T., Schipper, F. Modification of cathode materials for Li batteries Physical Sciences Reviews [Online] DOI: 10.1515/psr-2018-0037

158

4 Battery Materials

of the cathode materials improves properties such as capacity, cycle life, thermal stability and safety [2, 3]. A distinction between modifications of the crystal lattice by doping as well as by the crystal surface via coating can be drawn.

4.4.2 Surface coating Surface coating (1) induces improvement of the ion transfer on the surface, (2) improves the electrochemical performance, for example rate capacity, charge and discharge behavior or cycling stability and (3) reduces side reaction with the electrolyte [3, 4]. For coating, materials like carbon [5–9], metal oxides [10–17], metal carbonates [18], metal aluminates [19], metal fluorides [16, 20], metal phosphates [21, 22], metal oxyflourides [23], metal hydroxides [24], silica [25] and lithium borate glass [26, 27] can be used. As an example, coating with carbon shows an improvement of the electrical performance for olivines. The charge transfer and electronic conductivity are increased as well as extra electron-conducting pathways are provided [5–8]. Disadvantageously, during the thermal treatment of the carbon coating, reductive environments can be generated which reduce especially transition metal ions and have a negative effect of their electrochemical performance. Therefore, for cathodes based on lithium transition metal oxides carbon coating can be only used with special coating techniques [9, 28, 19]. Furthermore, coating layered cathode materials with lithium borate glass allows for faster Li conductivity, improved discharge capacity and better cycling performance [26]. Coating of metal oxide layers works as a protection layer of the cathode material. The active material is shielded from direct contact with the electrolyte. This leads to a suppression of side reactions with the electrolyte and hence improved electrochemical performance, cycle life and thermal stability as well as a suppression of phase transitions [3, 4, 30]. The coating can be performed by mechano-thermal treatment [25], sol-gel methods [9, 14], chemical vapor deposition and especially atomic layer deposition (ALD) [15, 17, 31]. Mechano-thermal treatment induces very rough coatings, with areas of higher and lower amount of the coating material [10, 32]. With ALD techniques ultrathin, pin-hole free coating layers can be generated [33]. Thicknesses in a range from single atom layer of about 1 Å to a few nm can be realized. This is suitable for cathode material coating layers, since the thin layer of approximately ten ALD cycles do not hinder the Li diffusion [15, 17, 31, 34]. Thinner layers do not fully cover the cathode material and thicker ones reduce the ion diffusion [15]. Figure 4.4.2 shows a comparison of the average discharge capacity of an uncoated and with ten cycles ALD (around 1 nm) coated layered transition metal oxide. An improvement of the longterm stability and higher discharge capacities are clearly visible. To achieve high stability and electrochemical performance, so called core-shell structures are investigated [3, 35, 36]. Herein, a core material with high capacity (e.g.

Average discharge capacity C (mAhg-1)

4.4 Modification of cathode materials for Li batteries

220

159

C/15 C/10 C/3 C/10 C/3

200 180 160 0.8C1C 2C 4C 140 120

10 cycles ALD coating at 30°C uncoated at 30°C

100 0

20

40

60

80

100

Cycle number Figure 4.4.2: Comparison of the average discharge capacity measured by different C-rate tests of a coated (ten cycles ALD coating, thickness is around 1 nm) and uncoated layered lithium transition metal oxide at 30 °C. The results show higher average discharge capacities and a better long-term stability (lesser decrease during the C/3 test over 80 cycles).

LiNi0.8Co0.1Mn0.1O2) is completely covered with a thick layer of a low reactive and stable shell material (e.g. LiMn2O2). The shell protects the core material and improves the capacity and cycle life [35, 36]. To reduce strain on the core shell interface, a concentration gradient strategy with different shell layers of various material concentration has been used [36, 37]. A further approach used a Li[Ni0.54Co0.12Mn0.34]O2 core material and Li-spinel structure as shell material [38]. This structure shows excellent rate capacity and thermal stability. Therefore coating morphologies can be distinguished in rough, core-shell and ultra-thin coating, which is shown in Figure 4.4.3. The here discussed coating methods can be applied for other battery types, like Na, Mg or Al ion batteries, as well [39]. 4.4.3 Doping Doping is the integration of extrinsic metal ions into the crystal lattice of the cathode material, which enhance the structural stability as well as the electrochemical performance [2, 30]. The interpretation of doping effects can be complicated due to the interrelation between doping, microstructure and morphology [2]. Commonly used dopants are Al [40–45], Mg [41, 46–48], Fe [45, 49, 50], Zr [48, 51 – 53], Ti [48, 54 –57], Ga [58, 59], F [60–62] and Cr [41, 63– 67]. Methods for doping are conventional solid-state reactions [49, 6], wet-chemical [45, 56, 58, 64, 67] or sol-gel methods [40, 63, 65].

160

4 Battery Materials

rough coating

thin film coating

core-shell

Figure 4.4.3: Morphology of different surface coatings. A rough coating is characterized by a not completely covered surface. Whereby thin film coatings cover the surface by a thin layer completely. This thin layer prevents the cathode material for reaction with the electrolyte, but does not hinder Li diffusion. The core-shell structure is coated with a thick layer and combines the good stability behavior of the shell material as well as the good electrochemical performance of the core material. Adapted from Ref. [3] with permission from The Royal Society of Chemistry.

For example, Fe [49] or Cr [63] doping of Li3V2(PO4)3 olivine cathode materials improves the electrochemical performance. Fe doped material, with a small dopant concentration from 2% to 4%, shows an enhancement of retention rate of discharge capacity of about 71%, in comparison to only 58% for the undoped material. This results from the enhanced conductivity and structural stability induced by Fe doping [49]. Similar results can be obtained by Cr doping. Here, the specific capacity can be held at 81% for 1% Cr doping, in comparison to only 66% for the undoped material after 100 cycles at 2C [63]. It has to be noted that large concentrations of dopant can induce changes in the crystal structure and structural instabilities, which impairs the electrochemical performance [49, 63]. Figure 4.4.4 shows the discharge capacity for an Fe-doped Li3FexV2−x(PO4)3 cathode material with different dopant concentrations. Furthermore, Cr stabilizes the spinel structure of spinel-based cathode materials, due to the larger Mn3+ ion, which is substituted by a smaller Cr3+ ion. This reduces the M3+O6 octahedron distortion [65]. Another example is doping with F, which stabilizes the structure of layered cathode materials [61], suppresses the layered-to-spinelphase transition and enhances the stability of the electrode-electrolyte surface [62]. Furthermore, various other dopants have different impacts on the material. For example in Ni-rich-layered transition metal oxides (LiNi0.8Co0.1Mn0.1O2), Ti has a stabilization effect, because the Ti4+ ion can substitute the Co3+ ion. To compensate this, Mn is reduced from valence state Mn4+ to Mn3+. Additionally, a Ti doping prevents the migration of Ni2+ from the transition metal site to the Li site [54]. Mg substitute Li, which has a similar ion radius as Mg. This so-called “stable pillar effect” improves the structural and cycling stability [46]. Doping with Zr shows improved electrochemical performance because Zr ions occupy the Li site in the layered structure and destabilize the Ni in the Li site as well as reduce the concentration of “Jahn–Teller” active Ni3+ ions [51].

4.4 Modification of cathode materials for Li batteries

161

180

Discharge capacity (mAhg-1)

160 140 120 100 a 80

b c d e

60 40 0

20

40

60

80

Cycle number Figure 4.4.4: Discharge capacity of different amounts of Fe doping for Li3FexV2−x(PO4)3, with x = 0.00 (a), x = 0.01 (b), x = 0.02 (c), x = 0.04 (d) and x = 0.06 (e). The figure shows that doping with amounts of x = 0.02–0.04 improves the discharge capacity. Small amounts of Fe have only a slight effect on the electrochemical performance (b), but too high concentrations induce structural changes and phase instabilities, which has a negative influence on the discharge capacity (e). Redrawn from publication Ren, Manman et al. “Preparation and electrochemical studies of Fe-doped Li3V2(PO4)3 cathode materials for lithium-ion batteries”, Journal of Power Sources 162 (2006): 1357 – 1362, Copyright 2017, with permission from Elsevier.

A co-doping combines the advantages of two different dopants. As an example, a doping of F, together with Al, results in a suppression of the layered-to-spinel-phase transition of layered cathode materials as well as reduces capacity fading and voltage decay [68]. Furthermore, a combination of both modification techniques can be used, as well. Hereby, the cathode material is first doped with metal ions, and second, the electrode is coated [39]. Otherwise, the cathode material was coated and ensuing annealed. Hereby ions diffuse into the crystal lattice and dope the material. A thin layer on the surface still remains unchanged [69]. This results in an enhanced cycling stability and electrochemical performance.

4.4.4 Summary This overview shows that purposeful modifications of the cathode material for lithium ion batteries optimize the electrochemical behavior. Hereby, an additional layer is directly coated on the cathode surface (surface coating) or extrinsic ions are

162

4 Battery Materials

implemented in the material (doping). Thus the electrochemical performance can be improved and higher structural, thermal or cycling stability is guaranteed. Different methods of coating as well as different kinds of dopants are well investigated and proved for standard cathode materials for lithium ion batteries, such as spinel, olivine and layered transition oxides. The here shown modification approaches can be combined and used for post-lithium battery technologies as well. Funding: TW would like to thank Federmann Enterprises LTD for financial support by the Federmann Scholarship.

References [1]

Etacheri V, Marom R, Elazari R, Salitra G, Aurbach D. Challenges in the development of advanced Li-Ion batteries: a review. Energy Environ Sci. 2011;4:3243–3. DOI: 10.1039/ C1EE01598B. [2] Fergus Jeffrey W. Recent developments in cathode materials for lithium ion batteries. J Power Sources. 2010;195:939–54. DOI: 10.1016/j.jpowsour.2009.08.089. [3] Chen Z, Qin Y, Amine K, Sun YK. Role of surface coating on cathode materials for Lithium-Ion batteries. J Mater Chem. 2010;20:7606–6. DOI: 10.1039/c0jm00154f. [4] Li C, Zhang HP, Fu LJ, Liu H, Wu YP, Rahm E, et al. Cathode materials modified by surface coating for lithium ion batteries. Electrochim Acta. 2006;51:3872–83. DOI: 10.1016/j.electacta.2005.11.015. [5] Belharouak I, Johnson C, Amine K. Synthesis and electrochemical analysis of vapor-deposited carbon-coated LiFePO4. Electrochem Commun. 2005;7:983–8. DOI: 10.1016/j.elecom.2005. 06.019. [6] Chen Z, Dahn JR. Reducing carbon in liFePO4/C composite electrodes to maximize specific energy, volumetric energy, and tap density. J Electrochem Soc. 2002;149:A1184–4. DOI: 10.1149/1.1498255. [7] Li HH, Jin J, Wei JP, Zhou Z, Yan J. Fast synthesis of Core-Shell LiCoPO4/C nanocomposite via microwave heating and its electrochemical Li intercalation performances. Electrochem Commun. 2009;11:95–8. DOI: 10.1016/j.elecom.2008.10.025. [8] Zhang Y, Sun CS, Zhou Z. Sol–gel preparation and electrochemical performances of LiFe1/3Mn1/3Co1/3PO4/C composites with Core–Shell nanostructure. Electrochem Commun. 2009;11:1183–6. DOI: 10.1016/j.elecom.2009.03.044. [9] Cao Q, Zhang HP, Wang GJ, Xia Q, Wu YP, Wu HQ, et al. A novel carbon-coated LiCoO2 as cathode material for lithium ion battery. Electrochem Commun. 2007;9:1228–32. DOI: 10.1016/j.elecom. 2007.01.017. [10] Chen Z, Dahn JR. Studies of LiCoO2 coated with metal oxides. Electrochem Solid-State Lett. 2003;6:A221–1. DOI: 10.1149/1.1611731. [11] Cho J, Kim Y, Park B. Novel LiCoO2 cathode material with Al2O3 coating for a Li Ion cell. Chem Mater. 2000;12:3788–91. [12] Guo S, Yu H, Liu P, Liu X, Li D, Chen M, et al. Surface coating of Lithium–Manganese-rich layered oxides with delaminated MnO2 nanosheets as cathode materials for Li-Ion batteries. J Mater Chem A. 2014;2:4422–2. DOI: 10.1039/C3TA15206E.

References

163

[13] Meng H, Li L, Liu J, Han X, Zhang W, Liu X, et al. Surface modification of Li-rich layered Li [Li0.17Ni0.17Co0.10Mn0.56]O2 oxide with LiV 3 O 8 as a cathode material for Li-Ion batteries. J Alloys Compd. 2017;690:256–66. DOI: 10.1016/j.jallcom.2016.08.089. [14] Sun YK, Lee YS, Yoshio M, Amine K. Synthesis and electrochemical properties of ZnO-coated LiNi0.5Mn1.5O4Spinel as 5 V cathode material for lithium secondary batteries. Electrochem Solid-State Lett. 2002;5:A99–9. DOI: 10.1149/1.1465375. [15] Seok J, Cavanagh AS, Yan Y, George SM, Manthiram A. Effects of atomic layer deposition of Al2O3 on the Li[Li0.20Mn0.54Ni0.13Co0.13]O2 cathode for Lithium-Ion batteries. J Electrochem Soc. 2011;158:A1298–8. DOI: 10.1149/2.030112jes. [16] Sun Y, Han J, Myung S, Lee S, Amine K. Significant improvement of high voltage cycling behavior AlF3-coated LiCoO2 cathode. Electrochem Commun. 2006;8:821–6. DOI: 10.1016/j.elecom. 2006.03.040. [17] Laskar Masihhur R, Jackson David HK, Guan Y, Xu S, Fang S, Dreibelbis M, et al. Atomic layer deposition of Al2O3–Ga2O3 alloy coatings for Li[Ni0.5Mn0.3Co0.2]O2 cathode to improve rate performance in Li-Ion battery. ACS Appl Mater Interfaces. 2016;8:10572–80. DOI: 10.1021/ acsami.5b11878. [18] Zhang J, Xiang YJ, Yu Y, Xie S, Jiang GS, Chen CH, et al. Electrochemical evaluation and modification of commercial lithium cobalt oxide powders. J Power Sources. 2004;132:187–94. DOI: 10.1016/j.jpowsour.2004.01.035. [19] Cao H, Xia B, Zhang Y, Xu N. LiAlO-coated LiCoO as cathode material for lithium ion batteries. Solid State Ionics. 2005;176:911–4. DOI: 10.1016/j.ssi.2004.12.001. [20] Sun YK, Cho SW, Lee SW, Yoon CS, Amine K. AlF3-coating to improve high voltage cycling performance of Li[N1/3Co1/3Mn1/3]O2 cathode materials for lithium secondary batteries. J Electrochem Soc. 2007;154:A168–8. DOI: 10.1149/1.2422890. [21] Bai Y, Liu N, Liu J, Wang Z, Chen L. Coating material-induced acidic electrolyte improves LiCoO2 performances. Electrochem Solid-State Lett. 2006;9:A552–2. DOI:10.1149/1.2357982. [22] Appapillai AT, Mansour AN, Cho J, Shao-Horn Y. Microstructure of LiCoO2 with and without “AlPO4” nanoparticle coating: combined STEM and XPS studies. Chem Mater. 2007;19: 5748–57. DOI: 10.1021/cm0715390. [23] Lee KS, Myung ST, Amine K, Yashiro H, Sun YK. Dual functioned BiOF-coated Li [Li0.1Al0.05Mn1.85]O4 for lithium batteries. J Mater Chem. 2009;19:1995–5. DOI: 10.1039/ B819224C. [24] Jang SB, Kang SH, Amine K, Bae YC, Sun YK. Synthesis and improved electrochemical performance of Al (OH)3-coated Li[Ni1/3Mn1/3Co1/3]O2 cathode materials at elevated temperature. Electrochim Acta. 2005;50:4168–73. DOI: 10.1016/j.electacta.2005.01.037. [25] Fergus Jeffrey W. Recent developments in cathode materials for lithium ion batteries. J Power Sources. 2010;195:939–54. DOI: 10.1016/j.jpowsour.2009.08.089. [26] Ying J, Wan C, Jiang C. Surface treatment of LiNi0.8Co0.2O2 cathode material for lithium secondary batteries. J Power Sources. 2001;102:162–6. DOI:10.1016/j.jpowsour.2004.01.033. [27] Amatucci GG, Blyr A, Sigala C, Alfonse P, Tarascon JM. Surface treatments of Li1+xMn2−xO4 spinels for improved elevated temperature performance. Solid State Ion. 1997;104:13–25. DOI: 10.1016/S0167-2738(97)00407-4. [28] Kim HS, Kong M, Kim K, Kim IJ, Gu HB. Effect of carbon coating on LiNi1/3Mn1/3Co1/3O2 cathode material for lithium secondary batteries. J Power Sources. 2007;171:917–21. DOI: 10.1016/j.jpowsour.2007.06.028. [29] Kim J, Kim B, Lee JG, Cho J, Park B. Direct carbon-black coating on LiCoO2 cathode using surfactant for high-density Li-Ion cell. J Power Sources. 2005;139:289–94. DOI: 10.1016/j.jpowsour.2004.07.008.

164

4 Battery Materials

[30] Liu W, Oh P, Liu X, Lee MJ, Cho W, Chae S, et al. Nickel-rich layered lithium transition-metal oxide for high-energy Lithium-Ion batteries. Angew Chem. 2015;54:4440–57. DOI: 10.1002/ anie.201409262. [31] Jung YS, Cavanagh AS, Dillon AC, Groner MD, George SM, Lee SH, et al. Enhanced stability of LiCoO2 cathodes in Lithium-Ion batteries using surface modification by atomic layer deposition. J Electrochem Soc. 2010;157:A75–5. DOI: 10.1149/1.3258274. [32] Chen Z, Dahn JR. Effect of a ZrO2 coating on the structure and electrochemistry of LixCoO2 when cycled to 4.5 V. Electrochem Solid-State Lett. 2002;5:A213–3. DOI: 10.1149/1.1503202. [33] Badot JC, Mantoux A, Baffier N, Dubrunfaut O, Lincot D. Electrical properties of V2O5 thin films obtained by atomic layer deposition (ALD). J Mater Chem. 2004;14:3411–1. DOI: 10.1039/ B410324F. [34] Shi Y, Zhang M, Qian D, Meng YS. Ultrathin Al2O3 coatings for improved cycling performance and thermal stability of LiNi0.5Co0.2Mn0.3O2 cathode material. Electrochim Acta. 2016;203:154– 61. DOI: 10.1016/j.electacta.2016.03.185. [35] Sun YK, Myung ST, Kim MH, Prakash J, Amine K. Synthesis and characterization of Li [(Ni0.8Co0.1Mn0.1)0.8(Ni0.5Mn0.5)0.2]O2 with the microscale Core−Shell structure as the positive electrode material for lithium batteries. J Am Chem Soc. 2005;127:13411–8. DOI: 10.1021/ ja053675g. [36] Sun YK, Myung ST, Park BC, Prakash J, Belharouak I, Amine K, et al. High-energy cathode material for long-life and safe lithium batteries. Nat Materials. 2009;8:320–4. DOI: 10.1038/ nmat2418. [37] Koenig Gary M, Belharouak I, Deng H, Sun YK, Amine K. Composition-tailored synthesis of gradient transition metal precursor particles for Lithium-Ion battery cathode materials. Chem Mater. 2011;23:1954–63. DOI: 10.1021/cm200058c. [38] Cho Y, Lee S, Lee Y, Hong T, Cho J. Spinel-layered Core-Shell cathode materials for Li-Ion batteries. Adv Energy Mater. 2011;1:821–8. DOI: 10.1002/aenm.201100239. [39] Kalluri S, Yoon M, Jo M, Liu HK, Dou SX, Cho J, et al. Adv Mater (Deerfield Beach, Fla.), 2017. [40] Park SH, Park KS, Sun YK, Nahm KS, Lee YS, Yoshio M, et al. Structural and electrochemical characterization of lithium excess and Al-doped nickel oxides synthesized by the sol–gel method. Electrochim Acta. 2001;46:1215–22. DOI: 10.1016/S0013-4686(00)00710-6. [41] Liu L, Sun K, Zhang N, Yang T. Improvement of high-voltage cycling behavior of Li(Ni1/3Co1/3Mn1/3)O2 cathodes by Mg, Cr, and Al substitution. J Solid State Electrochem. 2008;13:1381–6. DOI: 10.1007/ s10008-008-0695-z. [42] Nayak PK, Grinblat J, Levi M, Levi E, Kim S, Choi JW, et al. Al doping for mitigating the capacity fading and voltage decay of layered Li and Mn-rich cathodes for Li-Ion batteries. Adv Energy Mater. 2016;6:1502398–8. DOI:10.1002/aenm.201502398. [43] Zhou Fu, Zhao Xuemei, Dahn JR. Synthesis, electrochemical properties, and thermal stability of Al-doped LiNi1∕3Mn1∕3Co[sub (1∕3−z)]AlzO2 positive electrode materials. J Electrochem Soc. 2009;156:A343–3. DOI:10.1149/1.3079424. [44] Xiao L, Zhao Y, Yang Y, Cao Y, Ai X, Yang H, et al. Enhanced electrochemical stability of Al-doped LiMn2O4 synthesized by a polymer-pyrolysis method. Electrochim Acta. 2008;54:545–50. DOI: 10.1016/j.electacta.2008.07.037. [45] Liu D, Wang Z, Chen L. Comparison of structure and electrochemistry of Al- and Fe-doped LiNi1/3Co1/3Mn1/3O2. Electrochim Acta. 2006;51:4199–203. DOI: 10.1016/j. electacta.2005.11.045. [46] Pouillerie C, Perton F, Biensan Ph, Pérès JP, Broussely M, Delmas C, et al. Effect of magnesium substitution on the cycling behavior of lithium nickel cobalt oxide. J Power Sources. 2001;96:293–302. DOI: 10.1016/S0378-7753(00)00653-4.

References

165

[47] Liu Z, Zhang X, Hong L. Preparation and electrochemical properties of spherical LiFePO4 and LiFe0.9Mg0.1PO4 cathode materials for lithium rechargeable batteries. J Appl Electrochem. 2009;39:2433–8. DOI: 10.1007/s10800-009-9931-1. [48] Shiratsuchi T, Okada S, Doi T, Yamaki J. Cathodic performance of LiMn1−xMxPO4 (M=Ti, Mg and Zr) annealed in an inert atmosphere. Electrochim Acta. 2009;54:3145–51. DOI: 10.1016/j. electacta.2008.11.069. [49] Ren M, Zhou Z, Li Y, Gao XP, Yan J. Preparation and electrochemical studies of Fe-doped Li3V2 (PO4)3 cathode materials for Lithium-Ion batteries. J Power Sources. 2006;162:1357–62. DOI: 10.1016/j.jpowsour.2006.08.008. [50] Song MY, Kwon IH, Song J, Shim S. Electrochemical properties of Li1−z (Ni1−y Fey)1+z O2 synthesized by the combustion method in an air atmosphere. J Appl Electrochem. 2008;39:617–25. DOI: 10.1007/s10800-008-9699-8. [51] Schipper F, Dixit M, Kovacheva D, Talianker M, Haik O, Grinblat J, et al. Stabilizing nickel-rich layered cathode materials by a high-charge cation doping strategy: zirconium-doped LiNi0.6Co0.2Mn0.2O2. J Mater Chem A. 2016;4:16073–84. DOI: 10.1039/C6TA06740A. [52] Liu H, Li C, Cao Q, Wu YP, Holze R. Effects of heteroatoms on doped LiFePO4/C composites. J Solid State Electrochem. 2007;12:1017–20. DOI:10.1007/s10008-007-0480-4. [53] Sivaprakash S, Majumder SB. Understanding the role of Zr4+ cation in improving the cycleability of LiNi0.8Co0.15Zr0.05O2 cathodes for Li ion rechargeable batteries. J Alloys Compd. 2009;479:561–8. DOI: 10.1016/j.jallcom.2008.12.129. [54] Kim J, Amine K. The effect of tetravalent titanium substitution in LiNi1−xTixO2 (0.025⩽x⩽0.2) system. Electrochem Commun. 2001;3:52–5. DOI: 10.1016/S1388-2481(00)00151-X. [55] Wu SH, Chen MS, Chien CJ, Fu YP. Preparation and characterization of Ti4+-doped LiFePO4 cathode materials for lithium-ion batteries. J Power Sources. 2009;189:440–4. DOI: 10.1016/ j.jpowsour.2009.01.015. [56] Liu GQ, Yuan WS, Liu GY, Tian YW. The electrochemical properties of LiNi0.5Mn1.2Ti0.3O4 compound. J Alloys Compd. 2009;484:567–9. DOI: 10.1016/j.jallcom.2009.04.141. [57] Wilcox J, Patoux S, Doeff M. Structure and electrochemistry of LiNi1/3Co1∕3−yMyMn1∕3O2 (M=Ti, Al, Fe) positive electrode materials. J Electrochem Soc. 2009;156:A192–2. DOI: 10.1149/ 1.3056109. [58] Song MY, Mumm DR, Park CK, Park HR. Cycling performance of LiNi1−yMyO2 (M=Ni, Ga, Al and/or Ti) synthesized by wet milling and solid-state method. Met Mater Int. 2012;18:465–72. DOI: 10.1007/s12540-012-3013-3. [59] Nishida Y, Nakane K, Satoh T. Synthesis and properties of gallium-doped LiNiO2 as the cathode material for lithium secondary batteries. J Power Sources. 1997;68:561–4. DOI: 10.1016/S03787753(97)02535-4. [60] Du G, NuLi Y, Yang J, Wang J. Fluorine-doped LiNi0.5Mn1.5O4 for 5V cathode materials of Lithium-Ion battery. Mater Res Bull. 2008;43:3607–13. DOI: 10.1016/j. materresbull.2008.02.025. [61] Wang D, Wang Z, Li X, Guo H, Xu Y, Fan Y, et al. Effect of surface fluorine substitution on high voltage electrochemical performances of layered LiNi0.5Co0.2Mn0.3O2 cathode materials. Appl Surf Sci. 2016;371:172–9. DOI: 10.1016/j.apsusc.2016.02.224. [62] Zheng J, Wu X, Yang Y. Improved electrochemical performance of Li[Li0.2Mn0.54Ni0.13Co0.13]O2 cathode material by fluorine incorporation. Electrochim Acta. 2013;105:200–8. DOI: 10.1016/j. electacta.2013.04.150. [63] Chen Y, Zhao Y, An X, Liu J, Dong Y, Chen L, et al. Preparation and electrochemical performance studies on Cr-doped Li3V2(PO4)3 as cathode materials for Lithium-Ion batteries. Electrochim Acta. 2009;54:5844–50. DOI: 10.1016/j.electacta.2009.05.041.

166

4 Battery Materials

[64] Amarilla JM, Petrov K, Picó F, Avdeev G, Rojo JM, Rojas RM, et al. Sucrose-aided combustion synthesis of nanosized LiMn1.99−yLiyM0.01O4 (M=Al3+, Ni2+, Cr3+, Co3+, y=0.01 and 0.06) spinels. J Power Sources. 2009;191:591–600. DOI: 10.1016/j. jpowsour.2009.02.026. [65] Churikov AV, Kachibaya EI, Sycheva VO, Ivanishcheva IA, Imnadze RI, Paikidze TV, et al. Electrochemical properties of LiMn2−yMeyO4 (Me = Cr, Co, Ni) spinels as cathodic materials for Lithium-Ion batteries. Russ Electrochem. 2009;45:175–82. DOI: 10.1134/ S1023193509020086. [66] Kobayashi Y, Mita Y, Seki S, Ohno Y, Miyashiro H, Nakayama M, et al. Configurational entropy of lithium manganese oxide and related materials, LiCryMn2−yO4 (y=0, 0.3). J Electrochem Soc. 2008;155:A14–4. DOI: 10.1149/1.2799069. [67] Mohan P, Kumar KA, Kalaignan GP, Muralidharan VS. Improved electrochemical properties of chromium substituted in LiCr1 − xNixO2 cathode materials for rechargeable Lithium-Ion batteries. J Solid State Electrochem. 2012;16:3695–702. DOI: 10.1007/ s10008-012-1808-2. [68] Guo B, Zhao J, Fan X, Zhang W, Li S, Yang Z, et al. Aluminum and fluorine co-doping for promotion of stability and safety of lithium-rich layered cathode material. Electrochim Acta. 2017;236:171–9. DOI: 10.1016/j.electacta.2017.03.133. [69] Schipper F, et al. From surface ZrO2 coating to bulk Zr doping by high temperature annealing of Nickel-rich lithiated oxides and their enhanced electrochemical performance in Lithium Ion batteries. Adv Energy Mater. 2017;8:1701682–2. DOI: 10.1002/ aenm.201701682.

4.5 Positive electrodes based on Ion-implanted SrTiO3 Max Stöber and Charaf Cherkouk Abstract: An O2-electrode was fabricated using a metal ion implanted SrTiO3 single crystal. The time resolved oxygen exchange rate of ion implanted strontium titanate (SrTiO3) single crystals was studied by means of oxygen solid electrolyte coulometry (OSEC). Transmission electron microscopy (TEM) was performed in order to determine structural changes after ion implantation. Moreover, theoretical modelling based on defect chemistry under equilibrium conditions was applied for determining of effective rate constants. OSEC measurements turn out to be a damage and calibration free method, which was used for the first time in order to characterize kinetic parameters of oxygen exchange on single crystalline surfaces. Keywords: Ion Implantation, Oxygen Exchange, Oxygen Solid Electrolyte Coloumetry (OSEC), Oxygen Vacancies, Strontium Titanate

4.5 Positive electrodes based on Ion-implanted SrTiO3

167

4.5.1 Introduction The search for powerful energy storage systems in order to manage excess grid power and decrease the need to burn hydrocarbon fuel is seen as one of the key challenges for mitigating the climate change [1]. Energy storage systems, which are using ambient oxygen as air electrode, like hydrogen fuel cells or metal air batteries, are able to reach significantly higher gravimetric and volumetric energy densities compared to lithium ion batteries [2, 3]. Strontium titanate (SrTiO3) is, due to its ability to incorporate and transport oxygen, a promising candidate for such systems [4–8]. Furthermore, the unique defect chemistry of SrTiO3 is known to be responsible for peculiar new properties like pyroelectricity [9] and acting as an all-solid battery [8, 10]. The oxygen vacancies in SrTiO3 are sufficiently mobile even at ambient temperature [8, 11]. Metal oxide ceramics are attractive materials for solid electrolytes, because of their low electronic conductivity combined with a considerable oxygen ion conductivity at elevated temperatures [12]. Most materials of this class also offer advantages in chemical and mechanical robustness [9] and a very high melting temperature [13]. Table 4.5.1 shows a small fraction from the vast diversity of this material class.

Table 4.5.1: Examples of functional metal oxide materials considered in literature for different applications. Application

Metal oxide materials

Metal air battery [5] Solid oxide fuel cell (electrolyte) [14]

Co3O4, MnO2, Mn2O3, LaNiO3, LaMnO3 Yttria-stabilized zirconia (YSZ), La2O3/GaO/MgO/SrO, (LGSM), SrTiO3, CeO2, Ce2O3 La(1-x) SrxMnO3(LSM), La(1-x)SrxCoO3(LSX), La(1-x)SrxTiO3, GdBaMn2O5 SrTiO3, CeO2, ZrO2, La SrTiO3,YSZ TiO2, ZnO, Ag3PO4, Mn2O3, Cu2O, WO3, SrTiO3, CoFeO4 Ga2O3, SrTiO3, SrTiFeO3, TiO2, Nb2O5, ZnO

Solid oxide fuel cell (cathode) [8, 15, 16] Solid oxide fuel cell (anode) [17, 18] Photocatalyst (water splitting) [19, 20] Gas sensing: oxygen (chemresistor/ semiconductor type) [18] Gas sensing: oxygen (amperometric) [18, 19]

ZrO2, Bi2O3/MoO3, YSZ

Even though the fields of applications differ widely, the key requirements are always similar: combining high reactivity with oxygen or oxygen containing compounds with fast conductivity of oxygen ions on the materials surface or through the bulk crystal. It is remarkable, that SrTiO3 has been investigated not only as a cathode, but also as anode and electrolyte for energy storage devices, which makes it very universal (Table 4.5.1).

168

4 Battery Materials

4.5.2 Synthesis and structural characterization of SrTiO3 single crystal O2-electrodes The ideal cubic perovskite structure ABO3 can be visualised as a network of BO6 octahedra, with A-metal ions in the interstitials [21]. When at ambient conditions, a pure SrTiO3 crystal condenses in this highly symmetric form. The symmetry group of this lattice is Pm-3 m. There are many ways to break this symmetry into a less symmetric structure, for example metal ion doping [10], ion implantation [22], annealing or application of an electric field [13]. The symmetry, and therefore the number of energetically equivalent pathways trough the structure, is playing a very significant role for the transport of ionic species. Single crystalline SrTiO3 samples with (100)-orientation were used. Prior to ion implantation the samples were preheated in N2 atmosphere at 700°C for 5 h to increase electronic conductivity. This was necessary in order to connect the samples to ground potential avoiding charging during ion implantation. The samples were implanted with Ni+, N+, O+ and Ag+ ions at an energy of 180 keV. To avoid channelling effects, the ion beam was adjusted to 7 tilt from the perpendicular axis of the sample’s surface. The ion implantation was performed at the Ion Beam Centre (IBC) of the Helmholtz-Zentrum Dresden-Rossendorf, Germany. Ag+ dopants were used in order to compare the catalytic effect of the nickel ions from point of view of the electronic configuration and a different atomic radius (Ag 145 pm to Ni 124 pm). Both dopants of N+ and O+ ions are volatile in SrTiO3 and, therefore, the change of oxygen exchange caused by the positional disorder from ion implantation can be better understood. It has been shown by transmission electron microscopy (TEM) investigations (Figure 4.5.1) that the ion implantation has caused amorphization in the first 180 nm of the sample implanted with Ni+-ions at 180 keV. In literature, this is described as a metastable glass phase [23], meaning that the layer may be caused to regrow and

glue outside of crystal

implanted near surface

bulk crystal 1 μm

100 nm

20 nm

Figure 4.5.1: TEM images of Ni implanted SrTiO3, which has been thinned by ion beam etching. The amorphous layer (bright area) is 180 nm thick and is showing no sign of metal nanoclusters greater than the resolution limit of roughly 3 nm.

4.5 Positive electrodes based on Ion-implanted SrTiO3

169

form new phases using thermal annealing. This process is depending on exact temperature [24] and humidity conditions [25]. It is evident that the metal ion implantation had drastic impact on the first 180 nm but did not change the remaining bulk of the material. Therefore, defect chemical calculations and considerations of oxygen transport in single crystals remain valid for the vast majority of the sample. The implanted layer, as it is directly after ion implantation, shows no signs of metal nanoclusters. Such clusters were expected to occur according to investigations of Xiang et al. on 1015 cm−2 fluence Ni doped Al2O3. The spatial resolution of the microscope can be estimated to about (2–3) nm.

4.5.3 Oxygen diffusion and defect chemistry in strontium titanate For SrTiO3 single crystals at elevated temperatures, bulk diffusion is significantly faster compared to diffusion trough a phase boundary [26]. Therefore, the whole process can be seen as limited by the surface reaction – meaning the diffusion of oxygen through the phase boundary:   jth = kδ  ½V O€ðtÞ – ½V O€eq ðT, pO2 Þ ,

(4:5:1)

where jth is the theoretical oxygen flow through the boundary of the solid, ½V O€ðtÞ the oxygen concentration at the surface at a given time t ½V O€eq ðT, pO2 Þ the equilibrium concentration for a given temperature T and ambient oxygen partial pressure pO2 , and kδ the effective rate constant. From this equation, it is evident that the oxygen flow at the phase boundary is proportional to the difference between current and equilibrium oxygen concentration. This can be understood as analogous to Ficks first law, which describes diffusion as caused by a gradient in concentration. Contrary to the laws of diffusion in homogenous phases, an exchange over the phase boundary is also possible, when there is no change in concentration at the gas phase, because the equilibrium ½V O€eq ðT Þ is temperature dependent. The linear factor of this proportionality kδ is depending on temperature T by an Arrhenius law: 

 – EA , kδ = kδ, 0  exp kB  T

(4:5:2)

where kδ, 0 is the pre-exponential factor, kB the Boltzmann constant and EA as the activation energy. The exponential influence of the reciprocal temperature 1/T is very crucial to the behaviour of the exchange rate. That means temperature changes of 50 K may influence the exchange rate by a whole order of magnitude. At temperatures below 450 °C, the exchange is practically zero, a state which is often referred to as the frozen-in equilibrium [11, 20]. For temperatures above 450 °C one can calculate the theoretical oxygen flow jth , if the

170

4 Battery Materials

function of the equilibrium oxygen vacancy concentration ½V O€eq ðT, pO2 Þ is well known. It can be accessed by a defect chemistry model introduced by Denk et al. [27]. This leads to an equation system formed by several mass action laws, which for the sake of simplicity, was approximately solved by a third grade polynomial function as depicted in Figure 4.5.2.

18.4 18.2 18.0

log10 [VO¨] (cm−3)

18.6

17.8 17.6 –3

–2

–1

log

10

800 700 600 500

0

1

pO

2 (Pa)

2

3

4

200 5

6

0

300

400 C) T (°

100

Figure 4.5.2: Concentration of oxygen vacancies ½V O€ from the defect chemical model of equilibrium states as a function of temperature T and oxygen partial pressure pO2 . Bold black lines indicate the temperature-dependent paths for fixed oxygen partial pressures of 1 Pa and 21,000 Pa, being the partial pressures during measurement and during equilibration on synthetic air. The coloured area in-between marks the reachable values of ½V O€ during temperature changes. Every change in height corresponds to a certain exchange of oxygen.

4.5.4 Oxygen solid electrolyte coulometry on ion-implanted SrTiO3 single crystals Figure 4.5.3 shows the experimental (blue) and fitted (green) oxygen flow during the oxygen release (excorporation) on silver doped SrTiO3 single crystal using the OSEC setup as a function of time. The OSEC setup is described in the section 5.7. The experimental time-resolved oxygen flow jexp ðtÞ fits well with the theoretical model jth ðtÞ, which can be calculated by an iteration of equations (1) and (2) over small time steps. The parameters EA , kδ, 0 and ½V O€ðt = 0Þ were refined for best fit. The working principle of this model is explained in Figure 4.5.3, and is based on chemical equilibrium calculations according to equation (1). In the model calculation it is

171

4.5 Positive electrodes based on Ion-implanted SrTiO3

1200 40

Oxygen flow (mol/s·10−12)

800°C

800°C

800

20

600

experiment jexp(t)

10 0

40°C 10

0

2

4

400 200

simulation jth(t) 6

Temperature (° C)

1000 30

8

0

Oxygen vacancies (mol·10−9)

100 80 no exchange, T too low

60 40

equilibrium

0

2

4

6

8

Time (h)

Figure 4.5.3: An example of experimentally determined oxygen flow using OSEC (blue) and the corresponding fitted (green) oxygen release (excorporation) on silver doped SrTiO3 single crystal as a function of time under a temperature offset (red) at constant change rate of 600 K/h. Outline of the working principle used for calculation of simulated OSEC data (bottom): The lower section depicts the states of oxygen vacancy concentration at every given time ½V O€ðt Þ compared to the equilibrium state

V O, € eq (T) = [V O€](T, t → ∞) for the given temperature T. Changing T in a ramp-like scheme initiates an exchange reaction. This means a change of oxygen vacancy concentration, resulting in the calculated oxygen flow jth(t) (upper green graph). This can also be measured, resulting in OSEC curves (blue graph).

assumed that the exchange reaction is relatively slow compared to the diffusion, meaning that the oxygen vacancy concentration in the crystal’s volume has no gradient. Spatial distribution can, therefore, be neglected, and ½V O€ is just a function of time and not depending on space coordinates. The time axis was discretized in steps of 1 s, and an acceptor concentration (impurities/dopants) of 1.4 × 1020 cm−3 had to be assumed in order to obtain the right integrated exchange concentration (see Figure 4.5.3). From fitting the jexp ðtÞ curves to the iterative model by varying the exchange parameters kδ, 0 and EA , the results were obtained for a variety of samples implanted with different metal ions, summarized in Table 4.5.2. Some metallic elements like silver, iron, nickel or platinum have a high catalytic activity due to their electronic structure [28]. However, the properties of such atoms are to be discussed in a different way, when they are not forming a macroscopic metallic structure but are incorporated into an ionic metal oxide lattice. The implantation of Ni metal ions did cause an increase in the maximum of measured oxygen exchange rate jmax of 33.5% (Table 4.5.2), which means that the effective exchange rate has increased by approximately 67%, because only half of the crystal’s surface

172

4 Battery Materials

Table 4.5.2: Summary of the measured exchange characteristics and the respective fitting parameters for a set of OSEC O2 exchange experiments on SrTiO3 crystals with different doping species. Experiment Doping None Ag Ni O N

[O2]ex (nmol)

jmax (pmols−1)

23.2 27.7 28.6 25.4 24.5

10.7 ± 0.2 12.3 ± 0.2 14.3 ± 0.2 10.4 ± 0.2 9.6 ± 0.2

Simulation kδ,0 EA (eV) (10 cms−1) 8

2.45 2.45 2.45 2.45 2.45

2.5 2.49 2.48 2.52 2.52

Literature [22] 

kδ ð800 CÞ (cm−2s−1) 4.0×10–4 4.9×10–4 5.5×10–4 3.6×10–4 3.9×10–4

EA (eV)

kδ ð800 CÞ (cm−2s−1)

4.34 2.7 ± 0.7 – – – – – – – –

3.05×10–4 – – – –

kδ,0 (10 cms−1) 8

area (front side) was implanted. It has to be discussed, which mechanistic principles are causing those changes. Implantation with Ag did hardly at all increase the exchange rate. This can likely be explained by the different electronic configuration or the different atomic radius (Ag 145 pm to Ni 124 pm), which makes incorporation into the tight perovskite lattice less likely. Also, the implantation with N+ and O+ ions did not significantly change the exchange characteristics, probably because those elements are volatile in this material system and did not form a stable doped layer during temperature cycles. On the other hand, this leads to a quite important implication that the increased oxygen exchange was not entirely caused by the disorder from ion implantation. As it is evident from the data presented in Figure 4.5.3, some of the peaks are not fitting to the theoretical model. It can be assumed, that higher precision of the measurement would lead to a smoother graph. The presented setup will be improved soon with a new reactor, which uses metallic gaskets instead of glass and sealing paste. Also, the newly constructed reactor minimizes dead volumes and, therefore, also the inner surface where gases can adsorb on.

4.5.5 Summary Surface and near surface investigations of especially Ni ion implanted SrTiO3 single crystals have revealed a 180 nm thick disordered layer, which is significantly different from the bulk in terms of structure and elementary composition. Measurements of oxygen exchange using OSEC have shown that this layer indeed benefits the exchange rate and, therefore, the kinetic parameters governing this reaction. The simulated OSEC curves fit reasonably well with the experimental data of pristine crystals, with an activation energy for oxygen exchange estimated at 2.5 eV and a preexponential factor at 2.45 × 108 cms−1, which is in agreement with data from literature. Ni ion implantation has caused a 67% increase in total exchange rate compared

References

173

to unperturbed samples. Even though a more conventional way of reactivity increase, using Ag thick film brush deposition, was proven to be more effective, the idea of using metal ion implantation to accelerate the oxygen reduction reaction has proven to be legit. For future research it is worth investigating how this treatment may benefit other structures in terms of oxygen exchange reactions. Also, an investigation of the catalytic activity with the same material but using hydrogen instead of oxygen would be worth considering. Coulometric titration has proven to be a very powerful tool for such measurements and presents, as yet, an untapped potential for the research on crystalline materials.

References [1] [2] [3] [4]

[5]

[6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16]

Gallo A, Simoes-Moreira J, Costa H, Santos M, Dos Santos EM. Energy storage in the energy transition context: a technology review. Renew Sust Energ Rev. 2016;65:800–22. Cheng F, Chen J. Metal-air batteries: from oxygen reduction electrochemistry to cathode catalysts. Chem Soc Rev. 2012;41:2502172–92. Van Noorden R. The rechargeable revolution: a better battery. Nat. 2014;507:26–8. Ma Q, Iwanschitz B, Dashjav E, Baumann S, Sebold D, Raj IA, et al. Microstructural variations and their influence on the performance of solid oxide fuel cells based on yttrium-substituted strontium titanate ceramic anodes. J Power Sources. 2015;279:678–85. Ruiz-Morales JC, Canales-Vazquez J, Savaniu C, Marrero-López D, Zhou W, Irvine JT. Disruption of extended defects in solid oxide fuel cell anodes for methane oxidation. Nat. 2006;439: 568–71. Miller DN, Irvine JT. B-site doping of lanthanum strontium titanate for solid oxide fuel cell anodes. J Power Sources. 2011;196:7323–7. Gerblinger J, Meixner H. Fast oxygen sensors based on sputtered strontium titanate, Sensors and Actuators B. Chem. 1991;4: 99–102. 265. Hanzig J, Zschornak M, Nentwich M, Hanzig F, Gemming S, Leisegang T, et al. Strontium titanate: an all-in-one rechargeable energy storage material. J Power Sources. 2014;267:700–5. Hanzig J, Mehner E, Jachalke S, Hanzig F, Zschornak M, Richter C, et al. Dielectric to pyroelectric phase transition induced by defect migration. New J Phys. 2015;17:023036. Stöcker H, Hanzig J, Zschornak M, Mehner E, Jachalke S, Richter C, et al. Strontium titanate: from symmetry changes to functionality. Cryst Res Technol. 2016;1:1600222. Hanzig J, Zschornak M, Mehner E, Hanzig F, Münchgesang W, Leisegang T, et al. The anisotropy of oxygen vacancy migration in SrTiO3. J Phys Condens Matter. 2016;28:225001. Nassau K, Miller A. Strontium titanate: an index to the literature on properties and the growth of single crystals. J Crystal Growth. 1988;91:373–81. De Souza R, Zehnpfenning J, Martin M, Maier J. Determining oxygen isotope profiles in oxides with time-of-flight SIMS. Solid State Ionics. 2005;176:1465–71. Mahato N, Banerjee A, Gupta A, Omar S, Balani K. Progress in material selection for solid oxide fuel cell technology: a review. Prog Mater Sci. 2015;72:141–337. Kilner JA, Burriel M. Materials for intermediate-temperature solid oxide fuel cells. Annu Rev Mater Res. 2014;44:365–93. Ruiz-Morales JC, Canales-Vázquez J, Savaniu C, López DM, Zhou W, Irvine JTS. Disruption of extended defects in solid oxide fuel cell anodes for methane oxidation. Nat. 2006;439:568–71.

174

4 Battery Materials

[17] Marina OA, Canfield NL, Stevenson JW. Thermal, electrical, and electrocatalytical properties of lanthanum-doped strontium titanate. Solid State Ionics. 2002;149:21–8. [18] Shen X, Sasaki K. Robust SOFC anode materials with la doped SrTiO3 backbone structure. Int J Hydrogen Energy. 2016;41:17044–52. [19] Upadhyay RK, Soin N, Roy SS. Role of graphene/metal oxide composites as photocatalysts, adsorbents and disinfectants in water treatment: a review. Reaserch Adv. 2014;4:3823–51. [20] Hertkorn D, Benkler M, Gleißner U, Büker F, Megnin C, Müller C, et al. Morphology and oxygen vacancy investigation of strontium titanate-based photo electrochemical cells. J Mater Sci. 2015;50:40–8. [21] Johnsson M, Lemmens P. Crystallography and Chemistry of Perovskites. In: Kronmüller H, Parki S., editor(s). Handbook of Magnetism and Advanced Magnetic MaterialsNovel Materials Vol. 4. Chichester, Uk: John Wiley & Sons Ltd, 2006:2098–106. [22] Stöcker H, Zschornak M, Richter C, Hanzig J, Hanzig F, Hinze A, et al. Surface-near modifications of SrTiO3 local symmetry due to nitrogen implantation investigated by grazing incidence XANES. Scripta Mater. 2014;86:1–4. [23] Bernas H. Ion beam-induced amorphization: A crystal-to-glass transition? In: Sigmund P., editor(s). Ion beam Science: Solved and Unsolved Problems, Matematisk-fysiske Meddelelser ed. Vol. 52. Copenhagen: The Royal Danish Academy of Science and Letters, 2006:383–403. [24] Liedtke R, Hoffmann S, Waser R. Recrystallization of oxygen ion implanted Ba 0.7Sr0.3TiO3 thin films. J Am Ceram Soc. 2000;83:436–8. [25] Simpson TW. Recrystallisation of strontium titanate, PhD thesis, University of Western Ontario, 1993. [26] Merkle R, Maier J. Wie wird sauerstoff in oxide eingebaut? Kinetische studie einer “simplen” Feststoffreaktion am modellmaterial SrTiO3. Angew Chem. 2008;120:3936–58. [27] Denk I, Münch W, Maier J. Partial conductivities in SrTiO3: bulk polarization experiments, oxygen concentration cell measurements, and defect-chemical modeling. J Am Ceram Soc. 1995;78:3265–72. [28] Roque-Malherbe RMA. In: Roque-Malherbe RMA., editor(s). The physical chemistry of materials, CRC Press ed. Boca Raton London New York: Taylor&Francis Group, 2010978113811 7709.

4.6 Separators and electrolytes for rechargeable batteries: Fundamentals and perspectives Tina Nestler, Elsa Roedern, Nikolai F. Uvarov, Juliane Hanzig, Giuseppe Antonio Elia and Mateo de Vivanco Abstract: Separators and electrolytes provide electronic blockage and ion permeability between the electrodes in electrochemical cells. Nowadays, their performance and cost is often even more crucial to the commercial use of common and future electrochemical cells than the chosen electrode materials. Hence, at the present, many efforts are directed towards finding safe and reliable solid

This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Nestler, T., Roedern, E., Uvarov, N. F., Hanzig, J., Elia, G. A., de Vivanco, M. Separators and electrolytes for rechargeable batteries: Fundamentals and perspectives Physical Sciences Reviews [Online] DOI: 10.1515/psr-2018-0115

4.6 Separators and electrolytes for rechargeable batteries

175

electrolytes or liquid electrolyte/separator combinations. With this comprehensive review, the reader is provided with recent approaches on this field and the fundamental knowledge that can be helpful to understand and push forward the developments of new electrolytes for rechargeable batteries. After presenting different types of separators as well as the main hurdles that are associated with them, this work focuses on promising material classes and concepts for next-generation batteries. First, chemical and crystallographic concepts and models for the description and improvement of the ionic conductivity of bulk and composite solid electrolytes are outlined. To demonstrate recent perspectives, research highlights have been included in this work: magnesium borohydride-based complexes for solidstate Mg batteries as well as all-in-one rechargeable SrTiO3 single-crystal energy storage. Furthermore, ionic liquids pose a promising safe alternative for future battery cells. An overview on their basic principles and use is given, demonstrating their applicability for Li-ion systems as well as for so-called post-Li chemistries, such as Mg- and Al-ion batteries. Keywords: solid electrolytes, ionic liquids, Mg-ion battery, Al-ion battery, solid electrolyte–electrode interface, separators

4.6.1 Introduction While the electrodes are the electrochemical active materials that determine the overall reachable energy density and efficiency, the choice of the separator and the electrolyte is not less important. In fact, they mainly influence cycle stability and safety. For instance, rechargeable Li cells commonly utilize polymeric separators whose pores are filled with liquid flammable electrolyte, since it is cheap and guarantees high power densities. However, problems arise from different failure mechanisms during cell operation, which can affect the integrity and functionality of these separators. In the case of excessive heating or mechanical damage, polymeric separators can become an incalculable security risk [1]. In the recent past, this impaired the reputation of companies as Tesla [2] and caused losses of billions of euros for Samsung [3]. In order to minimize these risks, the so-called composite separators have been suggested that, however, still struggle with problems due to the organic electrolyte. A higher temperature stability and reliable battery operation can be achieved with full ceramic electrolytes. Such solid electrolytes have been commercially mainly used for high-temperature operation systems up to now, due to generally low ion conductivities at room temperature. The scientific progress in the last years and the demand on higher safety, cycleability and volumetric energy density pushes the release of all-solid-state Li cells. Starting with a short topical overview of available separator materials and technologies by summarizing the authors previous review on this topic [1], different solid electrolyte classes and

176

4 Battery Materials

approaches to improve their performance are illuminated here. Furthermore, ionic liquids (ILs) offer high safety and improved cell performance and are thus discussed here as alternative for conventional electrolytes in Li-ion batteries, as well as enabler for post-lithium systems.

4.6.2 Fundamentals and categorization The oldest known separator materials originate from the first experiments by A. Volta at the end of the eighteenth century. Volta demonstrated the generation of electricity from his voltaic pile using a cloth as separator, which was soaked in a sodium chloride solution [4]. Later, a variety of other separator materials became known, manufactured from cedar shingles and sausage casings, cellulosic papers, wood, cellophane, nonwoven fabrics, foams or microporous polymeric membranes [5–8]. The development of dedicated separators in the second half of the twentieth century was particularly driven by the emerging chemical industry [8]. A detailed overview of separator materials and technologies for batteries is given by Arora et al. [7] and Daniel and Besenhard [8]. A review, dedicated to the technology of Li-ion batteries only, has been published by Huang [9]. The term separator emphasizes its functionality to spatially isolate the anode from the cathode. Its function is to prevent electric contact of the electrodes, which would cause self-discharge, but simultaneously allows ion flow as basis for energy delivery. The ion flow can be conducted by the separator itself or by liquid electrolytes that fill the pores of a separator membrane. Thus, in this regard, separators can be categorized as ion conductive (solid electrolytes) or ion permeable (previous membranes) (Figure 4.6.1). This presented subdivision is based on characteristics such as physical properties, crystallographic features, morphology and composition and has been already published in Ref. [1]. Other separator categorizations can be found elsewhere [7]. Prominent examples of the first category in Figure 4.6.1 are industrially produced microporous films made from polymers as polypropylene (PP), polyethylene (PE) or polytetrafluoroethylene (PTFE), naturally abundant materials (e.g. cellulose [10] or rubber [7]), as well as nonwovens. Nonwovens are defined as sheets, webs or matts manufactured from fibres that are bonded by adhesion, cohesion, chemical bonds, friction or heat treatment. However, paper and woven, sewn or tufted products are excluded [7]. Commercially used nonwovens have been made of PE, PTFE, polyethylene terephthalate (PET) or polyvinylidene difluoride (PVDF), mainly produced by a wet-laid process [9]. In commercial batteries, separators made from microporous polymeric films dominate [11], which are filled with Li salts in organic solvents. As these materials and production methods are well known [7], in the following paragraphs, the focus will be on the emerging technologies and compounds in the field of solid electrolytes and ILs instead.

177

4.6 Separators and electrolytes for rechargeable batteries

Separator Spatial and electrical isolation of the electrodes

Ion conducting

Ion permeable

Solid electrolyte

Pervious membrane

Ion exchange Enables redox reaction at the electrodes

Polymer electrolyte

Ceramic

Glass

Porous polymer membrane

Composite

gel

crystalline

amorphous

organic

inorganic + organic

Others e. g. nonwovens wovens naturally abundant materials

separator is ion conductive liquid electrolyte needed to provide ion conductivity

thermal stability

mech. stability and impermeability

cycle life

ion conductivity

flexibility

production costs

Figure 4.6.1: Categorization scheme of separator materials. A ‘traffic light system’ denotes advantages and disadvantages of the respective materials. Reprinted from Ref. [1], with the permission of AIP Publishing.

In a more detailed view, separators should fulfil a set of specifications. On the one hand, they should preferably enable good ion conductivity to ensure high energy and power densities. For example, the ion conductivity of polymer membranes, based on poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) co-polymer hosts, filled with liquid electrolyte is 3.5  10 – 3 S/cm for Li-ion battery systems at room temperature [12]. On the other hand, separators should exhibit an insignificant electronic conductivity to prevent self-discharge. For instance, a typical value of the PVDF electronic conductivity is 10 – 13 S/cm [13]. For the common case of ion permeable separators, the ion conductivity greatly depends on the porosity of the separator material and the soaked in electrolyte, respectively [12]. Furthermore, the ability of efficient wetting with the liquid electrolyte is crucial. Additional conditions have to be met in order to ensure a reliable and safe cell operation, which apply to every type of separator:

178

– – – – –

4 Battery Materials

chemically inert towards the electrolyte and the electrodes (resistance to oxidation/reduction), sufficient puncture and tensile strength, bendability, structural integrity over the operating temperature range of the electrochemical cell, prevention of thermal runaways in the case of a short-circuited battery.

The term thermal runaway denotes the situation, when an increased temperature leads to the release of thermal energy, which in turn, accelerates the underlying process. For instance, production failures can lead to tiny short circuits in battery cells. These spots will experience extremely high current densities that can lead to the destruction (melting) of the surrounding separator material and thus even higher current flows. This positive feedback mechanism may eventually result in temperatures that are sufficient to inflame organic electrolytes and cause explosions of sealed batteries. From the production point of view, the most important properties are mechanical stability and simultaneous bendability, which are needed to let separators withstand the cell assembly process, as well as the volume changes of up to 300 % in case of operating Li-ion batteries [14]. In order to increase the power and energy density of batteries, the thickness of the separator is reduced to a minimum, which can compromise mechanical stability and structural integrity. Different failure mechanisms, as the growth of metallic dendrites or chemical attack and oxidation, may lead to the perforation of the separator and finally a short circuit and potentially thermal runaway [15]. Depending on the particular storage chemistry and technology, separators are exposed to different stress levels affecting the separator properties and thus cell performance. First of all, changes in environmental conditions can be mentioned, e.g. variations in temperature of up to 100 °C, as is typical for the operation of leadacid batteries in automobiles. Li-ion cells can reach temperatures of up to 65 °C [16] during discharge, whereas in the case of a short circuit a temperature increase of even more than a 100 °C within a few seconds may occur [17]. For the operation of sodiumsulfur batteries or high-temperature solid oxide fuel cells, temperatures of around 300 °C [18] and T > 500 °C [19], respectively, are necessary for normal operation, so that high-temperature stability of the separator is required. With regard to reports about burning electric vehicles [2] and smart-phones [3], the safety of battery systems has turned into focus even more. One strategy to improve fire safety is to utilize high melting temperature polymers, such as polyamide and polyimide, or the combination of different polymers [9]. The latter could prevent thermal runaways by improving the shutdown ability of the separator while simultaneously sustaining mechanical integrity [9]: If the battery overheats during

4.6 Separators and electrolytes for rechargeable batteries

179

operation, the polymer with the lower melting point softens and closes the pores of the polymer membrane, thus shutting down the ion conduction and thus battery operation. Simultaneously, the polymer with the higher melting point ensures stability and suppresses short circuits. However, the shutdown is often incomplete and does not block all conductivity. Thus, the cell temperature may continue to rise, which further degrades the stability of the separator. An option to minimize the risk of thermal damage of polymer-based separators is the blending or coating with ceramic particles as Al2O3, SiO2 and ZrO2 [11, 15, 20]. These materials are termed as composite separators or ceramic enhanced separators [9]. Reported advantageous properties are short circuit prevention up to 220 °C [21] and negligible shrinkage at high temperatures [15], excellent wettability and comparable high thermal conductivity. The latter promotes fast heat dissipation, preventing damage of the electrodes [9, 15, 22]. Additionally, a higher overcharge abuse tolerance was observed. Even though the risk of a thermal damage is significantly reduced by composite ceramic separators, a high-temperature use is still limited by the melting point of the organic matrix and degradation mechanism connected to the liquid electrolyte. Solid ion conductive separators, in contrast, eliminate the risk of thermal runaways, are less prone to side reactions with other battery components and also allow for highvoltage electrodes. Thus, they have become attractive for the next generation of Li cells [23, 24]. These materials can be subdivided into polymer, ceramic and glass solid electrolytes and will be thoroughly discussed here. In summary, a thorough selection of separator materials and their careful processing during fabrication and cell assembly play a key role in the operation of electrochemical storage devices. The ideal separator has to be selected with regard to the electrolyte and electrode requirements and the field of application. There will often be a trade-off between high power/energy density and a long and safe service life with constantly high coulomb efficiencies.

4.6.3 Solid electrolytes Solid electrolytes are materials with fast ionic transport of one ionic species, while there is no or no significant electronic conductivity.1 Fast ionic conductivity at ambient temperatures in solids was found for the first time in α-AgI [25] at the beginning of the last century. As has been already stated in Ref. [1], today, there is a broad variety of single-crystalline, polycrystalline and amorphous materials to choose from, which most commonly belong to the group of oxides (e.g. β-Al2O3 [18, 26]), sulfides (e.g. Li10GeP2S12 [27]) and phosphates (e.g. LiZr2P3O12 [28]). There are

1 If there is an additional comparably high electronic conductivity, they are called mixed conductors, which are good candidates for intercalation electrodes but not as separator.

180

4 Battery Materials

numerous reports on materials that conduct mono- and divalent ions such as Ag + [29–31], Na + [18, 32–35], Li + [36, 37], K + [38], H3 O + [39], Tl + [40], NH4+ [41], Cu + [42], F – [43–45], OH – [46], O2 – [39, 47, 48], Ca2 + [49], Sr2 + [49, 50], Ba2 + [49, 50], Zn2 + [49, 51], Cd2 + [49], Pb2 + [49] and Hg2 + [49]. In contrast, compounds for tri- or even tetravalent ions as Gd3 + [52–54], Al3 + [55] or Zr4 + [55] are controversial [56, 57] and not unequivocally proven, which will be briefly discussed below. Solid electrolytes are of prime importance for batteries, fuel cells and sensors. The recent advent of solid electrolytes, however, can only be explained with regard to the challenges faced by current Li and emerging post-Li systems. Currently, small mobile consumer electronics, hybrid and electric cars and even some stationary batteries for the storage of renewable energies employ Li-ion batteries that utilize liquid electrolytes with flammable organic solvents. Within the last years, this caused occasionally fire or even explosion, damaging the reputation and profit of wellknown companies [2, 3]. The safety of solid electrolytes is expected to be guaranteed, since they are mostly non-flammable. The even more important driver for the research on solid electrolytes is the demand on batteries with higher energy densities and longer lifetime for the development of mobile devices and especially the spread of electric vehicles. As already mentioned, solid electrolytes are much less vulnerable to cause side reactions as conventionally used carbonate-based electrolytes permit only negligible self-discharge, and cannot leak, boil, freeze or ignite. All-solid-state batteries hold a promise to increase both the cycle life due to their stability [58–60] and also the energy density because of generally higher electrochemical windows. Both aspects allow a more flexible choice of electrodes with higher voltages or capacity, as pure Li anodes [28] and polyanion cathodes [61]. Furthermore, metalair batteries [37, 62], which are becoming especially important for EVs, could benefit from solid electrolytes, as they would prevent the reaction between the metal anode and the electrolyte or unwanted side products [62, 63]. Another advantage from a technical point of view is that there is no need for the time-consuming, electrolytefilling step during battery assembly [64]. Additionally, simplified battery structures can be designed: In the case of cell stacks, all batteries can be mounted in one container, instead of connecting individual containers as necessary for the use of liquid electrolytes [37]. On the other hand, the rigid nature of ceramics poses problems to cell winding and assembly. It also leads to stress during charge and discharge due to the associated volume change of the electrodes, which can result in cracking of the separator and in delamination of the sintered electrode layers. It should be noted that this issue is less relevant for intercalation electrodes, as they do not cause large volume changes [37]. Moreover, extremely thin solid electrolyte separators (< 0.5 μm) can become flexible enough for winding. Further disadvantages of ceramic separators are the reduced electrode/electrolyte contact area and the overall lower ion conductivity at room temperature in comparison to liquid electrolytes. However, e.g. in the case of

4.6 Separators and electrolytes for rechargeable batteries

181

Li ions, a ceramic material (Li10GeP2S12) with a conductivity comparable to conventional organic electrolytes (10 – 2 S/cm) has been found [27]. Moreover, a strategy to avoid the high-resistance problem of the electrode/solid electrolyte interface (SEI) and to simultaneously keep the high-temperature stability is to use an IL as wetting agent, which has been already demonstrated by Kim et al. [65]. Plenty of Li-conducting materials have been investigated so far, and there has been a significant progress recently. Several companies promote the application of solid electrolytes in the next generation of Li batteries in electronic devices, including Cymbet, Front Edge Technology, Infinite Power Solutions, Sakti3, SEEO, Toyota, Planar Energy, SolidEnergy Systems, Quantum Scape, etc. Nevertheless, only a few compounds meet the required properties for commercialization simultaneously: sufficient room temperature conductivity, low-resistance grain boundaries and chemical inertness towards the electrodes, low production cost and environmental impact [66]. While for microbatteries they have been already commercialized, allsolid-state batteries for portable devices or automotive applications are still in R&D or pilot state. In the following sections, we point out the most important principles of crystallographic material design for solid electrolytes to hint on potential research directions to overcome current hurdles.

4.6.3.1 Ionic conduction in solids: Fundamentals and crystallographic requirements Fast ion transport requires high concentrations of defects (intrinsic or extrinsic), as the ions need empty spots to jump into. The motion of ions in solids can occur via interstitial sites, vacancies or along grain boundaries. The transport of point defects can be described as random walk, giving a tracer or self-diffusion coefficient of the ions (e.g. in [67]) D* =

Γ < r2 > f, 2d

(4:6:1)

where r is the jump distance between occupied and vacant site, Γ is the frequency of ion hopping between two sites and d is the dimensionality of ion conduction (1, 2 or 3). Additionally, a correlation factor f (1≤f > 0) is needed, as the direction of successive jumps is not perfectly random. This fact can be easily understood by imaging a tracer ion that just jumped into a vacancy: the chance of the reverse jump is much higher than into the partially occupied other neighbouring sites. Thus, f needs to take into account the topology of the path, the mechanism and also the defect density [68, 69]. Furthermore, in crystals, the diffusion can be imagined as thermally activated Brownian motion of ions across a periodic landscape of potential barriers. Generally, the ion conductivity increases with higher temperatures. On the one hand, the concentration of charge carriers may be enhanced due to the additional thermal energy that might be needed for defect creation in the first place. On the other

182

4 Battery Materials

hand, the ions possess more energy to overcome the activation barrier Emig for ion hopping inside the migration channel. Both the hopping frequency Γ and the concentration of charge carriers c thus follow Arrhenius law [70, 71]: Γ = pνe

c = c0 e

E – mig ,

–

kB T

(4:6:2)

Ef kB T ,

(4:6:3)

with kB being the Boltzmann constant, T the absolute temperature, p the number of potential jumping paths, c0 the charge carrier concentration at infinite temperature and Ef the defect/charge carrier formation energy. The hopping attempt frequency ν can be estimated by the Debye frequency. Finally, these considerations connect to the measured ion conductivity σ within an electric field via the Nernst–Einstein/ Einstein–Smoluchowski relation σ=

D* HR cQ2 , kB T

(4:6:4)

where Q is the electric charge of the conduction ion and HR is the Haven ratio, which is close to 1 [69]. This factor is needed to include the deviation of D* describing an undirected diffusion to the rather concerted walk: When an electric field is applied, the ions perform successive jumps, since they cannot only jump in the nearest vacant site, but also in (formerly) occupied sites. By finally combining Eqs. 1–4, it can be shown that the conductivity σ follows the temperature behaviour σ=

σ0 – kEaT e B , T

(4:6:5)

with σ0 combining the described constants and Ea = Emig + Ef . Thus, by measuring the temperature-dependent conductivity and plotting logσ versus 1/T, one can determine the activation energy for ion conduction Ea , which is in the range of 0.25 eV to 1 eV for sodium conductors [18]. Such plots are shown for several solid Li conductors and organic liquid electrolytes in Figure 4.6.2. In the following paragraphs, chemical and crystallographic conditions that facilitate or enhance ionic conductivity in solid materials in general are discussed. Main parts of this discussion have been already published by the authors in [57]. To improve ion conductivity, two types of doping have been applied to solid electrolytes: homogeneous and heterogeneous doping. For the broadly used homogeneous doping, small amounts of foreign atoms are dissolved in the compound and substitute intrinsic ions. In this way, aliovalent ions can be incorporated in the ion conductor. Depending on the charge difference between the doped and replaced ion, interstitials or vacancies are

183

4.6 Separators and electrolytes for rechargeable batteries

800 0

500

LISICON Li14Zn(GeO4)4

200

T (°C) 27

100

Ionic liquid electrolyte 1 M LiBF4/EMIBF4 Glass-ceramic electrolyte Li7P3S11

log[σ (S cm-1)]

Doped Li3N Li3.25Ge0.25P0.75S4

Li-β-alumina

-3

Li3.6Si0.6P0.4O4

Gel electrolyte 1M LiPF6 EC-PC(50:50 vol.%) + PVDF-HFP (10 wt%)

Li3N

Glass electrolyte Li2S-P2S5

-4

La0.5Li0.5TiO3

Polymer electrolyte PEO-LiCIO4 (10 wt% added TiO2)

-5

Organic electrolyte 1M LiPF6/EC-PC (50:50 vol%)

-6 Polymer electrolyte LiN(CF3SO2)2/(CH2CH2O)n (n=8)

-7 1

-100

New Li10GeP2S12 solid electrolyte Glass electrolyte Li2S-SiS2-Li3PO4

-1 -2

-30

2

3

LIPON

4

5

6

103 T -1 (K-1)

Figure 4.6.2: Thermal evolution of ion conductivity of Li solid electrolytes, organic liquid electrolytes, polymer electrolytes, ionic liquids and gel electrolytes (data sources are indicated in [27]). Li10GeP2S12 shows an ion conductivity that is comparable to those of the liquid electrolytes at room temperature and higher at lower temperatures, which is important for the application in electric cars. Reprinted by permission from Macmillan Publishers Ltd: Nature materials [27], copyright 2011.

formed to maintain electroneutrality. Tailored homogeneous doping can lead to higher conductivity, since either vacancies or mobile charge carriers needed for ion migration are created or Ea is lowered [45, 72–75]. For cation conductors, doping the cation sublattice, either with immobile (see i.e. in [74]) or mobile ions [72, 73, 75], can be used to engineer the site energies and thus hopping barriers. An instructive example for the latter case is the garnet-type Li conductor Li3+xLa3M2O12 (M = Te, Nb, Zr) [73]: Ab initio calculations reveal that the activation energy sensitively depends on the amount of Li, as the occupancy of the Li(1) and Li(2) site depends on the overall Li concentration. These occupancies, in turn, determine which ion migration routes are dominant. Higher Li occupancies favour a migration path with lower activation energy (Figure 4.6.3), which is consistent with experimental findings [73]. Furthermore, doping of the counterion sublattice can be utilized as well. For example, the number of point defects that participate in ionic motion in the fluoride-ion conductor BaF2 is increased, when it is doped with trivalent cations [45]. The second approach, heterogeneous doping, is typically realized by creating interfaces to other electrolytes or insulators or even to the same compound. In most cases, grain boundaries in polycrystalline materials increase the resistance.

184

4 Battery Materials

Δ E (eV)

1.6

LaO6

1.2 0.8 0.4

Li(2)

Li(1)

Li(1)

0.0

Li(2)

1.2

MO8

Li(2)

(a) Li3 La3Te2O12

Li(2)

Δ E (eV)

(b) Li5 La3Nb2O12 0.8

Li(1)

Route A

0.4 Li(2)

Li(2)

0.4

The Li Sites

24d

Δ E (eV)

LiO6 (Li2) 96h

48g

(a)

Li(2)

(c) Li7 La3Zr2O12

0.3

Route B

0.2

Li(1)

0.1 0.0

24d

Li(2)

0.0

LiO4 (Li1)

Li(2)

Li(1) corner

Li(2)

Migration Path (b)

Figure 4.6.3: Left: Crystal structure of garnet-type Li-ion conductors, showing two different Wyckoff positions for Li (Li(1) and Li(2)). Right: Calculated energy barriers for the dominant Li migration routes (blue arrows) for garnet-type compounds with different Li concentration and resulting occupancies of Li(1) and Li(2): (a) In the case of low Li concentrations, the tetrahedral sites are occupied. In order to migrate through the crystal, the ions have to hop to the neighbouring octahedral site, which shows a significantly higher site energy and is thus not likely. (b) For higher Li concentrations, Li ions already occupy the octahedral position and the migration can occur via the interstice between the octahedrons. (c) A very high Li stoichiometry results into vacancies in the tetrahedral site and different site energies in comparison to (a) and (b) due to Li-Li interaction. Starting from the octahedral site, Li is thus allowed to move via the tetrahedral position, bridging a lower activation energy. Reprinted figures with permission from [73].

However, if the interfaces provide appropriate defects or even better paths for ion hopping due to favourable ion redistribution in the space charge region, the conductivity can be enhanced in comparison to the bulk material (e.g. [30, 45, 76–80]). For instance, this is the case for the grain boundaries in the already mentioned Fconductor BaF2 [45]. As an example for a two-phase system, AgCl-Al2O3 shows increased Ag+ conductivity in comparison to bulk AgCl, due to an increased amount of cation defects at the interface [30]. For an overview on this topic and comprehensive discussion, the reader is referred to Ref. [77]. More recent studies illuminate the role of lattice-mismatch strain [78] as well as dislocations [80] for increased oxygen ion conductivity.

4.6 Separators and electrolytes for rechargeable batteries

185

For solid electrolytes for Li and other monovalent and divalent ions, numerous works and reviews have been published discussing favourable crystal structure motifs for ion migration and how doping is used to improve the ion diffusivity in inorganic solids (e.g. [75, 76, 81–88]). From all these studies as well as heuristic considerations, several factors can be extracted that provide high ion mobility in solids: 1. High symmetry, low occupancy: Vacant sites in close vicinity to each other with no or only small differences of site energies are required, in order not to form high barriers in between. Ideally, they would form a stable 3D network of tunnels, not only to provide ion conductivity in every direction, but also since 1D conductivity can be easily blocked by defects. Such features are commonly found in structures with highly symmetrical space groups and low occupancies of the atomic sites (with high multiplicity) of the mobile species. 2. Right-sized tunnels: The voids that surround the vacant sites and the bottlenecks of the connecting tunnels should be of a suitable (sufficient but also not too large) spatial dimension for the considered migrating ion. That means the formed channel should be smooth without too tight or wide sections and severe coordination changes. 3. Mobile-ion ratio: There should be a balanced ratio of the number of mobile ions and connected vacant sites, which can be optimized e.g. by aliovalent doping. 4. High-valent framework cations: To reduce the energy barrier of the channel that connects the vacant sites, the migration path should be accompanied by preferably screened counterions in the first coordination sphere. Such a screening of attractive interactions can be favoured by ions in the lattice with a higher same-sign charge as the considered mobile ion. They may form strong competitive bonds. 5. Polarizable anions The counterions should be polarizable to further reduce electrostatic interaction with the mobile ions. As stated in the first point, the sites in the migration channel need to possess similar energies. For instance, fast (/superionic) Li conductors were found to show paths that straightly connect two tetrahedral sites instead of crossing an energetically higher octahedral site [75]. Structures with this feature were found to possess bodycentred, cubic like anion frameworks (Figure 4.6.4). It was also shown that a higher volume of these anion sublattices, resulting in a higher Li-anion distance, would favour low barriers [75], as coinciding with criteria 2. Hints for the structural design can be also found by looking at insertion materials: Rong et al. [87] discovered for intercalation electrodes that the hopping barrier tends to be lower if the inserted ions occupy sites that do not show their preferred coordination. This means, in the case of Li, octahedral instead of tetrahedral sites. By this way, they already possess a high site energy, which potentially lowers the energetic difference toward the

186

4 Battery Materials

(a)

0.5 bcc (T–T) 0.4 Energy (eV)

T2 bcc T1

0.3

T2

T1 0.2

c 0.1

b

a 0.0 Li–ion migration path

(b)

0.5 fcc (T–O–T)

Energy (eV)

0.4 O1 fcc

T1

c b

T2

0.3

0.1

a

O1

0.2

T2

T1

0.0 Li–ion migration path

(c)

0.5 0.4

O1

T3

hcp T1

Energy (eV)

O2

hcp (T–O–T) hcp (O–O) hcp (T–T)

0.3 O2 0.2 T3

O1 0.1

T1

T2

c b a

T2

0.0 Li–ion migration path

Figure 4.6.4: Li-ion migration pathways (left column) in (a) bcc-, (b) fcc- and (c) hcp-type sulfur lattices with the corresponding migration barriers calculated by density functional theory (right column). Different Li paths are coloured red, green and blue, sulfur yellow, LiS4 tetrahedra green and LiS6 octahedra red. Reprinted by permission from Macmillan Publishers Ltd: Nature Materials [75], copyright 2015.

bottleneck they have to pass during diffusion. This is also in good agreement with the example given earlier concerning garnet-type Li conductors [73] (Figure 4.6.3). In general, a minimal change of coordination between the stable and intermediate sites appears to be advantageous for high-valent ions [87]. This is why structures with close-packed anion sublattices such as spinels or olivines might be penalized [87], as the change in the local environment during ionic motion is more drastic than for other anionic packings.

4.6 Separators and electrolytes for rechargeable batteries

187

Even though the second criteria (spatial fit of migration channel and mobile ion) might appear logical, for decades it was generally accepted that the mobile ions need to be small to show good conductivity [81] and the wider the opening of the migration channels the better. From this point of view, Al3 + with an effective ion radius of 39 pm in comparison to 59 pm for Li+ (both at the coordination number four [89]) would seem to easily migrate through the channels in crystal structures. Many authors indeed argued this property would facilitate finding intercalation compounds (e.g. [90]). However, small size does not guarantee best ion conductivity: Among monovalent ions in β-alumina, it is not Li+, but the medium-sized Na+ and Ag+ ions that display the lowest activation barriers (Figure 4.6.5). In fact, the small size and thus high charge density of Al3+ is expected to cause trapping in the lattice and at defects due to low polarizability. Indeed, finding materials was revealed to be a tough task, with only one unequivocally identified reversible intercalation material so far [91]. In general, a high charge will lead to stronger Coulomb interactions and therefore comparably higher activation energies. Apart from these considerations, how to find a suitable environment for the considered mobile species? Unfortunately, the suitable spatial dimension cannot be simply estimated from the ionic size. For instance, even though Mg2+ and Zn2+ possess almost the same ionic radii (72 pm and 74 pm [89]), very different diffusion barriers have been calculated in the same

0,40

0,35

Li+

0,30 EA (eV)

Rb+ 0,25

K+

0,20

0,15 Ag+

Na+ 0,10 0,7

0,8

0,9

1,0

1,1

1,2

1,3

1,4

1,5

1,6

rCN=6 (Å) Figure 4.6.5: The activation energies Ea for the diffusion of different monovalent ions in β-alumina [31] versus the effective ionic radii of the respective mobile ions in octahedral coordination rCN = 6 [89]. The dashed line is meant to solely serve as guide to the eyes.

188

4 Battery Materials

well-known oxidic intercalation materials (spinel Mn2O4, olivine FePO4, NiO2, δV2O5) [87]. Rong et al. ascribe this to different preferences of the ions for octahedral or tetrahedral sites, which results in opposed preferred topologies of the pathway as described above. The preferred ionic surroundings can be taken into account by bond valence (BV) analysis, as outlined in [92] or [57]. It utilizes the experimental finding that the ideal bond length can be described as depending on the ion–counterion pair and coordination number, applying data-mined constants from the topological analysis of thousands of crystal structures. It can be used to find and evaluate potential migration channels, similarly to the topological analysis by Voronoi Dirichlet partitioning [92]. The listed criteria are also relevant to the local atomic environment of amorphous or vitreous conductors. Furthermore, they must be met by compounds for high-valent (|charge| >2) solid electrolytes as well. Especially the screening of the attractive coulomb interactions by ions in the lattice with a higher same-sign charge as the considered mobile ion should be an important prerequisite. For cation conductors, this can be achieved by incorporating W6+, Mo6+, Mn4+, P5+ or Si4+. Furthermore, the presence of other ions with a lower net charge in the compound has to be examined critically due to their higher likelihood of migration. These comprehensible but still heuristic considerations, however, could not be conclusively proven so far. In contrast to mono- and divalent species, the rather controversial question, whether trivalent ions can be conducted in crystalline solids, has not been unambiguously clarified yet. As already mentioned, since the higher charge results in increased coulombic interactions, high-valent ions are expected to much more likely stay fixed on their crystallographic site. Nevertheless, the work groups of Farrington [54] and later on Imanaka [55] claimed to have synthesized compounds that are able to conduct trivalent or even tetravalent ions. However, it appears that there is still a lack of direct and unequivocal experimental evidence for this phenomenon. For example, measurements of X2 (BO4)3-type compounds (X = Sc, Al, In, Lu, Yb, Tm, Er; B = Mo, W) have been interpreted as proving X3 + conduction by Imanaka et al. since 1995 [55, 93, 94]. However, the performed experiments were not suited to doubtlessly demonstrate that. Eventually, several theoretical simulations [95–100] as well as experiments [56] employing the more reliable Tubandt method proved this wrong and showed anion, dominantly O2 – , conduction instead [56]. A comparison between the potential pathways/accessible sites for different ionic species calculated by BV and molecular dynamics is shown in Figure 4.6.6, showing that Sc3+ conduction is rather improbable in Sc2(WO4)3 [96]. Up to now, high-valent ion conduction is still under debate, given that there are only few works on this field and the direct verification is challenging. With respect to the promising applications of such materials, in aluminium-ion batteries, however, this topic is starting to get into focus [57]. To sum up, for superionic transport, there must be a multitude of energetic similar and spatially close sites, which show a low occupancy of the considered mobile ionic species and smooth pathways in between. Despite the structural aspects

189

4.6 Separators and electrolytes for rechargeable batteries

(a)

(b) 1.5

MSD / Å

WO421.0 Sc3+ 0.5 T=1073K 0.0

(c)

(d)

0

50

100 t / ps

150

Figure 4.6.6: Isosurfaces for a bond valence mismatch of 0.35 valence units for (a) Sc3+, (b) O2- and (c) WO42- in Sc2(WO4)3. These isosurfaces can be interpreted as showing the potentially accessible sites and pathways for the respective ion. (d) The mean squared displacement of WO42- in a molecular dynamics simulation of Sc2(WO4)3 confirms its mobility, in contrast to Sc3+ ions. Reprinted from Solid State Ionics, Stefan Adams, From bond valence maps to energy landscapes for mobile ions in ionconducting solids, 1625–1630, Copyright (2006), with permission from Elsevier.

of the bulk material, the performance of batteries with solid electrolytes will be even more determined by the stability and ion transport properties of the electrolyte/ electrode interface [101–105]. Therefore, the next section will focus on this aspect of solid conductors, which has been only marginally considered for a long time. 4.6.3.2 The solid electrolyte – electrode interface As already mentioned in the introduction, introducing ceramic particles to polymerbased separators leads to a significant improvement of their properties. Such composite separators or ceramic enhanced separators effectively prevent short circuit and enable fast heat dissipation and tolerance of overcharge. Nevertheless, the presence of polymers in the composites limits the operating temperature range due to their comparatively low melting points. Other drawbacks include their often limited oxidative stability, etc. In this regard, solid electrolytes are preferable as they have high mechanical strength and excellent thermal stability and some of them have also high electrochemical and chemical stability when contacted with electrode materials, including metallic Li or Li-based anode electrode materials and high-voltage electrodes for Li batteries.

190

4 Battery Materials

Despite their evident advantages, the application of solid electrolytes for electrochemical energy storage devices is difficult. For instance, the properties of an ideal single-crystalline separator will depend on the orientation of the crystallographic direction relative to the direction of the electric field. Real solid materials consist of randomly oriented grains with different types of faces having different specific surface properties and surface area. Therefore, one must be aware that the properties of each particular solid material are determined by the statistical average properties of the ensemble of individual grains. In first approximation, a so-called brick-wall model, it can be proposed that all grains in the material have the same surface type, shape and surface area. Let us first consider the case of a single-phase solid-state separator in contact with the crystalline electrode (second phase) of the electrochemical device. Then several possible situations can occur (Figure 4.6.7):

(a)

(b) Solid Electrolyte

Electrode

(c)

Figure 4.6.7: Scheme of the electrode/electrolyte interface (red) in the case of (a) poor contact, (b) good contact and (c) the formation of an insulating phase (green) at the interface.

a. If the adhesion between the phases is poor, then the solid electrolyte/electrode interface comprises an ensemble of point contacts with a small total contact area diminishing over time. In this case, the exchange current is small and such electrodes cannot provide high and stable current densities, despite a high ionic conductivity of the solid electrolyte. b. If the adhesion between the phases is strong enough, an epitaxial contact layer forms at the interface. The properties of the interface layer depend on the type of the solid electrolyte surface that is in contact with the electrode, relative orientation of the grains and the misfit between the lattice parameters of the phases. c. When the adhesion between the phases is very strong, a chemical interaction may take place, resulting in the formation of interface phases. This situation can be observed for example for Li metal electrodes in contact with organic solvents

4.6 Separators and electrolytes for rechargeable batteries

191

where a so-called SEI forms [106], which covers the surface of the Li anode and prevents it from further reaction with the organic solution. However, such interfaces may only have low ionic conductivity. In the optimal case (b), the most stable contact can be found, when no insulating layer forms between the phases. However, even in this case, both the electrolyte/ electrode contact and the formed interface layer may be easily broken on cycling due to a strong volume change of the electrode material resulting from the electrochemical processes. Freshly formed cracks act as potential sources of dendrite growth on metallic anodes of batteries. An example of such behaviour is the Na-beta-alumina membrane for Na-S batteries with molten Na and liquid sodium polysulphides electrodes. These batteries suffer from dendrite formation and their penetration through the solid electrolyte membrane. Similar processes are observed in Li batteries [107]. Another problem of solid electrolytes having high ionic conductivities often is a rather high grain resistance. The reason of this effect is the existence of the surface potential, which can cause an electric double layer that is depleted of the respective charge carriers. For example, the formation of a subsurface so-called diffuse layer depleted of anionic vacancies is responsible for the high grain resistance of oxygen-ion conductivity in zirconia-based solid electrolytes [108–110]. It has been demonstrated that the surface potential in ionic compounds is determined by the adsorption energies of the dominant defects (VO ) and the segregation energy of the dopant ions (YZr’) [111, 112]. Both may be estimated by the Stern model in combination with the Gouy–Chapman and Mott–Schottky models for calculating the potential and concentration distributions in the diffuse layer. As already mentioned, one main approach for the development of solid electrolytes is to choose a basic compound with suitable crystal structure and doping this compound with different, usually heterovalent, dopants. As a result, the conductivity may be increased at some optimal concentration of defects. However, the dopants can easily segregate at the grain boundaries leading to an increased grain resistance. At present, the question remains open how to control the dopant segregation effect in solid electrolyte ceramics. Doping in high concentrations may cause the formation of impurity phases at grain boundaries that are another reason of the degradation of solid electrolytes. In order to obtain samples with negligible contributions of grain boundaries, one should obtain dense ceramics with large grain sizes. For example, dense conducting ceramics of solid oxygen-conducting electrolytes can be obtained only after prolonged sintering at temperatures not lower than 1,400–1,500 °C. However, sintering of Li-ion conducting ceramics based on compounds with LISICON, perovskite or garnet-type structure is accompanied by partial evaporation of Li oxide. This causes serious problems with the control of the chemical and phase composition. Hence, the preparation of solid electrolyte with high ionic conductivity and low grain boundaries resistance represents a challenging task.

192

4 Battery Materials

Alternative materials for separators in batteries are composite solid electrolytes. As opposed to superionic solid electrolytes, in composite solid electrolytes heterogeneous additives produce many interfaces, which enable enhanced conductivity of all non-superionic ionic compounds. A schematic representation of superionic ceramic materials, polycrystalline salts and composite solid electrolytes is presented in Figure 4.6.8.

(a)

(b)

(c)

Superionic ceramics

Ordinary ionic salts

Composite solid electrolytes

(d)

Nanocomposite solid electrolytes

Figure 4.6.8: Comparison of superionic ceramics, ordinary salts and composite solid electrolytes. Red areas relate to regions with a high concentration of defects and high ionic conductivity, regions coloured in orange have moderate conductivity, blue areas correspond to regions with low ionic conductivity and particles of the insulating additive are coloured in green.

The increased ionic conductivity can be explained by the space charge model proposed by Wagner and Maier [113, 114]. Accordingly, chemical adsorption of ions occurs at the surface of the heterogeneous component, leading to the formation of a dense adsorption layer at the interface and a diffuse layer of vacancies near the interface. This process can be described by the Stern model. Consequently, the conductivity of composites increases as the size of dopant particles decreases. Hence, composites with nanosized grains (about 10 nm) are of particular interest for practical applications. Adding such oxides to the ionic component may produce a nanocomposite with properties that strongly depend on the character of the interface interaction between the phases. For composites with coarse-grained additives, the presence of interfaces has no appreciable effect on the bulk structural or thermodynamic properties of the ionic salt. In nanocomposites, the salt may transform to a thermodynamically metastable (usually amorphous) phase, which is stabilized by the interface interaction. The concentration and the effective thickness of this phase can be estimated from the dependence of the phase transition enthalpy on the volume fraction of the additive. At a sufficiently high concentration of the additive, the ionic salt may completely transform to a new disordered phase [115]. Today, many composite solid electrolytes are known. Due to their high ionic conductivity, solid electrolytes of the oxide salt type attract great interest especially for the application in batteries and other electrochemical devices. The combination of high conductivity with enhanced mechanical strength, the absence of grain

4.6 Separators and electrolytes for rechargeable batteries

193

resistance and excellent manufacturability make these composites promising materials for commercial applications. For example, composite solid electrolytes based on LiI have been used in solid-state Li batteries for pacemakers [116]: The battery comprises a Li anode, a composite solid electrolyte LiI-Al2O3 with various additives and an iodine cathode. This primary battery provides a voltage of (2–2.8) V depending on the anode material, with the specific energy of the battery being in the range of (150–250) mWh/g. These batteries are characterized by durability with a shelf life of more than 20 years without significant self-discharge and high reliability [117]. The high reliability of solid-state Li batteries is evidenced by the fact that from 1972 to 2000, the Catalyst Research Corporation has produced 150,000 batteries for pacemakers, and none of the batteries failed during these 17 years. The only drawback of such batteries is a small current density. The use of composite solid electrolytes opens ways to avoid the problem of dendrite growth as the mechanical strains in the composite and nanocomposite materials easily relax due to their effective redistribution along the composite matrix. As a result, the material is much more stable against deformation and does not crack during the electrochemical charge–discharge cycling. A significant advantage of composite solid electrolytes is the possibility to enlarge the electrochemically active electrolyte/electrode area by several orders of magnitude, leading to a strong enhancement of the working current of the device. This requires a special morphology of the composite and the solid electrolyte/electrode intermediate layer. The morphology of the composites exerts a great influence on their electrical properties. For the description of the conductivity of the common composites of the ‘conductor–insulator’ type, a mixing equation was proposed [118]: σα = ð1 – f ÞσαMX + f σαA ,

(4:6:6)

where σMX is the conductivity of ionic conductor MX; σA is the conductivity of the insulating heterogeneous additive A; and f is the volume fraction of the additive. In traditional mixing rules, the parameter α is taken constant: α = 1 and –1 for oriented composites consisting of parallel layers of the components when the conductivity is measured in parallel and perpendicular directions, respectively; α = 0 and 1/3 correspond to the Lichtenecker and Landau–Lifshitz equations, respectively. The analysis shows that the parameter α is determined by the composite morphology and may vary with the concentration [118]. It was assumed that α can be approximated by a linear dependence α = ð1 – f Þα1 + f α2 ,

(4:6:7)

where α1 and α2 are determined by the morphology of the composites in the diluted limit f ! 0 and f ! 1, respectively. The mixing equation may be also used for estimating the dielectric permittivity ɛ of composites

194

4 Battery Materials

ε = Reðε* Þ =

Imðσ* Þ , ωε0

(4:6:8)

where Re(ε* ) and Im(σ* ) are the real part of the complex dielectric permittivity and imaginary part of complex conductivity, respectively; ω is the angular frequency of the alternating electric field and ε0 is the dielectric constant of vacuum. The generalized mixing rule with the concentration-dependent parameter α provides a satisfactory description of the percolation-type behaviour. In particular, for the ensemble of randomly distributed isolated spherical particles of one component embedded in the matrix of another component α1 = 2/3 and α2 = -1/3. In this case, the dielectric permittivity measured at low frequencies goes through a maximum corresponding to the percolation threshold. For composite solid electrolytes, the mixing rule may be represented in the following form: σα = ð1 – f – fS ÞσαMX + fS σαS + f σαA

ðfS < f Þ,

(4:6:9)

where fS and σS are the total concentration (volume fraction) and the conductivity of the interfacial regions, respectively. The parameter α(f) depends on the concentration as mentioned above. As seen from this equation, additional contribution of the surface to the overall conductivity of the composite is determined by the second term. The conductivity of the interface regions depends on the interface interaction between the components and the concentration of the interfacial regions, which is determined by the morphology of the composite. For statistically distributed grains of both components, the fS value can be estimated to be [118] fS =

2βλ f ð1 – f Þ, LA

(4:6:10)

where β is the geometric factor (β = 3 for cubic particles); λ is the thickness of the interface layer and LA is the particle size of the heterogeneous additive. If σS > σMX , one should observe a conductivity maximum for a specific volume fraction (Figure 4.6.9), which is dependent on the particular values of the parameters σMX , α(f), fS and σS . Simultaneously, a dielectric permittivity maximum is observed at the concentration corresponding to the percolation threshold. Composites with another type of morphology form when the heterogeneous component penetrates the material in a way that it spreads along the space between the grain boundaries and forms a so-called non-autonomous interface phase, which is stabilized by the interface interaction [119]. As the volume fraction of space between the grain boundaries (equal to fS ) is small, all grain boundaries get filled at low concentrations of the additive. On further increase of the volume fraction, the

4.6 Separators and electrolytes for rechargeable batteries

195

900 4 10-6 10-7

3

10-8

2

3 300

2

10-9

0.5 f

10-10 1.0

2 1

1 0 0.0

(b)

600

ε/ε0

4

(a)

σ, S/cm

σ/σMX

4

4 3 2 1

0 0.0

0.5

1.0

f

Figure 4.6.9: Theoretical dependence of (a) the dc conductivity and (b) low-frequency dielectric permittivity on the volume fraction of the additive for a composite with σMX = 1  10 – 6 S/cm, σ A = 1  10 – 9 S/cm, σ S = 1  10 – 5 S/cm and β = 3. The ratio λ=L varies between 0, 0.0167, 0.0333 and 0.05 for the curves 1, 2, 3 and 4, respectively, with α1 = 0.667 and α2 = – 0.3337. The conductivity data in (a) are represented in linear (solid lines) and logarithmic (dashed lines) scales.

spreading process stops and the composite behaves as an ordinary statistical mixture. An example of such systems are MeWO4-WO3 composites, where Me = Ca, Sr and Ba [120–122], in which MeWO4 are typical dielectrics and WO3 is a good electronic conductor without any contribution of ionic conductivity. Adding low concentrations of WO3 (f < fS ) leads to a sharp increase in the ionic conductivity by three or four orders of magnitude. As a result, composite solid electrolytes are formed out of individual components that have no appreciable ionic conductivity. At higher concentrations of WO3, the concentration dependence of the conductivity is typical for composites of the insulator-conductor type and the character of the conductivity changes from ionic to electronic. Composites formed by filling porous matrixes with ionic salts have a specific morphology and can also be used as solid electrolytes. The pore size distribution and the pore volume are key morphological parameters defining the properties of the matrix. If the pore size is less then (1–3) nm and the adhesion between MX and the surface of A is sufficiently high, the ionic salt penetrates into the pores and its transport properties drastically change [115]. Hence, the part of the ionic salt located inside pores may be regarded as the surface high-conducting phase and its volume fraction is equal to fS . An estimation shows that the dependence of conductivity of such composites on the volume fraction of the porous matrix has a maximum at f = fmax , where fmax is the volume fraction of the pores [119]. The mixing equation may be adopted to predict the conductivity and dielectric properties of ternary composites of the conductor–electrolyte–dielectric type [123]. In this case, the equation may be expressed in a more general form

196

4 Battery Materials

N P

σ

i

αi fi

0 P 1 N N αi fi X B i C = fi A , @σ i

(4:6:11)

i

where σi and fi are the conductivity and volume fraction of the ith phase; the parameters αi have the same meaning as in eq. (4.6.7) and the summation is made over all the components of the N-phase composite (for a ternary system N = 3). It was shown that in ternary composites the choice of the parameters αi is ambiguous and possible solutions to this problem have been proposed. In three-phase composites, three percolation thresholds may occur. Their positions, as lines on the ternary plots, may be determined by the dielectric permittivity maxima. The absolute values of permittivity in the maximum, in turn, is determined by the ratio of the conductivities of insulating and conducting phases. Varying the concentration of the components and the morphology also opens wide possibilities for the improvement of the solid electrolyte/electrode contact. In particular, for the increase of the contact area, one may introduce an additional transient layer containing the electrode material as third phase. The conductivity of such composites can be estimated using a mixing equation for three-phase composite as well. All three phases will play a specific role. The ionic salt enables ionic transport, the additive of the electrode material phase will provide electronic transport to the interfaces and the dispersed insulating phase will serve as both the reinforcing matrix responsible for total integrity of the composite and the heterogeneous dopant generating ion-conducting interfaces. For instance, this idea was employed for the preparation of the composite electrode–electrolyte interlayer in solid-state supercapacitors [124].

4.6.3.3 Research highlight: Mg2 + solid electrolytes for all-solid-state batteries Magnesium batteries are one of the most promising candidates for post-Li-ion battery technologies [125]. Mg is a divalent ion and can consequently store two electrons per ion, which translates to a higher volumetric energy density of 3,833 mAh/cm3 compared to 2,036 mAh/cm3 for Li. In contrast to Li, Mg can be electroplated without forming dendrites, which can cause short circuits and battery failure [126]. Mg is also safer than Li when exposed to air and has a higher natural abundance in the earth crust, making it an excellent candidate for battery anodes. Rechargeable Mg batteries were invented in the early 2000s and have received significant attention since then [127–129]. However, big challenges remain in their development and many aspects of the underlying battery processes at the electrodes, electrolytes and their interface are still poorly understood. The Mg2+ ion is characterized by a very high charge density, owing to the divalent positive charge carried and the small ionic radius of only 72 pm (Figure 4.6.10) [89].

4.6 Separators and electrolytes for rechargeable batteries

Empirical atomic radii [pm]

Effective ionic radii [pm]

Li

Na

Mg

145

180

150

Li+

Na+

Mg2+

76

102

72

197

Figure 4.6.10: Comparison of empirical ionic radii and effective ionic radii for Li, Na and Mg (data taken from [89]).

This hampers the mobility of Mg in the electrolyte and in the cathode host materials, due to the greater electrostatic interaction, compared with monovalent cations such as Na+ and Li+. A wide range of intercalation cathode materials and the mechanisms of high-valent ion insertion and migration in those was recently reviewed and numerous studies on liquid Mg electrolytes are available [130–132]. As discussed in the previous section, solid-state batteries offer many advantages over batteries with liquid organic electrolytes. This research highlight will focus on the challenge of finding solid-state electrolyte materials with sufficiently high magnesium-ion conductivity for Mg batteries. Mg-ion conductivity in inorganic solids is rare, but was reported for doublemagnesium zirconium orthophosphate, MgZr4(PO4)6 [133]. These polycrystalline solid Mg conductors reach conductivities higher than 10 – 5 S/cm at temperatures above 400 °C, and their structure and properties were studied, especially in relation to their potential application as active materials in gas sensors [133–137]. However, in order for a solid-state ionic conductor to be a useful solid electrolyte for battery applications, the ionic conductivity should ideally be close to 10 – 3 S/cm at the operating temperature, in order to allow for current densities and thereby charge and discharge rates comparable to those of Li-ion batteries [138]. Polymer-electrolyte systems, typically based on poly(ethylene oxide) (PEO) and Mg salts, were considered as solid-state electrolytes, but the initially investigated salts Mg(SO3CF3)2 or Mg(N (SO2CF3)2)2 are known to form an insulating interphase with the Mg metal anode [139]. Reversible Mg plating/stripping was demonstrated for PEO/Mg(BH4])2 composites [140], and D. Aurbach and coworkers realized a rechargeable solid-state Mg battery with PVDF/Mg(AlCl2EtBu)2/tetraglyme as gel electrolyte, Mo6]S8 cathode, and AZ-31 magnesium alloy (3 % Al and 1 % Zn) as the anode material. Chevrel phase Mo6S8 can intercalate Mg reversibly at a potential of about 1.2 V and is often used as a benchmark cathode material, despite its low capacity, for the lack of other intercalation cathodes. Magnesium borohydride Mg(BH4)2 and derived materials were successfully used in liquid electrolytes, thanks to the reductive stability of the BH4– anion. Furthermore,

198

4 Battery Materials

they are also attractive as solid-state Mg ion conductors [141, 142]. However, the conductivity of pristine Mg(BH4)2 in the solid state is extremely low ( 10 % water can impede Zn reoxidation. In TFSI, water reduces the overpotential associated with deposition [236].

4.6.5 Conclusion Current organic electrolytes and thin separators in commercial Li-ion battery systems demonstrate crucial safety deficiencies. Furthermore, the urgent need for higher energy densities drives the development of alternative electrolyte chemistries: They are needed as key enabler for next-generation Li batteries as well as high-valent battery systems. Among them, solid electrolytes are promising candidates that have already overcome the hurdle of utilizable ion conductivity, at least in the case of Li

References

209

batteries. Employing the outlined appropriate structural motifs and framework ions as well as tailored doping may also lead to utilizable mobilities for other ion species. As an example for cell chemistries with other mobile species, current research on magnesium borohydride-based complexes as electrolytes for solid-state Mg-ion batteries has been presented. Mg batteries offer many advantages over Li batteries, but the high charge density of the Mg2+ ion also poses serious challenges for the search of suitable electrolyte and cathode materials. Higher conductivities can be also reached by composite solid electrolytes, which can be described by the so-called mixing equation. However, in many cases not the bulk properties of solid electrolytes, but rather the intergrain interfaces and the solid electrolyte-electrode interface dominate the performance. Thus, the main issue is to optimize the electrode–electrolyte interface to provide a low-resistant and durable contact upon cycling. A concept to bypass the interface problem is the presented all-in-one rechargeable energy storage solely utilizing an oxygen-deficient single crystal of SrTiO3, serving as anode, cathode as well as electrolyte. This work may potentially inspire similar approaches, utilizing a field-induced redistribution of vacancies and defect separation, establishing a non-equilibrium state accompanied by an electromotive force. Another alternative, offering high safety, while not dealing with the interface problems of solid electrolytes, are ILs. Their tuneable properties, such as high ionic conductivity, low volatility, reduced flammability and high electrochemical and chemical stability offer new possibilities in battery research. At the same time, their current drawbacks, as in the case of Al-ion batteries – high prices and corrosivity against other cell components – need to be solved. Funding: TN, JH, and MdV are grateful for financial support of the Federal Ministry of Education and Research (CryPhysConcept (03EK3029A) and R2RBattery (03SF0542A)). Furthermore, ER thanks the Swiss National Science Foundation for financial support within the Sinergia project ‘Novel ionic conductors’ (CRSII2_160749/1).

References [1]

[2] [3] [4] [5] [6] [7]

Nestler T, Schmid R, Münchgesang W, Bazhenov V, Schilm J, Leisegang T, et al. Separatorstechnology review: ceramic based separators for secondary batteries. In: AIP Conference Proceedings, volume 1597. AIP, 2014: 155–184. https://electrek.co/2016/08/15/tesla-model-s-catches-fire-test-drive-france/ (2016). https://www.cnet.com/news/why-is-samsung-galaxy-note-7-exploding-overheating/ (2016). Volta A. On the Electricity excited by the mere Contact of conducting Substances of different kinds. Philos Trans R Soc 1800;2:403 Washburn E. Thirds to f. US Patent 482,444 (1892). Randell CF, White NC. Separators for electrochemical cells. WO Patent App. PCT/GB1996/ 001,318 (1996). Arora P, Zhang Z. Battery separators. Chem Rev. 2004;104:4419.

210

[8] [9] [10] [11]

[12]

[13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25] [26] [27] [28]

[29]

[30]

4 Battery Materials

Daniel C, Besenhard JO. Handbook of battery materials. Weinheim, Germany: Wiley, 2012. Huang X. Separator technologies for lithium-ion batteries. J Solid State Electrochem. 2011;15:649. Dafler JR. The ultimate polymer application: Resin-bonded cellulose separators for lead–acid batteries. J Appl Polym Sci. 1977;2551:21. Weber CJ, Geiger S, Falusi S, Roth M. International Freiberg Conference on Electrochemical Storage Materials - ESTORM 2013. American Institute of Physics: AIP Conference Proceedings, 2014. Costa CM, Gomez Ribelles J, Lanceros-Méndez S, Appetecchi G, Scrosati B. Poly (vinylidene fluoride)-based, co-polymer separator electrolyte membranes for lithium-ion battery systems. J Power Sour. 2014;245:779. Ji Y, Jiang Y. Increasing the electrical conductivity of poly (vinylidene fluoride) by KrF excimer laser irradiation. Appl Phys Lett, 2006;89:221103. Bruce PG, Scrosati B, Tarascon JM. Nanomaterials for rechargeable lithium batteries. Angew Chem Int Ed. 2008;47:2930. Zhang SS. A review on the separators of liquid electrolyte Li-ion batteries. J Power Sour. 2007;164:351. Wu MS, Liu KH, Wang YY, Wan CC. Heat dissipation design for lithium-ion batteries. J Power Sour. 2002;109:160. Barnett B, Ofer D, Sriramulu S, Stringfellow R. Lithium-ion batteries, safety. In: Batteries for sustainability. New York: Springer, 2013:285–318. Hueso KB, Armand M, Rojo T. High temperature sodium batteries: status, challenges and future trends. Energy & Environ Sci. 2013;6:734. Choudhury A, Chandra H, Arora A. Application of solid oxide fuel cell technology for power generation a review. Renewable Sustainable Energy Rev. 2013;20:430. Frost, Sullivan. High-tech materials alert 15th March 2013. Frost & Sullivan, 2013. Roth EP, Doughty DH, Pile DL. Effects of separator breakdown on abuse response of 18650 Liion cells. J Power Sour. 2007;174:579. Orendorff CJ. The role of separators in Lithium-ion cell safety. Interface-Electrochem Soc. 2012;21:61. Kim JG, Son B, Mukherjee S, Schuppert N, Bates A, Kwon O, et al. A review of lithium and nonlithium based solid state batteries. J Power Sour. 2015;282:299. Sun C, Liu J, Gong Y, Wilkinson DP, Zhang J. Recent advances in all-solid-state rechargeable lithium batteries. Nano Energy. 2017;33:363. Tubandt C, Lorenz E. Molekularzustand und elektrisches LeitvermoÈgen kristallisierter Salze. Z Phys Chem. 1914;24:513. Wen Z, Hu Y, Wu X, Han J, Gu Z. Main challenges for high performance NAS battery: materials and interfaces. Adv Funct Mater. 2013;23:1005. Kamaya N, Homma K, Yamakawa Y, Hirayama M, Kanno R, Yonemura M, et al. A lithium superionic conductor. Nat Mater. 2011;10:682. Fotheringham U, Schneider M, Hochrein O, Biedenbender S, Lauer S, Kluge M. Lithium ionconductive glass ceramic, method for producing a lithium ion-conductive glass ceramic, ion conductor and use of the ion conductor. WO Patent App. PCT/EP2013/054,732 (2013). Uneme Y, Tamura S, Kawato T, Imanaka N. Moderate temperature operative sulfur dioxide gas sensor based on an Ag+ ion conducting solid. Electrochem Solid-State Lett. 2011; 14:J38. Maier J. Space charge regions in solid two-phase systems and their conduction contribution I. Conductance enhancement in the system ionic conductor-inertphase and application on AgCl: Al2O3 and AgCl: SiO2. J Phys Chem Solids. 1985;46:309.

References

211

[31] Yao YFY, Kummer J. Ion exchange properties of and rates of ionic diffusion in beta-alumina. J Inorg Nucl Chem. 1967;29:2453. [32] de Jongh P, Blanchard D, Matsuo M, Udovic T, Orimo S. Complex hydrides as room-temperature solid electrolytes for rechargeable batteries. Appl Phys A. 2016;122:1. [33] Lu X, Xia G, Lemmon JP, Yang Z. Advanced materials for sodium-beta alumina batteries: status, challenges and perspectives. J Power Sour. 2010;195:2431. [34] Anantharamulu N, Rao KK, Rambabu G, Kumar BV, Radha V, Vithal M. A wide-ranging review on Nasicon type materials. J Mater Sci. 2011;46:2821. [35] Udovic TJ, Matsuo M, Tang WS, Wu H, Stavila V, Soloninin AV, et al. Exceptional superionic conductivity in disordered sodium Decahydro-closo-decaborate. Adv Mater. 2014;26:7622. [36] Knauth P. Inorganic solid Li ion conductors: an overview. Solid State Ionics. 2009;180:911. [37] Takada K. Progress and prospective of solid-state lithium batteries. Acta Mater. 2013;61:759. [38] Menezes PV, Martin J, Schaefer M, Staesche H, Roling B, Weitzel KM. Bombardment induced ion transport Part II. Experimental potassium ion conductivities in borosilicate glass. Phys Chem Chem Phys. 2011;13:20123. [39] Malavasi L, Fisher CAJ, Islam MS. Oxide-ion and proton conducting electrolyte materials for clean energy applications: structural and mechanistic features. Chem Soc Rev. 2010;39:4370. [40] Maier J, Reichert B. Ionic transport in heterogeneously and homogeneously Doped Thallium (I)Chloride. Berichte der Bunsengesellschaft für physikalische Chemie. 1986;90:666. [41] Imanaka N, Tamura S, Adachi G. Ammonia sensor based on ionically exchanged NH4+ gallate solid electrolytes. Electrochem Solid-State Lett. 1998;1:282. [42] Banno N, Sakamoto T, Iguchi N, Sunamura H, Terabe K, Hasegawa T, et al. Diffusivity of Cu ions in solid electrolyte and its effect on the performance of nanometer-scale switch. Electron Devices, IEEE Trans. 2008;55:3283. [43] Reddy MA, Fichtner M. Batteries based on fluoride shuttle. J Mater Chem. 2011;21:17059. [44] Patro L, Hariharan K. Fast fluoride ion conducting materials in solid state ionics: an overview. Solid State Ionics. 2013;239:41. [45] Rongeat C, Reddy MA, Witter R, Fichtner M. Nanostructured fluorite-type fluorides as electrolytes for fluoride ion batteries. J Phys Chem C. 2013;117:4943. [46] Hibino T, Kobayashi K, Nagao M. An all-solid-state rechargeable aluminum–air battery with a hydroxide ion-conducting Sb(v)-doped SnP2 O7 electrolyte. J Mater Chem A. 2013;1:14844. [47] Kharton VV, Marques FMB, Atkinson A. Transport properties of solid oxide electrolyte ceramics: a brief review. Solid State Ionics. 2004;174:135. [48] Fujii K, Esaki Y, Omoto K, Yashima M, Hoshikawa A, Ishigaki T, et al. New perovskite-related structure family of oxide-ion conducting materials NdBaInO4. Chem Mater. 2014;26:2488. [49] Farrington G, Dunn B. Divalent beta”-aluminas: high conductivity solid electrolytes for divalent cations. Solid State Ionics. 1982;7:267. [50] Dunn B, Farrington G. Fast divalent ion conduction in Ba++, Cd++ and Sr++ beta” aluminas. Mater Res Bull. 1980;15:1773. [51] Ikeda S, Takahashi M, Ishikawa J, Ito K. Solid electrolytes with multivalent cation conduction. 1. Conducting species in Mg Zr PO4 system. Solid State Ionics. 1987;23:125. [52] Carrillo-Cabrera W, Thomas JO, Farrington GC. The ionic distribution in trivalent Gd3+β”-alumina. Solid State Ionics. 1983;9:245. [53] Dunn B, Farrington G. Trivalent ion exchange in beta” alumina. Solid State Ionics. 1983;9:223. [54] Sattar S, Ghosal B, Underwood M, Mertwoy H, Saltzberg M, Frydrych W, et al. Synthesis of Diand trivalent β”-aluminas by ion exchange. J Solid State Chem. 1986;65:231. [55] Imanaka N, Tamura S, Nunotani N. Multivalence cation conductors. Encycl Appl Electrochem. 2014;1334–9.

212

4 Battery Materials

[56] Neiman AY, Pestereva N, Zhou Y, Nechaev D, Koteneva E, Vanec K, et al. The nature and the mechanism of ion transfer in tungstates Me2+WO4(Ca, Sr, Ba) and Me3+2WO3(Al, Sc, In) according to the data acquired by the tubandt method. Russ J Electrochem. 2013;49:895. [57] Nestler T, Fedotov S, Leisegang T, Meyer DC. Towards Al3+ mobility in crystalline solids: critical review and analysis. Crit Rev Solid State Mater Sci. 2018. [58] Hayashi A, Ohtomo T, Mizuno F, Tadanaga M, Tatsumisago K. All-solid-state Li/S batteries with highly conductive glass–ceramic electrolytes. Electrochem Commun. 2003;5:701. [59] Machida N, Kobayashi K, Nishikawa Y, Shigematsu T. Electrochemical properties of sulfur as cathode materials in a solid-state lithium battery with inorganic solid electrolytes. Solid State Ionics. 2004;175:247. [60] Hayashi A, Ohtsubo R, Ohtomo T, Mizuno F, Tatsumisago M. All-solid-state rechargeable lithium batteries with Li2S as a positive electrode material. J Power Sour. 2008; 183:422. [61] Hautier G, Jain A, Chen H, Moore C, Ong SP, Ceder G. Novel mixed polyanions lithium-ion battery cathode materials predicted by high-throughput ab initio computations. J Mater Chem. 2011;21:17147. [62] Mori R. A new structured aluminium–air secondary battery with a ceramic aluminium ion conductor. RSC Adv. 2013:3:11547. [63] Shimonishi Y, Zhang T, Imanishi N, Im D, Lee DJ, Hirano A, et al. A study on lithium/air secondary batteriesstability of the NASICON-type lithium ion conducting solid electrolyte in alkaline aqueous solutions. J Power Sour. 2011;196:5128. [64] Technology and market forecast of separators for rechargeable lithium ion batteries (2009– 2014). Bundang-gu, Seongnam-si, Korea: Solar & Energy Co., Ltd, 2010. [65] Kim H, Ding Y, Kohl PA. LiSICON–ionic liquid electrolyte for lithium ion battery. J Power Sour. 2012;198:281. [66] Gao J, Chu G, He M, Zhang S, Xiao R, Li H, et al. Screening possible solid electrolytes by calculating the conduction pathways using bond valence method. Sci China Phys Mech Astron. 2014;57:1526. [67] Barber MN, Ninham BW. Random and restricted walks: theory and applications, volume 10. New York: CRC Press, 1970. [68] Koiwa M, Ishioka S. Random walk properties of lattices and correlation factors for diffusion via the vacancy mechanism in crystals. J Stat Phys 1983;30:477. [69] Mehrer H. Diffusion in solids: fundamentals, methods, materials, diffusion-controlled processes, volume 155. Berlin, Heidelberg, Germany: Springer Science & Business Media, 2007. [70] Almond D, Duncan G, West A. The determination of hopping rates and carrier concentrations in ionic conductors by a new analysis of ac conductivity. Solid State Ionics. 1983;8:159. [71] Balluffi RW, Allen S, Carter WC. Kinetics of materials. Hoboken, USA: John Wiley & Sons, 2005. [72] Xie H, Alonso JA, Li Y, Fernandez-Diaz MT, Goodenough JB. Lithium distribution in aluminumfree cubic Li7La3Zr2O12. Chem Mater. 2011;23:3587. [73] Xu M, Park MS, Lee JM, Kim TY, Park YS, Ma E. Mechanisms of Li+ transport in garnet-type cubic Li3+xLa3M2O12 (M= Te, Nb, Zr). Phys Rev B. 2012;85:052301. [74] Whiteley JM, Woo JH, Hu E, Nam KW, Lee SH. Empowering the lithium metal battery through a silicon-based superionic conductor. J Electrochem Soc. 2014;161:A1812. [75] Wang Y, Richards WD, Ong SP, Miara LJ, Kim JC, Mo Y, et al. Design principles for solid-state lithium superionic conductors. Nat Mater. 2015;14:1026. [76] Sata N, Eberman K, Eberl K, Maier J. Mesoscopic fast ion conduction in nanometre-scale planar heterostructures. Nature. 2000;408:946. [77] Maier J. Nanoionics: ion transport and electrochemical storage in confined systems. Nat Mater. 2005;4:805.

References

213

[78] Pennycook TJ, Beck MJ, Varga K, Varela M, Pennycook SJ, Pantelides ST. Origin of colossal ionic conductivity in oxide multilayers: interface induced sublattice disorder. Phys Rev Lett. 2010;104:115901. [79] Adams S. Ultrafast lithium migration in surface modified LiFePO4 by heterogeneous doping. Appl Energy. 2012;90:323. [80] Adepalli KK, Kelsch M, Merkle R, Maier J. Enhanced ionic conductivity in polycrystalline TiO2 by one-dimensional doping. Phys Chem Chem Phys. 2014;16:4942. [81] Armstrong R, Bulmer R, Dickinson T. Fast ion transport in solids. Amsterdam: North Holland Publishing, 1973. [82] Goodenough J, Hong HP, Kafalas J. Fast Na+-ion transport in skeleton structures. Mater Res Bull. 1976;11:203. [83] Adachi G, Imanaka N, Aono H. Fast Li+ conducting ceramic electrolytes. Adv Mater. 1996; 8:127. [84] Levi E, Levi M, Chasid O, Aurbach D. A review on the problems of the solid state ions diffusion in cathodes for rechargeable Mg batteries. J Electroceram. 2009;22:13. [85] Adams S, Rao RP. Structural requirements for fast lithium ion migration in Li10GeP2S12. J Mater Chem. 2012;22:7687. [86] Guin M, Tietz F. Survey of the transport properties of sodium superionic conductor materials for use in sodium batteries. J Power Sour. 2015;273:1056. [87] Rong Z, Malik R, Canepa P, Sai Gautam G, Liu M, Jain A, et al. Materials design rules for multivalent ion mobility in intercalation structures. Chem Mater. 2015;27:6016. [88] Meutzner F, Münchgesang W, Leisegang T, Schmid R, Zschornak M, Ureña de Vivanco M, et al. Identification of solid oxygen-containing Na-electrolytes: an assessment based on crystallographic and economic parameters. Cryst Res Technol. 2017;52:1. [89] Shannon RT. Revised effective ionic radii and systematic studies of interatomic distances in halides and chalcogenides. Acta Crystallogr Sect A: Cryst Phys Diffr Theor Gen Crystallogr. 1976;32:751. [90] Liu S, Hu J, Yan N, Pan G, Li G, Gao X. Aluminum storage behavior of anatase TiO2 nanotube arrays in aqueous solution for aluminum ion batteries. Energy Environ Sci. 2012;5:9743. [91] Geng L, Lv G, Xing X, Guo J. Reversible electrochemical intercalation of aluminum in Mo6S8. Chem Mater. 2015;27:4926. [92] Meutzner F, Nestler T, Zschornak M, Canepa P, Gautam GS, Leoni S, et al. Theoretical methods for battery material identification and analysis in electrochemical storage materials: from crystallography to manufacturing technology. Berlin: De Gruyter Oldenbourg Publishing House, 2018. [93] Imanaka N, Kobayashi Y, Adachi GY. A direct evidence for trivalent ion conduction in solids. Chem Lett. 1995;24:433. [94] Imanaka N, Tamura S. Development of multivalent ion conducting solid electrolytes. Bull Chem Soc Jpn. 2011;84:353. [95] Driscoll DJ, Islam MS, Slater PR. Simulation and conductivity studies of defects and ion transport in Sc2(WO4)3. Solid State Ionics. 2005;176;539. [96] Adams S. From bond valence maps to energy landscapes for mobile ions in ion-conducting solids. Solid State Ionics. 2006;177:1625. [97] Zhou Y, Adams S, Rao RP, Edwards DD, Neiman A, Pestereva N. Charge transport by polyatomic anion diffusion in Sc2(WO4)3. Chem Mater. 2008;20:6335. [98] Zhou Y, Rao RP, Adams S. Mechanism of defect formation and polyanion transport in solid scandium tungstate type oxides. Monatshefte für Chemie-Chem Mon. 2009;140:1017. [99] Zhou Y, Rao RP, Adams S. Intrinsic polyatomic defects in Sc2(WO4)3. Solid State Ionics. 2011;192:34.

214

4 Battery Materials

[100] Zhou Y, Neiman A, Adams S. Novel polyanion conduction in Sc2(WO4)3 type negative thermal expansion oxides. Phys Status Solidi (b). 2011;248:130. [101] Yada C, Ohmori A, Ide K, Yamasaki H, Kato T, Saito T, et al. Dielectric modification of 5V-class cathodes for high-voltage all-solid-state lithium batteries. Adv Energy Mater. 2014;4:9. [102] Takada K, Ohta N, Tateyama Y. Recent progress in interfacial nanoarchitectonics in solid-state batteries. J Inorg Organomet Polym Mater. 2015;25:205. [103] Lepley N, Holzwarth N. Modeling interfaces between solids: application to Li battery materials. Phys Rev B. 2015;92:214201. [104] Takada K, Ohno T. Experimental and computational approaches to interfacial resistance in solid-state batteries. Front Energy Res. 2016;4:10. [105] Wenzel S, Leichtweiß T, Weber DA, Sann J, Zeier WG, Janek J. Interfacial reactivity benchmarking of the sodium ion conductors Na3PS4 and sodium β-alumina for protected sodium metal anodes and sodium all-solid-state batteries. ACS Appl Mater Interface. 2016. [106] Aurbach D. The role of surface films on electrodes in Li-ion batteries. In: Advances in Lithiumion batteries. New York: Springer, 2002: 7–77. [107] Zhang JG, Xu W, Henderson WA. Lithium metal anodes and rechargeable lithium metal batteries. Switzerland: Springer, 2017. [108] Guo X, Maier J. Grain boundary blocking effect in zirconia: a Schottky barrier analysis. J Electrochem Soc. 2001;148:E121. [109] Guo X, Waser R. Electrical properties of the grain boundaries of oxygen ion conductors: acceptor-doped zirconia and ceria. Prog Mater Sci. 2006;51:151. [110] Guo X, Sigle W, Fleig J, Maier J. Role of space charge in the grain boundary blocking effect in doped zirconia. Solid State Ionics. 2002;154:555. [111] Uvarov N. Surface disordering of classic and superionic crystals: a description in the framework of the Stern model. Russ J Electrochem. 2007;43:368. [112] Uvarov N. Estimation of the surface potential in superionic oxide conductors using the Stern model. Solid State Ionics. 2008;179:783. [113] Wagner J, Jr. Composite materials as solid electrolytes. In: Solid state batteries. Dordrecht, Netherlands: Springer, 1985: 77–90. [114] Maier J. Ionic conduction in space charge regions. Prog Solid State Chem. 1995;23:171. [115] Uvarov N. Composite solid electrolytes: recent advances and design strategies. J Solid State Electrochem. 2011;15:367. [116] Liang C, Joshi A, Hamilton N. Solid-state storage batteries. J Appl Electrochem. 1978;8:445. [117] Crompton TP. Battery reference book. Oxford, Great Britain: Newnes, 2000. eBook ISBN: 9780080499956. [118] Uvarov N. Estimation of composites conductivity using a general mixing rule. Solid State Ionics. 2000;136:1267. [119] Uvarov NF. Estimation of electrical properties of composite solid electrolytes of different morphologies. Solid State Ionics. 2017;302:19. [120] Konisheva E, Neiman A, Gorbunova E.. Transport processes and surface transformation at the CaWO4|WO3 interface. Solid State Ionics. 2003;157:45. [121] Neiman AY, Pestereva N, Sharafutdinov A, Kostikov YP. Conduction and transport numbers in metacomposites MeWO 4£ WO 3 (Me= Ca, Sr, Ba). Russ J Electrochem. 2005;41:598. [122] Neiman AY, Uvarov N, Pestereva N. Solid state surface and interface spreading: an experimental study. Solid State Ionics. 2007;177:3361. [123] Uvarov N. Phenomenological description of transport properties of three-phase composites. Russ J Electrochem. 2017;53:700. [124] Ulihin A, Mateyshina YG, Uvarov N. All-solid-state asymmetric supercapacitors with solid composite electrolytes. Solid State Ionics. 2013;251:62.

References

215

[125] Choi JW, Aurbach D. Promise and reality of post-lithium-ion batteries with high energy densities. Nat Rev Mater. 2016;1:16013. [126] Ling C, Banerjee D, Matsui M. Study of the electrochemical deposition of Mg in the atomic level: why it prefers the non-dendritic morphology. Electrochim Acta. 2012;76:270. [127] Aurbach D, Lu Z, Schechter A, Gofer Y, Gizbar H, Turgeman R, et al. Prototype systems for rechargeable magnesium batteries. Nature. 2000;407:724. [128] Aurbach D, Suresh GS, Levi E, Mitelman A, Mizrahi O, Chusid O, et al. Progress in rechargeable magnesium battery technology. Adv Mater. 2007;19:4260. [129] Chusid O, Gofer Y, Gizbar H, Vestfrid Y, Levi E, Aurbach D, et al. Solid-state rechargeable magnesium batteries. Adv Mater. 2003;15:627. [130] Canepa P, Sai Gautam G, Hannah DC, Malik R, Liu M, Gallagher KG, et al. Odyssey of multivalent cathode materials: open questions and future challenges. Chem Rev. 2017;117:4287. [131] Muldoon J, Bucur CB, Gregory T. Quest for nonaqueous multivalent secondary batteries: magnesium and beyond. Chem Rev. 2014;114:11683. [132] Yoo HD, Shterenberg I, Gofer Y, Gershinsky G, Pour N, Aurbach D. Mg rechargeable batteries: an on-going challenge. Energy Environ Sci. 2013;6:2265. [133] Ikeda S, Takahashi M, Ishikawa J, Ito K. Solid electrolytes with multivalent cation conduction. 1. Conducting species in MgZrPO4 system. Solid State Ionics. 1987;23:125. [134] Imanaka N, Okazaki Y, Adachi GY. Divalent magnesium ion conducting characteristics in phosphate based solid electrolyte composites. J Mater Chem. 2000;10:1431. [135] Imanaka N, Okazaki Y, Adachi G. Optimization of divalent magnesium ion conduction in phosphate based polycrystalline solid electrolytes. Ionics. 2001;7:440. [136] Mielewczyk A, Molin S, Gdula K, Jasiński G, Kusz B, Jasiński P, et al. Structure and electric properties of double magnesium zirconium orthophosphate. Mater Ceram. 2010;62:477. [137] Gobechiya E, Sukhanov M, Petkov V, Kabalov YK. Crystal structure of the double magnesium zirconium orthophosphate at temperatures of 298 and 1023 K. Crystallogr Rep. 2008;53:53. [138] Takada K. Solid electrolytes and solid-state batteries. In: AIP Conference Proceedings, volume 1765. AIP Publishing, 2016: 020008. [139] Mohtadi R, Mizuno F. Magnesium batteries: current state of the art, issues and future perspectives. Beilstein J Nanotechnol. 2014;5:1291. [140] Shao Y, Rajput NN, Hu J, Hu M, Liu T, Wei Z, et al. Nanocomposite polymer electrolyte for rechargeable magnesium batteries. Nano Energy. 2015;12:750. [141] Mohtadi R, Matsui M, Arthur TS, Hwang SJ. Magnesium borohydride: from hydrogen storage to magnesium battery. Angew Chem Int Ed. 2012;51:9780. [142] Lu Z, Ciucci F. Metal borohydrides as electrolytes for solid-state Li, Na, Mg, and Ca batteries: a first-principles study. Chem Mater. 2017;29:9308. [143] Ikeshoji T, Tsuchida E, Takagi S, Matsuo M, Orimo SI. Magnesium ion dynamics in Mg(BH4)2 (1- x) X2x (X= Cl or AlH4) from first-principles molecular dynamics simulations. RSC Adv. 2014;4:1366. [144] Higashi S, Miwa K, Aoki M, Takechi K. A novel inorganic solid state ion conductor for rechargeable Mg batteries. Chem Commun. 2014;50:1320. [145] Roedern E, Kühnel RS, Remhof A, Battaglia C. Magnesium ethylenediamine borohydride as solid-state electrolyte for magnesium batteries. Sci Rep. 2017;7. [146] Chen J, Chua YS, Wu H, Xiong Z, He T, Zhou W, et al. Synthesis, structures and dehydrogenation of magnesium borohydride–ethylenediamine composites. Int J Hydrogen Energy. 2015;40:412. [147] Lide D. CRC handbook of chemistry and physics, 88th Ed. Boca Raton: CRC Press, 2007. [148] Giorgini MG, Pelletti MR, Paliani G, Cataliotti RS. Vibrational spectra and assignments of ethylene-diamine and its deuterated derivatives. J Raman Spectrosc. 1983;14:16. [149] Iwamoto T, Shriver DF. Vibrational spectra of catena-µ-Ethylenediamine complexes of Zinc(II), Cadmium(II), and Mercury(II) with the formula M(en)X2. Inorg Chem. 1971;10:2428.

216

4 Battery Materials

[150] Allen A, Senoff C. Infrared spectra of tris-ethylenediamine complexes of Ruthenium(II). Can J Chem. 1965;43:888. [151] Bennett AM, Foulds GA, Thornton DA, Watkins GM. The infrared spectra of ethylenediamine complexesII. Tris-, bis-and mono (ethylenediamine) complexes of metal(II) halides. Spectrochim Acta Part A: Mol Spectrosc. 1990;46:13. [152] Skripov AV, Soloninin AV, Ley MB, Jensen TR, Filinchuk Y. Nuclear magnetic resonance studies of BH4 reorientations and Li diffusion in LiLa(BH4)3Cl. J Phys Chem C. 2013;117:14965. [153] Hansen BR, Paskevicius M, Li HW, Akiba E, Jensen TR. Metal boranes: progress and applications. Coord Chem Rev. 2016;323:60. [154] Mannhart J, Schlom DG. Oxide–Tausendsassas für die Elektronik. Phys J. 2005;4:45. [155] Lu X, Xia G, Lemmon JP, Yang Z. Advanced materials for sodium-beta alumina batteries: status, challenges and perspectives. J Power Sour. 2010;195:2431. [156] Tucker MC. Progress in metal-supported solid oxide fuel cells: a review. J Power Sour. 2010;195:4570. [157] Kraytsberg A, Ein Eli Y. Higher, stronger, better: a review of 5 V cathode materials for advanced Lithium-ion batteries. Adv Energy Mater. 2012;2:922. [158] Zhang Y, Feng H, Wu X, Wang L, Zhang A, Xia T, et al. Progress of electrochemical capacitor electrode materials: a review. Int J Hydrogen Energy. 2009;34:4889. [159] Cheng F, Chen J. Metal–air batteries: from oxygen reduction electrochemistry to cathode catalysts. Chem Soc Rev. 2012;41:2172. [160] Hanzig J, Zschornak M, Nentwich M, Hanzig F, Gemming S, Leisegang T, et al. Strontium titanate: an all-in-one rechargeable energy storage material. J Power Sour. 2014;267:700. [161] Valov I, Linn E, Tappertzhofen S, Schmelzer J, Van den Hurk S, Lentz F, et al. Nanobatteries in redox-based resistive switches require extension of memristor theory. Nat Commun. 2013;4:1771. [162] Hanzig J, Zschornak M, Hanzig F, Mehner E, Stöcker H, Abendroth B, et al. Migration-induced field-stabilized polar phase in strontium titanate single crystals at room temperature. Phys Rev B. 2013;88:024104. [163] Hanzig J, Mehner S, Jachalke E, Hanzig F, Zschornak M, Richter C, et al. Dielectric to pyroelectric phase transition induced by defect migration. J Phys. 2015;17:023036. [164] Kröger F. A., Vink H. J. Solid-State Physics. In F. Seitz and D. Turnbull (eds). New York: Academic Press Inc. 1956;3:307. [165] Waser R, Dittmann R, Staikov G, Szot K. Redox-based resistive switching memories–nanoionic mechanisms, prospects, and challenges. Adv Mater. 2009;21:2632. [166] Blanc J, Staebler DL. Electrocoloration in SrTiO3: vacancy drift and oxidation-reduction of transition metals. Phys Rev B. 1971;4:3548. [167] Mohapatra SK, Wagner S. Electrochromism in nickel-doped strontium titanate. J Appl Phys. 1979;50:5001. [168] Hanzig J, Zschornak M, Mehner E, Hanzig F, Münchgesang W, et al. The anisotropy of oxygen vacancy migration in SrTiO3. J Phys: Condens Matter. 2016;28:225001. [169] Xu K. Nonaqueous liquid electrolytes for lithium-based rechargeable batteries. Chem Rev. 2004;104:4303. [170] Xu K. Electrolytes and interphases in Li-ion batteries and beyond. Chem Rev. 2014;114:11503. [171] Armand M, Endres F, MacFarlane DR, Ohno H, Scrosati B. Ionic-liquid materials for the electrochemical challenges of the future. Nat Mater. 2009;8:621. [172] Sakaebe H, Matsumoto H, Tatsumi K. Application of room temperature ionic liquids to Li batteries. Electrochim Acta. 2007;53:1048. [173] Balducci A, Dugas R, Taberna P, Simon P, Plée D, Mastragostino M, et al. High temperature carboncarbon supercapacitor using ionic liquid as electrolyte. J Power Sour. 2007;165:922.

References

217

[174] Brandt A, Balducci A. Theoretical and practical energy limitations of organic and ionic liquidbased electrolytes for high voltage electrochemical double layer capacitors. J Power Sour. 2014;250:343. [175] Appetecchi GB, Montanino M, Passerini S. Ionic liquid-based electrolytes for high energy, safer Lithium batteries. In: ACS Symposium Series, volume 1117. ACS Publications, 2012: 67–128. [176] Eshetu GG, Grugeon S, Kim H, Jeong S, Wu L, Gachot G, et al. Comprehensive insights into the reactivity of electrolytes based on sodium ions. ChemSusChem. 2016;9:462. [177] Chen Y, Zhang X, Zhang D, Yu P, Ma Y. High performance supercapacitors based on reduced graphene oxide in aqueous and ionic liquid electrolytes. Carbon. 2011;49:573. [178] Xu B, Wu F, Chen R, Cao G, Chen S, Yang Y. Mesoporous activated carbon fiber as electrode material for high-performance electrochemical double layer capacitors with ionic liquid electrolyte. J Power Sour. 2010;195:2118. [179] Pettersson F, Keskinen J, Remonen T, von Hertzen L, Jansson E, Tappura K, et al. Printed environmentally friendly supercapacitors with ionic liquid electrolytes on paper. J Power Sour. 2014;271:298. [180] Yamamoto T, Nohira T, Hagiwara R, Fukunaga A, Sakai S, Nitta K, et al. Chargedischarge behavior of tin negative electrode for a sodium secondary battery using intermediate temperature ionic liquid sodium bis(fluorosulfonyl)amidepotassium bis(fluorosulfonyl)amide. J Power Sour. 2012;217:479. [181] Hasa I, Passerini S, Hassoun J. Characteristics of an ionic liquid electrolyte for sodium-ion batteries. J Power Sour. 2016;303:203. [182] Monti D, Jónsson E, Palacín MR, Johansson P. Ionic liquid based electrolytes for sodium-ion batteries: Na+ solvation and ionic conductivity. J Power Sour. 2014;245:630. [183] Menne S, Schroeder M, Vogl T, Balducci A. Carbonaceous anodes for lithium-ion batteries in combination with protic ionic liquids-based electrolytes. J Power Sour. 2014;266:208. [184] Menne S, Vogl T, Balducci A. The synthesis and electrochemical characterization of bis(fluorosulfonyl)imide-based protic ionic liquids. Chem Commun. 2015;51:3656. [185] Reale P, Fernicola A, Scrosati B. Compatibility of the Py24TFSILiTFSI ionic liquid solution with Li4Ti5O12 and LiFePO4 lithium ion battery electrodes. J Power Sour. 2009;194:182. [186] Montanino M, Moreno M, Carewska M, Maresca G, Simonetti E, Lo Presti R, et al. Mixed organic compound-ionic liquid electrolytes for lithium battery electrolyte systems. J Power Sour. 2014;269:608. [187] Elia GA, Bernhard R, Hassoun J. A lithium-ion oxygen battery using a polyethylene glyme electrolyte mixed with an ionic liquid. RSC Adv. 2015;5:21360. [188] Agostini M, Ulissi U, Di Lecce D, Ahiara Y, Ito S, Hassoun J, et al. A Lithium-ion battery based on an Ionic liquid electrolyte, Tin-Carbon nanostructured anode, and Li2O-ZrO2coated Li[Ni0.8Co0.15Al0.05]O2 Cathode. Energy Technol. 2015;3:632. [189] Di Lecce D, Brutti S, Panero S, Hassoun J. A new Sn-C/LiFe0.1Co0.9PO4 full lithium-ion cell with ionic liquid-based electrolyte. Mater Lett. 2015;139:329. [190] Nádherná M, Reiter J, Moškon J, Dominko R. Lithium bis(fluorosulfonyl)imidePYR14TFSI ionic liquid electrolyte compatible with graphite. J Power Sour. 2011;196:7700. [191] Reiter J, Nádherná M, Dominko R. Graphite and LiCo1/3Mn1/3Ni1/3O2 electrodes with piperidinium ionic liquid and lithium bis(fluorosulfonyl)imide for Li-ion batteries. J Power Sour. 2012;205:402. [192] Chagas LG, Buchholz D, Wu L, Vortmann B, Passerini S. Unexpected performance of layered sodium-ion cathode material inionic liquid-based electrolyte. J Power Sour. 2014;247:377. [193] Kim JK, Mueller F, Kim H, Jeong S, Park JS, Passerini S, et al. Eco-friendly energy storage system: seawater and Ionic Liquid electrolyte. ChemSusChem. 2016;9:42.

218

4 Battery Materials

[194] Wang P, Zakeeruddin SM, Comte P, Exnar I, Grätzel M. Gelation of ionic liquid-based electrolytes with silica nanoparticles for quasi-solid-state dye-sensitized solar cells. J Am Chem Soc. 2003;125:1166. [195] Galiński M, Lewandowski A, Stȩpniak I. Ionic liquids as electrolytes. Electrochim Acta. 2006;51:5567. [196] Xue L, Padgett CW, DesMarteau DD, Pennington WT. Synthesis and structures of alkali metal salts of bis [(trifluoromethyl) sulfonyl] imide. Solid State Sci. 2002;4:1535. [197] Zhou Q, Boyle PD, Malpezzi L, Mele A, Shin JH, Passerini S, et al. Phase behavior of Ionic LiquidLiX mixtures: Pyrrolidinium cations and TFSI anions linking structure to transport properties. Chem Mater. 2011;23:4331. [198] Eftekhari A, Liu Y, Chen P. Different roles of ionic liquids in lithium batteries. J Power Sour. 2016;334:221. [199] Bhatt AI, May I, Volkovich VA, Hetherington ME, Lewin B, Thied RC, et al. Group 15 quaternary alkyl bistriflimides: ionic liquids with potential application in electropositive metal deposition and as supporting electrolytes. J Chem Soc Dalton Trans. 2002;24:4532. [200] Matsumoto H, Yanagida M, Tanimoto K, Nomura M, Kitagawa Y, Miyazaki Y. Highly conductive room temperature molten salts based on small trimethylalkylammonium cations and bis (trifluoromethylsulfonyl) imide. Chem Lett. 2000;29:922. [201] Kim JK, Matic A, Ahn JH, Jacobsson P. An imidazolium based ionic liquid electrolyte for lithium batteries. J Power Sour. 2010;195:7639. [202] Seki S, Ohno Y, Kobayashi Y, Miyashiro H, Usami A, Mita Y, et al. Imidazolium-based roomtemperature Ionic liquid for Lithium secondary batteries. J Electrochem Soc. 2007;154:A173. [203] Seki S, Kobayashi Y, Miyashiro H, Ohno Y, Usami A, Mita Y, et al. Lithium secondary batteries using modified-imidazolium room-temperature ionic liquid. J Phys Chem B. 2006;110:10228. [204] Yim T, Lee HY, Kim HJ, Mun J, Kim S, Oh SM, et al. Synthesis and properties of pyrrolidinium and piperidinium bis (trifluoromethanesulfonyl) imide ionic liquids with allyl substituents. Bull Korean Chem Soc. 2007;28:1567. [205] Fang S, Yang L, Wei C, Peng C, Tachibana K, Kamijima K. Low-viscosity and low-melting point asymmetric trialkylsulfonium based ionic liquids as potential electrolytes. Electrochem Commun. 2007;9:2696. [206] Ye H, Huang J, Xu JJ, Khalfan A, Greenbaum SG. Li ion conducting polymer gel electrolytes based on ionic liquid/PVDF-HFP blends. J Electrochem Soc. 2007;154:A1048. [207] Tsunashima K, Yonekawa F, Sugiya M. Lithium secondary batteries using a Lithium Nickelatebased cathode and phosphonium ionic liquid electrolytes. Electrochem Solid-State Lett. 2009;12:A54. [208] Wilkes JS, Zaworotko MJ. Air and water stable 1-ethyl-3-methylimidazolium based ionic liquids. J Chem Soc Chem Commun. 1992;13:965. [209] Henderson WA, Passerini S. Phase behavior of ionic liquid LiX mixtures: Pyrrolidinium cations and TFSI - anions. Chem Mater. 2004;16:2881. [210] Elia GGA, Ulissi U, Mueller F, Reiter J, Tsiouvaras N, Sun YK, et al. A Long-life lithium ion battery with enhanced electrode/electrolyte interface by using an ionic liquid solution. Chem Eur J. 2016;22:6808. [211] Appel R, Becke-Goehring M, Eisenhauer G, Hartenstein J. Imidobisschwefelsäurechlorid. Chem Ber. 1962;95:625. [212] Michot C, Armand M, Sanchez JY, Choquette Y, Gauthier M. Materiau a conduction ionique presentant de bonnes proprietes anti-corrosion, 1995. [213] Elia GA, Ulissi U, Jeong S, Passerini S, Hassoun J. Exceptional long-life performance of lithiumion batteries using ionic liquid-based electrolytes. Energy Environ Sci. 2016;9:3210.

References

219

[214] Piper DM, Evans T, Leung K, Watkins T, Olson J, Kim SC, et al. Stable silicon-ionic liquid interface for next-generation lithium-ion batteries. Nat Commun. 2015;6:6230. [215] Larcher D, Tarascon JM. Towards greener and more sustainable batteries for electrical energy storage. Nat Chem. 2015;7:19. [216] Tarascon JM. Is lithium the new gold? Nat Chem. 2010;2:510. [217] Yabuuchi N, Kubota K, Dahbi M, Komaba S. Research development on sodium-ion batteries. Chem Rev. 2014;114:11636. [218] Nithya C, Gopukumar S. Sodium ion batteries: a newer electrochemical storage. Wiley Interdiscip Rev: Energy Environ. 2015;4:253. [219] Hasa I, Buchholz D, Passerini S, Hassoun J. A comparative study of layered transition metal oxide cathodes for application in sodium-ion battery. ACS Appl Mater Interface. 2015;7:5206. [220] Hasa I, Passerini S, Hassoun J. A rechargeable sodium-ion battery using a nanostructured SbC anode and P2-type layered [Na0.6Ni0.22Fe0.11Mn0.66O2 cathode. RSC Adv. 2015;5:48928. [221] Dunn B, Kamath H, Tarascon JM. Electrical energy storage for the grid: a battery of choices. Science (New York, N.Y.) 2011;334:928. [222] Zhao Q, Hu Y, Zhang K, Chen J. Potassium sulfur batteries: a new member of room-temperature rechargeable metal sulfur batteries. Inorg Chem. 2014;53:9000. [223] Ren X, Wu Y. A low-overpotential potassium oxygen battery based on potassium superoxide. J Am Chem Soc. 2013;135:2923. [224] Ponrouch A, Frontera C, Bardé F, Palacín MR. Towards a calcium-based rechargeable battery. Nat Mater. 2015;15:169. [225] See KA, Gerbec JA, Jun YSS, Wudl F, Stucky GD, Seshadri R. A high capacity calcium primary cell based on the Ca-S system. Adv Energy Mater. 2013;3:1056. [226] Aurbach D, Gizbar H, Schechter A, Chusid O, Gottlieb HE, Gofer Y, et al. Electrolyte solutions for rechargeable magnesium batteries based on organomagnesium chloroaluminate complexes. J Electrochem Soc. 2002;149:A115. [227] Lin MC, Gong M, Lu B, Wu Y, Wang DY, Guan M, et al. An ultrafast rechargeable aluminium-ion battery. Nature. 2015;520:325. [228] Sun H, Wang W, Yu Z, Yuan Y, Wang S, Jiao S. A new aluminium-ion battery with high voltage, high safety and low cost. Chem Commun. 2015;51:11892. [229] Shkolnikov E, Zhuk A, Vlaskin M. Aluminum as energy carrier: feasibility analysis and current technologies overview. Renewable Sustainable Energy Rev. 2011;15:4611. [230] Li Q, Bjerrum NJ. Aluminum as anode for energy storage and conversion: a review. J Power Sour. 2002;110:1. [231] U.S. Geological Survey. Mineral commodity summaries 2015: U.S. Geological Survey. 2015:196. http://dx.doi.org/10.3133/70140094 [232] Fleischer M. Recent estimates of the abundances of the elements in this earth’s crust. Geol Surv Circ. 1953;285. [233] MacFarlane DR, Tachikawa N, Forsyth M, Pringle JM, Howlett PC, Elliott GD, et al. Energy applications of ionic liquids. Energy Environ Sci. 2014;7:232. [234] Otaegui L, Goikolea E, Aguesse F, Armand M, Rojo T, Singh G. Effect of the electrolytic solvent and temperature on aluminium current collector stability: a case of sodium-ion battery cathode. J Power Sour. 2015;297:168. [235] Saha P, Datta MK, Velikokhatnyi OI, Manivannan A, Alman D, Kumta PN. Rechargeable magnesium battery: current status and key challenges for the future. Prog Mater Sci. 2014;66:1. [236] Giffin GA. Ionic liquid-based electrolytes for “beyond lithium” battery technologies. J Mater Chem A. 2016;4:13378. [237] Watkins T, Kumar A, Buttry DA. Designer ionic liquids for reversible electrochemical deposition/dissolution of magnesium. J Am Chem Soc. 2016;138:641.

220

4 Battery Materials

[238] Kamath G, Narayanan B, Sankaranarayanan SKRS. Atomistic origin of superior performance of ionic liquid electrolytes for Al-ion batteries. Phys Chem Chem Phys. 2014;16:20387. [239] Gifford PR. A substituted imidazolium chloroaluminate molten salt possessing an increased electrochemical window. J Electrochem Soc. 1987;134:610. [240] Zheng Y, Dong K, Wang Q, Zhang J, Lu X. Density, viscosity, and conductivity of Lewis acidic 1butyl- and 1-hydrogen-3-methylimidazolium chloroaluminate ionic liquids. J Chem Eng Data. 2013;58:32. [241] Lang CM, Kim K, Guerra L, Kohl PA. Cation electrochemical stability in chloroaluminate ionic liquids. J Phys Chem B. 2005;109:19454. [242] Melton TJ, Joyce J, Maloy JT, Boon JA, Wilkes JS. Electrochemical studies of sodium chloride as a lewis buffer for room temperature chloroaluminate molten salts. J Electrochem Soc. 1990;137:3865. [243] Vestergaard B, Bjerrum NJ, Petrushina I, Hjuler HA, Berg RW, Begtrup M. Molten triazolium chloride systems as new aluminum battery electrolytes. J Electrochem Soc. 1993;140:3108. [244] Fannin AA, King LA, Levisky JA, Wilkes JS. Properties of 1,3-dialkylimidazolium chloride-aluminum chloride ionic liquids. 1. Ion interactions by nuclear magnetic resonance spectroscopy. J Phys Chem. 1984;88:2609. [245] Wilkes JS, Levisky JA, Wilson RA, Hussey CL. Dialkylimidazolium chloroaluminate melts: a new class of room-temperature ionic liquids for electrochemistry, spectroscopy and synthesis. Inorg Chem. 1982;21:1263. [246] Wang H, Bai Y, Chen S, Luo X, Wu C, Wu F, et al. Binder-free V2O5 cathode for greener rechargeable aluminum battery. ACS Appl Mater Interface. 2015;7:80. [247] Jiang T, Chollier Brym M, Dubé G, Lasia A, Brisard G. Electrodeposition of aluminium from ionic liquids: part Ielectrodeposition and surface morphology of aluminium from aluminium chloride (AlCl3)1-ethyl-3-methylimidazolium chloride ([EMIm]Cl) ionic liquids. Surf Coat Technol. 2006;201:1. [248] Jiang T, Brym MC, Dubé G, Lasia A, Brisard G. Electrodeposition of aluminium from ionic liquids: Part II-studies on the electrodeposition of aluminum from aluminum chloride (AICl3])-trimethylphenylammonium chloride (TMPAC) ionic liquids. Surf Coat Technol. 2006;201:10. [249] Gale RJ, Osteryoung RA. Potentiometric investigation of dialuminum heptachloride formation in aluminum chloride-1-butylpyridinium chloride mixtures. Inorg Chem. 1979;18:1603. [250] Elia GA, Marquardt K, Hoeppner K, Fantini S, Lin R, Knipping E, et al. An overview and future perspectives of aluminum batteries. Adv Mater. 2016;28:7564. [251] Xu M, Ivey D, Qu W, Xie Z. Study of the mechanism for electrodeposition of dendrite-free zinc in an alkaline electrolyte modified with 1-ethyl-3-methylimidazolium dicyanamide. J Power Sour. 2015;274:1249.

4.7 Safer electrolyte components for rechargeable batteries Giovanni Battista Appetecchi Abstract: Among the electrochemical energy storage systems, rechargeable lithium batteries are considered very promising candidates for the next generation power sources because of their high gravimetric and volumetric energy density with respect This article has previously been published in the journal Physical Sciences Reviews. Please cite as: Appetecchi, G. B. Safer electrolyte components for rechargeable batteries Physical Sciences Reviews [Online] DOI: 10.1515/psr-2017-0150

4.7 Safer electrolyte components for rechargeable batteries

221

to other cell chemistries. The lithium-ion battery technology is based on the use of electrode materials able to reversibly intercalate lithium cations, which are continuously transferred between two host structures (negative and positive electrodes) during the charge and discharge processes. Commercial lithium-ion batteries commonly use liquid electrolytes based on suitable lithium salts (solute) and organic compounds (solvents). The latter, volatile and flammable, represent serious concerns for the safety of the electrochemical devices, this so far preventing their large diffusion in applications as automotive, storage from renewable sources, smart grids. One of the most appealing approaches is the partial or total replacement of the organic solvents with safer, less hazardous, electrolyte components. Here, a concise survey of ones of the most investigated types of alternative electrolyte components, proposed for safer and more reliable rechargeable lithium batteries, is reported. Graphical Abstract: Organic Solvents

Organic liquid electrolytes

Gel polymer electrolytes Organic-IL polymer electrolytes

lonic Liquids

Organic-ionic liquid electrolytes Lithium salts

lonic liquid electrolytes lonic liquid polymer electrolytes

Polymers Polymer electrolytes

Keywords: electrolyte components, ionic liquids, lithium salts, polymer hosts, rechargeable batteries

4.7.1 Introduction The use of energy is mainly based on the direct, but rather inefficient and polluting, conversion of the chemical energy of fossil fuels because of the lack of highly performing energy storage systems. Electrochemical power sources differ from other ones (e.g. thermal power plants) by the fact that the energy conversion occurs from chemical into electrical without any intermediate step. As a consequence, electrochemical systems show higher energy efficiencies besides to be environmentally friendly. Actually, since

222

4 Battery Materials

the stringent worldwide demand of a more efficient and rationale use of energy, rechargeable electrochemical devices are considered as well as novel power sources. In particular, highly efficient, rechargeable electrochemical storage systems are of topic importance for the development of electric/hybrid vehicles and for large-scale availability of electric energy from renewable sources. Also, the electronic consumer market is requiring more and more performing electrochemical energy storage devices. Lithium batteries are excellent candidates [1, 2] for the next generation power sources because of their much higher energy density with respect to other cell chemistries as shown in Figure 4.7.1. Actually, commercial Li-ion batteries are able to store a gravimetrical energy density value (150 W h kg−1) five times (30 W h kg−1) higher and almost twice (80 W h kg−1) with respect to the lead–acid and nickel–metal hydride systems, respectively. Near-future lithium batteries are expected to achieve or overcome 200 W h kg−1 whereas post lithium-ion systems as Li metal–air are potentially able to exceed 500 W h kg−1.

Gravimetrical energy ( W h kg-1 )

500

400

300

Lead-acid NiCd NiMH Li-ion Near future Li-ion Post Li-ion

200

100

0 Figure 4.7.1: Gravimetrical energy value of different electrochemical energy storage systems [from ref. 101].

Lithium is the most electropositive element, i.e. E0 (Li+/Li°) = −3.08 V vs. E0 (2H+/H2), allowing the realization of electrochemical cells operating between 3 and 4 V, this pushing the energy density of lithium batteries well above that of other electrochemical storage devices. In addition, lithium batteries may be fully discharged without depleting their cycling performance and do not exhibit the “memory effect” typical of nickel-based systems. The lithium battery technology uses electrode-active materials able to reversibly intercalate lithium cations (Li+). For instance, carbonaceous materials (generally graphite) are used as negative electrodes (anodes) whereas lithium transition metal oxides (LMOs) are the positive ones (cathodes). The overall electrochemical process is schematized as it follows:

4.7 Safer electrolyte components for rechargeable batteries

223

6C+Liy MOz $ Lix C6 +Liðy–xÞ MOz Despite the electrolyte is an inert component of the battery, it does play a key role in determining the behavior and the performance of the device. Therefore, an ideal electrolyte should tackle the following requirements: (i) from good to high ion transport properties, (ii) very low electronic conduction, (iii) high oxidation potential and low reduction potential, (iv) compatibility towards electrodes and other cell components, (v) good thermal stability, (vi) low toxicity, (vii) based on sustainable materials and (viii) low cost. Present lithium battery systems use electrolytes based on suitable lithium salts and organic solvents (generally alkyl carbonates), which represent major safety concerns [2–4]. In fact, the presence of flammable and volatile organics (even if trapped within inert polymer matrices as in a few commercial lithium-ion batteries) can lead to a dangerous chain of events such as heat generation, thermal runaway, cell venting, fire and, therefore, rapid cell disassembly [5–8]. For instance, an uncontrolled heat development (due to a series of highly exothermal reactions) might lead to a sharp temperature raise (T > 140 °C) and, therefore, causing combustion of the vaporized organic electrolyte (with oxygen becoming available from the decomposition of the delithiated positive electrode) and catastrophic events (explosion). In addition, it is noteworthy that the electrolyte burning causes decomposition of the lithium hexafluorophosphate, LiPF6, salt and the PVdF-based electrode binder, leading to development of highly toxic hydrofluoric acid (HF) gas. Figure 4.7.2 depicts photographs of a 20 A h lithium-ion battery after venting and burning

Figure 4.7.2: Front (left panel) and side (right panel) view of a 20 A h lithium-ion battery after venting and burning accidents.

224

4 Battery Materials

accidents. These safety drawbacks have prevented, so far, the diffusion and commercialization of lithium batteries for large-scale application as automotive, storage from renewable sources, etc. Several efforts were devoted toward this issue, such as incorporation of additives with protective (lithium carbonate, Li2CO3) SEI-forming ability onto anode materials [9–11] or use of safer cell chemistries (i.e. less oxidizing cathode materials) [12–14]. One of the most promising approaches to overcome the lithium battery safety limitations is the development of chemically stable and intrinsically safe electrolytes, e.g. the replacement of the organic solvents with non-volatile and non-flammable (liquid and/or solid) electrolyte components.

4.7.2 Components for lithium battery electrolytes 4.7.2.1 Solvents A liquid organic compound has to satisfy the following requirements to be used as solvent in lithium battery systems [1–3, 12, 13] such as high dielectric constant, low volatility and viscosity, and compatibility toward electrode materials. The physicochemical properties of the most common electrolyte solvents for lithium batteries are summarized in Table 4.7.1. Cyclic carbonates (namely, ethylene carbonate (EC), propylene carbonate (PC), γ-Butyrolactone (γ-BL)) are generally characterized by high dielectric constant values (which enables them to dissolve the lithium salts) and low volatility but, also, rather higher viscosity, this depleting their ion transport properties. Therefore, linear carbonates (usually displaying low viscosity) are added Table 4.7.1: Physicochemical properties (25 °C) of the most common electrolyte solvents for lithium battery systems [14]. Electrolyte solvent ACN γ-BL DEC DMC DME DMF DMSO EC EMC MF NM PC THF VC

Molecular weight

Boiling point /°C

Dielectric constant

Density /g cm−3

Viscosity /mPa s

41.05 86.09 118.13 90.08 90.12 73.10 78.13 88.06 104.1 60.06 61.04 102.1 72.12 86.05

81.6 204 126 90.1 84.0 158 189.0 248 109 31.5 101.2 241 65.0 162

35.95 39.1 2.84 3.1 7.20 36.71 46.45 89.6 2.4 8.5 35.94 64.4 7.39 127

0.777 1.13 0.975 1.07 0.859 0.944 1.095 1.322 1.00 0.974 1.131 1.19 0.880 1.36

0.341 1.75 0.81 2.4 0.455 0.796 1.991 1.85 (40°C) 0.65 0.333 0.694 2.53 0.46 n. a.

4.7 Safer electrolyte components for rechargeable batteries

225

in order to lower the viscosity of the resulting electrolyte. On the other hand, modest viscosity values result in low flash points (e.g. below 40 °C) [1–3, 12, 13], this enhancing the volatility and flammability of the solvent mixture, and leading to safety concerns. Additionally, organic compounds such as EC and vinylene carbonate (VC) show relevant film-forming ability (i.e. formation of the Solid Electrolyte Interface, SEI) onto carbonaceous anodes, this being considered mandatory for assuring good cycling performance in lithium battery systems [1–3, 12–14]. It is to note, however, that EC melts at 40 °C and needs to be blended with poorly viscous solvents. Therefore, commercial lithium-ion battery electrolytes are mainly based on mixtures (binary or ternary) of EC and linear carbonates, namely diethyl carbonate (DEC), dimethyl carbonate (DMC), and ethyl methyl carbonate (EMC), as the solvents [1–3, 12, 13]. 4.7.2.2 Lithium salts Commercial lithium-ion batteries use LiPF6 [1–3, 12–14] as the electrolyte salt because of its favorable properties, such as high ion conduction, wide electrochemical stability window and good SEI-forming ability, the latter property also preventing corrosion of the aluminum cathodic current collector. However, LiPF6 exhibits a few disadvantages such as poor thermal stability and easy development of HF (in presence of even traces of moisture and oxygen), which is able to dissolve transition metal ions from the cathode material. Therefore, various salts [1–3, 12–15] have been alternatively proposed for replacing LiPF6 such as LiAsF6, LiClO4, LiBF4, LiOSO2CF3 and imide salts (LiN(SO2F)2, LiN(SO2CF3)2, LiN(SO2C2F5)2) which, however, display unwelcome issues such as toxicity (LiAsF6), risk of explosion (LiClO4), low conduction in common solvents (LiBF4), modest or absent film-forming ability to protect the aluminum current collector (LiOSO2CF3 and all imides). Xiao et al. [16] synthesized derivative salts such as LiPF4(C2O2), which shows similar conductivity and higher thermal stability than LiPF6. Xu et al. [17] have reviewed different salt families, i.e. lithium fluoroalkyl phosphates [18–20], lithium chelatoborates [21– 23], lithium chelatophosphates [24, 25], lithium imidazolides [26] and lithium imidazolates [27]. A very interesting salt, initially proposed by Angell and coworkers [23, 28], is LiBOB. Its main peculiarities are (i) high thermal stability and safety, (ii) protective SEI-forming ability onto Al current collectors and graphite anodes, (iii) relatively low environmental impact of its decomposition products (B2O3 and CO2), (iv) capability to reduce Fe2+ and Mn2+ dissolution from LiFePO4 [29] and LiMn2O4 [30], respectively. Conversely, LiBOB displays disadvantages such as: (v) modest solubility in alkyl carbonates, (vi) lower conductivity (i.e. higher viscosity) of LiBOB-based solutions with respect to LiPF6-containing electrolytes [31], (vii) decomposition in presence of LiCoO2 and LiNixCoyMn1-x-yO2 [32], (viii) cell venting phenomena due to oxalate and carboxylate impurities resulting in CO2 formation [32].

226

4 Battery Materials

4.7.2.3 Additives The basic concept of the additive incorporation into lithium-ion battery electrolytes is the safety improvement of these systems without, at the same time, depleting their electric performance. Generally, these compounds are flame-retardants (FRs) or high flash point solvents and may also be used as “single” solvents. Mainly, five different FR additive families (Table 4.7.2) have been investigated: (i) alkyl phosphates, (ii) alkyl phosphonates, (iii) phosphazenes; (iv) hydrofluoroethers (HFEs) and (v) fluorinated esters.

Table 4.7.2: Physicochemical properties (25 °C) of the most common electrolyte additives for lithium battery systems. Electrolyte additive TMP [33] TEP [33] DMMP [36] DMMEMP [38] MFE [42] MFA [43]

Molecular weight

Dielectric constant

Boiling point /°C

Density /g cm−3

Viscosity /mPa s

140.07 182.15 124.08 198.15 250.06 110.06

n. a. n. a. 22.3 >77 n. a. n. a.

180 215 180 280 60 85

1.197 1.072 1.145 n. a. 1.529 1.272

2.0 1.6 1.75 4.85 n. a. n. a.

Xu et al. [33] demonstrated that alkyl phosphates such as trimethyl phosphate (TMP) and triethyl phosphate (TEP) are able to drastically reduce the flammability of common lithium-ion battery electrolytes. However, the non-flammability condition is matched only when the TMP or TEP volume content exceeds 40% whereas remarkable capacity fading is observed already at 5% due to the alkyl phosphate instability toward graphite anodes. In addition, a large alkyl phosphate content increases the electrolyte resistance. The cell performance decay can be mitigated through fluorination of the phosphate alkyl chains. Shim et al. [34] have investigated diphenyloctyl phosphate (DPOF) as additive, which enhances the thermal stability and lowers the irreversible capacity and charge-transfer resistance but, at the same time, promotes the growth of a thick SEI layer onto the anode. Alkyl phosphonates were investigated as either electrolyte additives [35, 36] or solvents [37, 38]. Among them, dimethyl methyl phosphonate (DMMP) and dimethyl (2-methoxyethoxy)methyl phosphonate (DMMEMP) were found by Dalavi et al. [36] and Zhang et al. [38], respectively, not depleting the ion transport properties of the conventional electrolytes and achieving high conductivities when used as single solvents with lithium salts because of their modest viscosity, high dielectric constant and low melting point. The phosphazenes were seen to effectively suppress the flammability of conventional electrolytes [33, 39, 40]. In particular, fluorinated phosphazenes behave as

4.7 Safer electrolyte components for rechargeable batteries

227

self-extinguishing additives. However, their high viscosity affects the ion transport properties of the electrolyte solutions, thus decreasing the cell performance. HFEs are suitable additives for producing safer lithium battery electrolytes because of their negligible flash point, low surface tension, modest viscosity, and low volatility and freezing temperature [41]. Among the HFE family, Arai et al. [42] have investigated methyl nonafluorobutyl ether (MFE). However, inflammable electrolyte solutions are obtained only for very large MFE contents (i.e. from 70 % to 90 %), this remarkably reducing their ionic conductivity. The use of branched HFEs with higher fluorine content, e.g. 2-trifluoromethyl-3-methoxyperfluoropentane (TMMP) and 2-(trifluoro-2-fluoro-3-difluoropropoxy)-3-difluoro-4-fluoro-5-trifluoropentane (TPTP), proposed by Naoi et al. [41] and by Shim et al. [34], respectively, allowed to reduce the amount of additive. Conversely, these compounds show modest miscibility with cyclic alkyl carbonates (EC) and behave as poor solvents. The most interesting among the fluorinated ester family is methyl difluoroacetate (MFA), reported by Tanaka et al. [43], which exhibits higher thermal stability (400 °C vs. 270 °C) in combination with faster ion transport properties with respect to conventional lithium-ion battery electrolytes.

4.7.3 Electrolyte confinement in polymer hosts 4.7.3.1 Basic concept A strategy for reducing the safety issue of lithium batteries is confinement of the organic electrolyte within proper polymer hosts capable to interact with the organic solvent and, therefore, to retain the electrolytic solution (which assures the Li+ conduction). The target is to obtain a quasi solid-state ion-conducting membrane (able to be processed and handled as a solid electrolyte material) which, however, exhibits ion transport properties typical of liquid electrolytes [44, 45]. One of the main peculiarities of the polymeric matrix is depleting the liquid leakage in order to minimize the amount of free solvent through the battery system and, therefore, to enhance the safety of the electrochemical device. Also, suitable polymer host should be electrochemically stable and compatible at the interface with electrodes [44, 45]. The chemical stability of these electrolyte systems, called gel polymer electrolytes (GPEs), is driven by the overall polymer host – solvent – lithium salt interactions. Based on the entity of such interactions GPEs can be classified in weak and strong gel electrolytes. A scheme of the macroscopic structure of a GPE system is depicted in Figure 4.7.3. 4.7.3.2 Weak gel electrolytes This GPE family is characterized by weak polymer-solvent interactions and, therefore, the role of the polymeric host is as well as an inert porous sponge (loaded with

228

4 Battery Materials

lithium cation

anion

solvent molecule

Polymer host

Figure 4.7.3: Schematized structure of a gel polymer electrolyte system.

liquid electrolyte, which is only physically retained by the polymer matrix). Among the most common polymer matrices for weak gel electrolytes, PVdF [46, 47] is the worthiest to be mentioned. Different approaches have been followed aiming to decrease the crystallinity degree of the PVdF matrix: (i) use of branched PVdF-HFP or PVdF-TrFE copolymers rather than linear PVdF monomers [48–69], (ii) incorporation of ceramic fillers such as BaTiO3 [48], Al2O3 [47, 48], SiO2 [47, 48, 50], and TiO2 [47, 51] (even for stabilizing the electrolyte/electrode interfacial compatibility and improving the mechanical properties), (iii) addition of different polymeric components to obtain blend polymer hosts [68–70]. PVdF-based gel electrolytes exhibit fast ion transport properties [71], e.g. room temperature conductivity values generally above 10–3 S cm−1 (Table 4.7.3), in combination with good mechanical characteristics. Conversely, such electrolyte systems are chemically unstable due to phase separation between the polymer host and the liquid electrolyte, this leading to liquid leakage and, therefore, depleting the gel performance and safety. 4.7.3.3 Strong gel electrolytes The establishment of strong polymer-solvent-Li+ interactions results in high retention of the liquid electrolyte, which is more than physically retained by the matrix itself [44, 45], e.g. the solvent molecules are able to enter in the intimate chemical structure of the polymer host (often leading to volume increase). This process,

229

4.7 Safer electrolyte components for rechargeable batteries

Table 4.7.3: Ionic conductivity values of a few gel polymer electrolyte (GPE) systems for lithium batteries. The GPE formulation is reported in mole ratios (if not differently indicated). (*) = −10 °C. GPE sample

P(AN)16.0(EC)56.5(PC)23.0(LiTFSI)4.5 [78] P(AN)16.0(EC)60.0(DMC)20.0(LiPF6)4.0 [79] P(MMA)30.0(EC)46.5(PC)19.0(LiClO4)4.5 [81] 25wt.%[P(EO)10(LiClO4)+10wt.%SiO2]75wt.% (EC1DEC3) [74] P(VDF-CTFE)33.0(EC)42.0(PC)19.0(LiC(SO2CF3)3)6.0 [57] 19wt.%P(VDF-HFP)-81wt.%(1 M LiPF6 in EC/DMC 1:1 wt. ratio) [67] 19wt.%P(VDF-TrFE)-84wt.%(1 M LiPF6 in EC/DMC 1:1 wt. ratio) [67]

Ionic conductivity/S cm−1 –20°C

20°C

60°C

–5

2.0 × 10 n. a. 1.0 × 10–4 7.0 × 10–4 (*)

–3

1.0 × 10 4.0 × 10–3 5.0 × 10–4 2.2 × 10–3

3.0 × 10–3 1.0 × 10–2 2.0 × 10–3 n. a.

n. a. n. a.

4.0 × 10–4 3.5 × 10–3

3.0 × 10–3 4.2 × 10–3

n. a.

2.6 × 10–3

n. a.

involving average solvent-polymer interactions, is generally named swelling and gives chemically/physically stable gel electrolytes, even if their mechanical properties are generally worse with respect to weak gel electrolytes. Strong GPEs based on high molecular weight PEO have been extensively investigated by Choi et al. [72], Kang et al. [73], Aihara et al. [74, 75] and Appetecchi et al. [76, 77], who have observed ion conduction values approaching 10–3 S cm−1 even at low temperatures (–10 °C). PEO gel electrolytes exhibit modest mechanical characteristics that, however, can be significantly improved by addition of ceramic fillers [74–77]. PAN-based gel electrolytes [78–80] have shown fast ion transport properties (4 × 10–3 S cm−1 at room temperature) in combination with a wide electrochemical stability window (>5 V). The incorporation of fillers [79, 80] was seen to improve the mechanical stability and solvent retention, resulting in time-stable ionic conductivity even upon prolonged storage times at medium-high temperatures in open cells [80]. GPEs based on PMMA were found to be much more chemically stable because of the stronger polymer– solvent interaction with respect to the PAN-based ones but, at the same time, exhibit slower transport properties [81–84]. Another approach, firstly proposed by Wieczorek et al. [85], is represented by the adoption of polymer blends, aiming to prepare gel electrolytes with improved properties with respect to those of the individual polymeric materials. Particularly, blend materials characterized by three-dimensionally interpenetrated, co-continuous microstructures are of most interest, as reported by Willemse et al. [86, 87] and Jordhamo et al. [88], because of the possibility of combining polymer materials having different peculiarities. In this scenario, Passerini et al. [89], Momma et al. [90] and Nara et al. [91] have developed gel electrolytes based on co-continuous, two-component polymer blends formed by polystyrene, i.e. assuring mechanical stability to the electrolyte membrane, and PEO, i.e. enabling ionic conductivity through swelling in the electrolytic solution. Room temperature

230

4 Battery Materials

conductivities approaching 10–3 S cm−1 were achieved by these electrolytes. Finally, gel electrolytes based on thermally cross-linked poly(fluorosilicone)-ethylene oxide matrices, investigated by Appetecchi et al. [92, 93], have exhibited high ion conductivities (>10–3 S cm−1 at room temperature) and liquid uptakes in combination with wide electrochemical stabilities and excellent mechanical properties. The room temperature conductivity values of various weak and strong GPE systems are summarized in Table 4.7.3.

4.7.4 Electrolytes based on ionic liquid media 4.7.4.1 Properties of ionic liquids Ionic liquids (ILs), salts molten at room temperatures or below, are a unique new fluid class since their numerous peculiarities such as extremely poor flammability, negligible vapor pressure, remarkable ionic conductivity, high thermal, chemical and electrochemical stabilities, low heat capacity, ability to dissolve inorganic (including lithium salts), organic and polymeric materials, hydrophobicity or hydrophilicity [94–97]. ILs have attracted growing attention as electrolyte components (solvents or additives) for replacing the organic compounds currently used (as solvents) not only in lithium batteries [98, 99], but also in a wide variety of electrochemical devices [100–105], resulting in improved safety in case of overheating/overcharging which can lead to venting/burning/explosion. The safety issues of ILs were proved through flammability and volatility tests illustrated in Figure 4.7.4. No flaming and cell volume change (due to internal pressure increase) were observed in ILs even upon prolonged exposition to fire or heating.

Organic electrolyte

lonic liquid electrolyte

Figure 4.7.4: Flammability (upper panels) and volatility (lower panels) tests carried out on conventional organic (left panels) and ionic liquid (right panels) electrolytes.

4.7 Safer electrolyte components for rechargeable batteries

231

The possibility to combine a wide variety of cations and anions gives the freedom to design the most suitable ILs for matching the requirements of the final electrochemical device. Viable lithium battery ILs are formed by alkylimidazolium, saturated alkyl quaternary ammonium [cyclic (pyrrolidinium, piperidinium) or acyclic (tetraalkylammonium)] cations in combination with hydrophobic perfluoroalkylsulfonylimide anions. The chemical structure of the most common IL cations and anions as lithium battery electrolyte solvents is depicted in Figure 4.7.5 whereas their physicochemical properties are summarized in Table 4.7.4.

R1 R2

N

R4 +

R1

R3 Tetraalkylammonium

N +

R2

Pyrrolidinium

O S

CnF2n+1

O R1

N

N +

N

O S

CnF2n+1

O

Per(fluoroalkylsulfonyl)imide

R2

Imidazolium

R1

N +

R2

n = 0, 1, 2, 3, 4

Piperidinium R1, R2, R3, R4 = methyl, ethyl, propyl ... dodecyl

Figure 4.7.5: Chemical structure of a few of the most viable IL cations and anions for lithium battery systems.

Table 4.7.4: Physicochemical properties (20 °C) of the most viable ionic liquids as electrolyte components for lithium battery systems. (*) = platinum metal as the working electrode. [a] http:// en.solvionic.com/products/1-ethyl-3-methylimidazolium-bisfluorosulfonylimide-99.9; [b] http:// www.sigmaaldrich.com/catalog/product/sial/11291. Ionic liquid PYR13FSI [107, 123] PYR13TFSI [152] PYR13BETI [106] PYR14FSI [107, 123] PYR14TFSI [106] PYR14BETI [106] PYR1(2O1)TFSI [106] PYR1(2S1)TFSI [108] EMIFSI [a,b] EMITFSI [106] EMIBETI [106]

Molecular weight

Melting point/°C

308.36 408.37 508.29 322.36 422.41 522.32 424.38 440.44 291.30 391.31 515.35

–13.6 6.1 6.5 –20.0 −6.5 8.9 < –40 < –40 –13.0 –9.7 –1.3

Density Viscosity /mPa s /g cm−3 1.343 1.432 1.517 1.310 1.399 1.482 1.459 n. a. n. a. 1.524 1.599

66 73 204 45 95 348 71 n. a. n. a. 36 98

Ionic conductivity/S cm−1 –10 °C 20°C 8.2 × 10–5